Value in Time: Better Trading through Effective 0470118733, 9780470118733

Citation preview

Value in Time

Founded in 1807, John Wiley & Sons is the oldest independent publishing company in the United States. With offices in North America, Europe, Australia, and Asia, Wiley is globally committed to developing and marketing print and electronic products and services for our customers’ professional and personal knowledge and understanding. The Wiley Trading series features books by traders who have survived the market’s ever changing temperament and have prospered—some by reinventing systems, others by getting back to basics. Whether a novice trader, professional, or somewhere in-between, these books will provide the advice and strategies needed to prosper today and well into the future. For a list of available titles, please visit our Web site at www.Wiley Finance.com.

Value in Time Better Trading through Effective Volume

PASCAL WILLAIN

John Wiley & Sons, Inc.

C 2008 by Pascal Willain. All rights reserved. Copyright


Published by John Wiley & Sons, Inc., Hoboken, New Jersey. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 646-8600, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permissions. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Willain, Pascal, 1959– Value in time : better trading through effective volume / Pascal Willain. p. cm. – (Wiley trading series) Includes bibliographical references and index. ISBN 978-0-470-11873-3 (cloth) 1. Investment analysis. 2. Stocks–Charts, diagrams, etc. I. Title. HG4529.W539 2008 2007049358 332.63′ 2042–dc22 Printed in the United States of America. 10

9

8

7

6

5

4

3

2

1

The world’s ever-growing population increasingly affects our daily lives. Like never before, this demographic challenge is forcing us to address critical issues: boosting the production of basic foods, finding new sources of energy, recycling base metals, and tackling environmental issues, to name a few. In my view, the stock market provides people with a simple way to take part in these high-growth sectors by investing in companies that offer solutions to these problems. However, because many people are too busy just trying to survive, they will not be able to adapt to or even recognize the coming changes. Simply dedicating this book to these people does not help much, but perhaps monetary donations will. I have decided to offer free of charge the Effective Volume tool described in Chapter 1. I welcome, however, donations to the Nello and Patrasche Foundation, a foundation that my wife and I created some years ago for the benefit of handicapped orphans.

Contents

Foreword

xi

Acknowledgments Introduction: Revolution Is at Your Doorstep PART ONE

The Set of Tools That Will Change Technical Analysis

CHAPTER 1 Effective Volume: An Open Window into the Market

xiii 1

9

11

Traders Get a Secret New Tool: A Brief Introduction to the Trading Mechanisms and the Market Players

14

Volume That Moves the Markets

20

Effective Volume

30

Practical Examples of Effective Volume Calculations

40

Technical Section: How to Calculate the Separation Volume

51

Improve Your Trading: Decide on the Big Picture

57

A Comparison with Traditional Tools

59

What We Learned Regarding Effective Volume

66

CHAPTER 2 Price and Value: The Active Boundaries Indicator

67

Buy Low

67

Traditional Measure of “Cheap"

69

Why Do Trends Exist?

81

Grandmothers Are Always Right!

101

vii

viii

CONTENTS

For Math Lovers: How to Calculate the Active Boundaries

109

What We Learned Regarding Active Boundaries

110

CHAPTER 3 When Volume Diverges from Price

113

Effective Volume: Two Arrows from One Bow

114

Price and Effective Volume Trends

118

Price-Volume Divergence Analysis

134

Examples of Divergence Analysis

144

How to Set the Optimal Analysis Window

176

Empty Trading Minutes

181

What We Learned Regarding Divergence Analysis

183

CHAPTER 4 Supply and Demand: The Key to Trading

185

Supply/Demand Equilibrium

186

Funds’ Strategies

205

Funds and Market Manipulation

211

What We Learned Regarding the Supply Analysis

218

PART TWO Trading Strategies

219

CHAPTER 5 Performance: The Risk/Return Balance

221

The Trading Strategy

223

Optimizing Profits

226

Minimizing Risks

245

Measures of Risk-Adjusted Performance: The Sharpe and Burke Ratios

260

What We Learned in This Chapter

262

CHAPTER 6 Automated Trading Systems

265

Production of Trading Signals

268

Trading Strategies

274

What We Learned in This Chapter

317

Contents

ix

PART THREE The Bonus Section

319

CHAPTER 7 The Market Is a Two-Way Street: Shorting Strategies

321

The Short Sale “Tick Test” Rule

321

How to Use This Book’s Tools for Short Trading

322

What We Learned in This Chapter

340

CHAPTER 8 Market and Sector Analysis

341

When Is the Market Becoming Expensive?

342

Sector Analysis

350

What We Learned in This Chapter

361

Conclusion

363

Data Providers

371

Sources

373

About the Author

375

Index

377

Foreword

ou have opened a revolutionary book that explodes the envelope of standard technical analysis. It introduces several new tools that can help you recognize when a trend is likely to reverse. It reveals new ways to profit from trends and their reversals. Pascal Willain used inexpensive off-the-shelf software to slice each trading day of a stock into one-minute segments, like cutting a long stick of salami into thin slices. He measures each slice to see whether prices rise or fall during that minute and discards the minutes with no changes. He finds the average one-minute volume for the day and separates the minutes with price changes into those with above-average or below-average volume. In each group, he adds up the volume of minutes with rising prices and subtracts the volume of minutes with falling prices. This gives him two cumulative volume lines: one for the minutes with above-average volume and the other minutes with for below-average volume. He named them Large and Small Effective Volume. Pascal explains that the minutes with above-average volume reflect the impact of the big money. He discovered that Large Effective Volume often has predictive value. When you find a condition in which the big money starts pushing up a stock while the small money remains negative or neutral, an upside reversal is in the cards. When the big money starts pushing the stock down while the small money is flat or buying, a downside reversal is more likely. Pascal compares his method to dropping down to the cell level and predicting the movement of the entire organism from the behavior of individual cells. He first described his concept of Effective Volume in the interview for my book Entries & Exits, but he goes much further in his new book. The author introduces another key concept, which he calls Active Boundaries. His research shows that the group of professional traders in any given stock is relatively stable and they shoot for relatively steady gains. When the returns from a stock over a period of time reach their Upper Active Boundary, the expectations for a further rise diminish and a

Y

xi

xii

FOREWORD

downside reversal is more likely. When a stock declines and hits its Lower Active Boundary, bullish expectations become high and the stock has a greater probability of an upside reversal. Numerous charts show how to catch reversals using these concepts. In addition to Effective Volume and Active Boundaries, Pascal describes several other concepts. He even provides what he calls “more complex examples for a second reading.” Pascal has a very rare ability to stand apart from the crowd, to question accepted concepts, and to come up with new ideas. For example, while acknowledging his debt to my Force Index, he stands the original concept on its head by asking why not have a Weakness Index, and even suggests its formula. This book abounds with examples of Pascal’s unorthodox approach. For example, he addresses a commonly heard rule, “You must buy when everyone else is selling,” and writes: For me, this is a sure recipe for financial disaster. There are only two clear times when you should buy: 1. You buy when everybody else is buying, but you do it early in the

trend. 2. You buy when everybody else has stopped selling. In other words, you buy when the supply of shares has dried up, when only a few shares are available for sale. You have to invest time and energy in reading this book. Pascal, like many original thinkers, follows his own train of thought, sometimes leaving less prepared readers behind. During the past year I have been receiving the analytic e-mails in which Pascal shares his research into current markets. It took me a little while to catch on to these concepts, and I hope that e-mails from readers will prompt Pascal to offer both his software and his analyses to the wider public. Publishing a book is like giving birth to a baby. This baby will require a bit of nurturing to grow and become strong enough to stand on its own. The reader must keep in mind that technical analysis alone is not enough to enable one to become a successful trader. Money management is essential for controlling risks, and you need good record keeping to learn from your profits and losses. I expect the concepts of Effective Volume, Active Boundaries, and others in this book to become accepted by many serious traders. As always, the early adopters will reap the greatest rewards. —D R . A LEXANDER E LDER New York City December 2007

Acknowledgments

f the Japanese government had not offered me a scholarship to study applied mathematics in the 1980s, I never would have been able to even think about creating new tools for the stock market. If my friend Bob had not advised me some years ago to study technical analysis and buy Dr. Alexander Elder’s book, I would certainly by now be leading another life. If Dr. Elder had not recognized the novelty of my approach, and had not written about it; if he had not first introduced me to his own agent, Ted Bonanno, who helped negotiate the publishing contract, and then to Matthew Kushinka, my copy editor, who helped me find the proper flow of words in the English language; if Dr. Elder had not advised me on the book structure, the style, the title—even the look and feel of the cover— this book, frankly, would not exist. I am truly grateful for Dr. Elder’s help, especially considering how busy he is. I also want to thank the four early readers of this book: Bob Grush, from the United States; Barry Silberman, also from the United States; ´ eric ´ Fred Snoy, from Belgium; and Thanassis Stathopoulos, from Greece. These four readers are independent traders who continuously look at improving their trading. Their comments and suggestions were of immense help to me, and I will always be grateful to them for their support. The last word is for my loving wife and life partner Michiko. Thank you for your continuous love and support.

I

xiii

INTRODUCTION

Revolution Is at Your Doorstep

n the eighteenth century, a Japanese rice trader named Munehisa Homma noticed that it was possible to predict the evolution of prices by studying certain patterns of past prices. He invented what is now called candlestick analysis, still one of the most widely used technical analysis tools. He assumed at that time that current prices represent all known information about the markets. This hypothesis is still shared by many professional traders, although we will see how limited it can be. In the twentieth century, many improvements in technical analysis appeared as new ideas emerged. These include Fibonacci retracements, Elliott wave analysis, moving average convergence/divergence (MACD) lines, and stochastics, to name a few. In 2001, lightning struck, but it went largely unnoticed. Why? As has often been the case throughout history, this revolution was the natural result of a change that had different causes. Everybody noticed the change, but very few noticed the revolution. It was similar to Louis Pasteur’s discovery of microbes. That was a revolution, but the real change that made that revolution possible was the invention of the microscope. The change that would bring about a revolution to the technical analysis of stock trading was decimalization. It happened on April 9, 2001, when traders began to measure stock prices to the penny instead of in sixteenths of a dollar (or 6.25 cents). The objective was to make the stock price fluctuations easier for the general public to understand (thereby attracting more retail investors), as well as to reduce the spread cost. On the contrary, as we will see later, this change had a large impact on the way institutional investors play the market. Decimalization killed market visibility and, as

I

1

2

VALUE IN TIME

some believe, may have encouraged price manipulation. At the same time, decimalization allowed the precise detection of traders’ movements. What you will find in this book is not just one tool, but a complete set of revolutionary tools. These tools are based on how market players act—not on their behavior or on their potential reactions, but on their real, tactical moves. These tools are so powerful that I believe they will eventually be programmed into your favorite stock-trading platform. You probably know that kids learn new languages much more quickly than grown-ups do. An adult’s brain is already formed (the synapses are already connected), so changing the brain takes more than just installing some new wiring. Similarly, it will be easier for beginning traders to read and understand this book than for confirmed traders. Most confirmed traders have years of experience that have crystallized their habits; they have automatisms that follow given patterns and chart formations. Bring them a new idea, and doubt will set in, endangering their whole trading system. However, if you are not already set in your trading ways, if you are not a monk who comes out of his cell to pray to the gods of trading at specific hours, then you will greatly enjoy this book. I will lead you down unexpected paths to a complete new vision of stock trading. Since the time I started doing technical analysis, I have been awestruck by the specific art of charting and chart interpretation. There are so many superb books on the market explaining how to interpret chart patterns. When I try to interpret chart patterns, I feel like an amateur musician who reads an unfinished sheet of music and tries to figure out how it could develop into a full concerto. Technical analysis has become a true art form, and I am thrilled when I meet artists who have mastered their art. Unfortunately, these great artists are on the verge of extinction. With the advent of computer trading and with the appearance of new tools such as those presented in this book, it is clear that traditional chartists will soon be forced to adapt in a new technical environment. In fact, everybody will have to adapt. The markets will be very harsh to those who do not. At the outset of this book, I would like to thank one of the great artists in chart reading: Dr. Alexander Elder. I came to technical analysis after I read two of Dr. Elder’s books: Come into My Trading Room and Trading for a Living. I applied his methods, but even if I turned a good profit, I was not totally satisfied. I first tried to improve upon Dr. Elder’s methods, but soon I realized that I had to start from scratch. I came up with a set of concepts and tools and developed them into a full trading method. I sent Dr. Elder a technical paper that explained my tools, and we met in Amsterdam. I was a bit nervous when we met: Would Dr. Elder listen to me? Would he like the ideas? To interest Dr. Elder in my work, I even told him that my method was a continuation of his work, when the truth was

Revolution Is at Your Doorstep

3

that even if I had learned the basics of technical analysis through his incredibly eye-opening books, I was coming up with totally new concepts. For 30 minutes I presented my tools to Dr. Elder, while he asked for more information: Show me this and that. How do you calculate this or that? How do you get your data? I replied, “Just as you wrote in your book,” or “It is like the method you explain in your book.” But Dr. Elder looked at me and said in a deep baritone, “Noooooo . . . this method stands on its own feet.” I was indeed lucky that Dr. Elder had looked at my tools the way a doctor diagnoses a patient—by concerning himself with the facts. Later, when Dr. Elder visited me in Belgium, he told me that I should not be so humble about my method. Instead of introducing it in a sort of technical manual, I should call it what it is—a revolutionary method—and explain it in terms that are as simple as possible. He then went through the structure of this book, simplifying and reorganizing it to make things easier to understand, and advising me on the types of figures I should use and how to present them. When I told him that he should become coauthor of the book, he then replied, “Noooooo . . . it is not my method. Just write ‘Thank you, Alex’ somewhere in the book. That’ll do it.” Well, here you are, Alex: Thank you!

WHAT IS THIS REVOLUTION ABOUT? First, I need to say that I am a dumb engineer, of sorts: I keep asking dumb questions, and I will continue to ask those questions until satisfactory answers are found that are also experimentally proven to be correct. Here are a few questions for which I could not find satisfactory answers in the literature:

r How can you find out what institutional players are doing? Are insiders r r r r

buying or selling? How can you see that news is coming? When is a stock cheap? When is it expensive? Why do trends exist? Where do they come from? What is the supply/demand equilibrium?

This book is therefore about finding out what insiders are doing, what large funds are doing, what traders’ expectations are, and how the equilibrium between supply and demand evolves. It is also about understanding when large funds are moving in and will eventually establish a new price trend, as well as knowing what buying power is necessary to support a trend, what will break trends, and when trends will be broken. In short, it is about reading the markets instead of guessing about them.

4

VALUE IN TIME

Traditional charts look very complex, because they are based mainly on prices. It is always difficult to make a decision based on one single piece of information (price), even if a price chart is supposed to include all the information about the market. The complexity lies in the guesswork, something at which traders need to become skilled if they want to be good traders. Many books talk about large funds as the “smart money.” Smart money is a term I have a hard time accepting, because it implies that the individual trader is not smart. I prefer to say that information leaks and price manipulations are routine occurrences in the markets. The objective of these leaks and manipulations is to take advantage of others. I would not call that smart. The tools that I developed will not allow you to become smart in the way that most traders define it. But they will allow you to see through manipulation. You will become smarter, then, if you learn to see the markets more clearly and if this transforms you into a better trader. I disagree with people who reduce the market to a competition between large players and individual traders, or who believe that market makers are behind the price moves. The market is much more complex than that, with an increasing number of connected, online traders scattered around the world, with 50 percent of the trades being computer-generated, and with large players often moving in opposite directions. You will, however, have to keep the following point in mind: The new tools for reading the market that I show you will not enable you to read the market exactly, all the time, and forever. Markets evolve, and I believe that all tools for analyzing the markets must evolve, too, including mine.

HOW THE BOOK IS ORGANIZED This book is divided in two regular parts, followed by a bonus section. The first part describes in detail the four new tools that I developed in order for each of them to find a solution to a specific problem. The second part integrates the various tools into trading strategies, and I show what works and what does not for either retail players or fund managers. Part Three, the bonus section, shows how to adapt the tools to sector analysis.

Part One: The Set of Tools That Will Change Technical Analysis I developed the sets of tools presented in this section because I needed them for my own trading. Each tool addresses a specific issue with only one goal in mind: to better understand the market. I am convinced that

Revolution Is at Your Doorstep

5

these new methods of measuring the market have the power to change technical analysis as we know it today. The industry will be slow to adapt, but in the end, I believe that the tools that provide a better measure of the market forces will prevail. Chapter 1: Effective Volume: An Open Window into the Market In Chapter 1, I show that the monitoring of the volume involved in small price changes from one trading minute to the next, which I define as the Effective Volume, is a very good tool to detect tactical moves by insiders and large players. The Effective Volume tool is excellent for detecting trendsetters and often allows the detection of coming price changes. I also review in comparison how traditional tools use volume data. Chapter 2: Price and Value: The Active Boundaries Indicator Chapter 2 deals with the monitoring of price trends. It is based on the hypothesis that the group of active traders who follow a stock is relatively stable and that their automatic trading tools use buy/sell strategies that do not evolve in time. The Active Boundaries indicator takes advantage of this stability to capture trends between boundaries of expectation: The price of a stock has a great probability to reverse up when it hits the Lower Boundary (where expectation is the highest) and to reverse down when it hits the Upper Boundary (where expectation is the lowest). Chapter 3: When Volume Diverges from Price In Chapter 3 I show that the historical comparison of price trends to Effective Volume trends allows detecting, for a specific stock, thresholds that define levels of high accumulation or distribution. This Divergence Analysis, after being adjusted for volatility discrepancies between price and volume, produces buy and sell signals that prove to be very effective. I then show how the combination of Active Boundaries and Divergence Analysis can lead to a set of trading rules that could be combined to form a trading strategy. Chapter 4: Supply and Demand: The Key to Trading Chapter 4 takes a hard look at the supply/demand equilibrium as the major market force. This study leads to the presentation of the Supply Analysis tool. The Supply Analysis tool is based on the calculation of the probability that a share will be offered for sale, depending on the price at which it was bought, the time elapsed since it was bought, and the price evolution since then. I then show in practical examples how the Supply Analysis tool, combined with the Effective Volume tool, can very effectively measure the supply/demand equilibrium and therefore lead to winning trades. This chapter then moves on to study how funds play in an illiquid environment. It shows that funds have great difficulty making money, primarily

6

VALUE IN TIME

because of the large size of the positions they must take. Finally, we discover that markets are very efficient and that therefore price manipulation by funds is not likely to occur. This leads us to the conclusion that traditional funds will not be able to beat the market.

Part Two: Trading Strategies After developing in the previous section a set of new tools that can be used independently, this section shows how these tools can be combined into various trading strategies. These trading strategies are tested against a buy/hold trading method not only in terms of risk/return balance, but also in terms of the total efforts that a trader must invest in order to sort out the best trading opportunities. Chapter 5: Performance: The Risk/Return Balance Chapter 5 shows that at the level of the trading strategy the risk/return balance is best measured using:

r For the risk: the expected monthly loss transferred (MLT) by the trading strategy to the portfolio.

r For the return: the yearly expected return (YER) of the trading strategy. A performing trading strategy must produce a YER higher than that produced by a standard buy-and-hold strategy. We will also see in Chapter 5 how the profit target, the stop loss, and the time limit parameters can be used to manage an opened trade. Chapter 6: Automated Trading Systems Chapter 6 first reviews the alert and production screens, two information displays used to alert the trader to the evolution of a set of stocks. These two screens summarize the material covered in Chapters 1 through 4. Chapter 6 then combines the tools introduced in Chapters 1 through 4 into various trading strategies. We gradually discover the characteristics of the three pillars of good trading strategies: the discovery of value, the selection of the right buying trigger, and the management of the trade evolution.

Part Three: The Bonus Section This part is called a bonus section because it opens the door to the understanding of different facets of trading using the trading tools introduced in Chapters 1 to 4.

Revolution Is at Your Doorstep

7

Chapter 7: The Market Is a Two-Way Street: Shorting Strategies After explaining the “tick test” rule, Chapter 7 studies how the various tools can be used for shorting strategies. Chapter 8: Market and Sector Analysis Chapter 8 shows how the application of the Active Boundaries tool to the general market trend can help in determining when the market is overpriced. It then studies how a modified version of the Effective Volume tool can be used to study sector movements.

READY FOR THE REVOLUTION The revolution of decimalization is here to stay. The only way that investors and traders can avoid becoming victims of insiders and manipulators is to use techniques that detect their moves. This is why I believe that tools such as those I present in this book will be widely used in the coming years. You will see that the different concepts introduced in this book are very simple in nature. The mathematics may look complex at first, but in fact it is mostly addition, subtraction, multiplication, and division. It is important that you understand what the math represents, what it measures, and how you can take advantage of what it tells you. But it is not the math that is doing the trading; it is the trader.

PART ONE

The Set of Tools That Will Change Technical Analysis

CHAPTER 1

Effective Volume An Open Window into the Market

W

hen you are considering a stock to trade, you have to view yourself as a doctor treating a patient. You have three points of view to help you in your diagnosis:

1. The patient’s general condition: age, gender, any preexisting condi-

tions, regular exercise or not, smoking or heavy drinking, and so on. 2. The patient’s symptoms: pain, fever, swelling, and the like. 3. The patient’s internal examination: a blood test, a scan, an X-ray, and

so on. When analyzing a stock, you may think that the general condition is the fundamental analysis: earnings, profit growth, and so on. It may disappoint you to learn that these are only external measures of value. Value itself is useless if not compared to how it is priced. How value is priced is also virtually useless if you do not know what the expectation of shareholders is. Indeed, being a shareholder means possessing equity (value) for which the shareholder expects a return. You will understand in Chapter 2 that the general condition of a stock is partly represented by its price trend. You will often see a price moving above and then below its price trend, indicating the evolving perception of value. Good trading requires you to catch this perception of value. I translate it into the measure of the expectation of active traders. You need to position your trades in harmony with this expectation: buy

11

12

VALUE IN TIME

when expectation is high, and sell when it is low. Chapter 2 explains this expectation concept and how to measure it. It is my first pillar for successful trading. The second thing to look for when diagnosing a stock is its symptoms. Today’s technical analysis is still performed at the level of the symptoms: Traders like to catch trends and their reversals, they will look for overbought and oversold situations, they will search for crowd movements, they will examine the demand/supply equilibrium, and so on. These traders are like doctors who look at a fever and know that after a few days the fever should dissipate. These traditional analysis tools are very useful if you master the art of interpreting them. Traders, like doctors, need a fair amount of experience to become truly skilled. Only then will they be able to see in the charts where a market or an individual stock is heading. These traditional tools require skills, training, and thinking. The great majority of traders use these tools with end-of-day data and react in unison during the following trading session. The “doctors” will see similar symptoms and will prescribe similar treatments (though this is not always the case). I also use these traditional technical analysis tools, because it is critical to see what others see to know where the technical analysis will push the crowd of traders. A doctor who has a doubt about a patient’s diagnosis will order a blood sample to be analyzed; the doctor can then diagnose the disease and prescribe the necessary medicine. Now, suppose that it were possible for a doctor to insert a tiny microscope inside the patient’s body, and that this microscope had a wireless communication with the doctor’s health monitoring station. The doctor could then monitor the fever not only after it appears (when the patient has already become sick), but before the fever appears, by monitoring any conditional change occurring at the microscopic level. The doctor could sort those changes and take into consideration only those that might cause a fever. Of course, this capability doesn’t yet exist in medicine. Similarly, what is lacking in today’s technical analysis is a way to detect micro changes that are strong enough to propagate over time into a full-blown sickness. A very useful tool that I present in this book therefore allows traders to reach what might be called the cell level. Going down to the cell level does not necessarily mean analyzing each transaction, looking over the trading book size, or studying all the coming orders. You need to look at the microbes through your microscope, but remember that you are more interested in seeing their propagation than their mere existence. When Pasteur discovered microorganisms such as viruses and bacteria, it was not finding out they existed that revolutionized medicine but rather the interpretation

Effective Volume

13

of how these organisms work. This interpretation led the way to vaccines that changed our everyday life. I love the work of Russian Nobel Prize winner Dr. Ilya Prigogine and his theories of dissipative structures. I have to confess that it was when I was looking for a way to discover a new tool that I remembered my readings of Prigogine during my student days. Although these theories would not apply to understanding how the stock market works, I found the principles strikingly close to how I believe the stock market functions. The work of Dr. Prigogine states that the dissipation of matter or energy is usually linked to the ideas of efficiency loss and to the evolution toward a larger disorder. However, far from the equilibrium of a structure, the dissipation could be at the origin of new states of matter. In short, life was created by dissipation that brought a system far from the equilibrium and forced it into a new state of order. Prigogine states: Far from the equilibrium, a state of operation can look like an organization because it results from the amplification of a microscopic deviation that at “the right timing” has privileged a reactive behavior as opposed to other reactive behaviors that were also possible. The individual behaviors can therefore in certain circumstances have a decisive role. —Translated by the author from La nouvelle Alliance, by Ilya Prigogine and Isabelle Stengers (Paris: Gallimard, 1979), page 237.

As you may now understand it, the market may be moving en masse, and this pattern has been greatly amplified by the advent of the Internet and fast communications. However, I will show you that many market movements are started at a much lower level and that the broad price trend changes are often triggered by only a fraction of the volume exchanged. Figure 1.1 shows the analogy between the stock market evolutions and the evolutions of an organism. An organism that is in a state of equilibrium first needs to be put out of equilibrium by an external trigger. This external trigger is strong enough to generate a micro change. If this trigger repeats itself for a period of time, it can propagate the change to the whole organism, which will then enter into a new state of equilibrium. I am not saying that we have to forget traditional technical analysis, but rather that traditional technical analysis is less and less adapted to fastmoving markets where information and manipulations are the basis of the market movements.

14

VALUE IN TIME

Living Organisms Strong trigger Micro change

Stock Market

Stable Organism

Strong volume Instability

Spread with time

Price change Spread with time

Organism with a different stablility

FIGURE 1.1 Change in market equilibrium. The price of a stock goes from one state of equilibrium to another. This change has triggered an abnormal increase in volume of transactions, one that is strong enough to trigger micro price changes whose spread will force a change in equilibrium.

TRADERS GET A SECRET NEW TOOL: A BRIEF INTRODUCTION TO THE TRADING MECHANISMS AND THE MARKET PLAYERS Before explaining how things changed in 2001, I would like to point out three basic rules that govern the largest stock markets (NASDAQ, New York Stock Exchange, etc.): 1. The price precedence rule says that if you offer to sell a stock at the

lowest price, your offer will be executed first. (If simultaneously John offers to sell his shares at $10, Jim at $10.01, and Martin at $9.99, Martin’s order will be executed first.) This guarantees that buyers also get the best price for the stock they purchase. Buy orders that offer the highest buying prices are also executed first. 2. The time precedence rule says that buy or sell orders that have the same price are ranked in their order of submission: The first to arrive is executed first. (If John offers to sell his shares at $10, and five seconds later Martin also offers his shares at $10, John’s order will be executed first, followed by Martin’s.) 3. A lesser-known rule is called the public order precedence rule. This rule states that members of a public exchange cannot execute their own orders ahead of orders from the general public that are standing at the same price. This rule was created to increase investor confidence that members of the exchange will not use their superior information to their advantage by trading ahead of the public.

Effective Volume

15

These three rules are applicable for orders only when they reach the market. However, before reaching the market, some orders go through a broker. The broker can just forward the order, or can take advantage of it and trade for himself ahead of the client’s order. This is seldom the case, but dishonest brokers do exist, and the bad behavior of a few is finally pushing human beings out of the loop in favor of electronic order-routing systems. As brokerage houses get a commission on each transaction, some have found out that it is more profitable to trade their clients’ accounts often, despite the fact that eventually they bankrupt their own clients and lose them. This reminds me of a friend who once told me that he had an e-mail exchange with a trading company in the United States; this company was ready to trade his account and, besides the traditional commissions on each transaction, take only 10 percent of his profits as commission. My friend thought that it was a good deal. Indeed, since he did not have to pay any management fee, he believed that this trading company would try to maximize its own profit, which was linked to my friend’s profit. My friend even asked me if I wanted to invest with him. I had to refuse, because I do not let other people manage my money. After only a few months, the first $25,000 he tested in that account had been reduced to almost nothing. He told me at that time that he was not very happy, but that he was calling the trading company on a regular basis; apparently they were always willing to give him an explanation on why they lost his money. Still later, when I met my friend again, he explained to me that after losing his first $25,000, he really wanted to get his investment back. He decided to get “really tough” with the trading company and gave them one last chance—he put in another $12,500. I remember telling him at the time that if the manager of his account was truly looking for a 10 percent profit, he never would have let the account go bust and he would certainly never have accepted my friend’s second investment. Why? The manager would have had to make a $25,000 profit to compensate for the first loss even before being paid one cent out of any profit from the additional $12,500. I advised my friend to take out whatever he could, because he was probably the victim of account churning, which is the term for when trading companies generate as many commissions as possible with useless trades. My friend did not follow my advice and thus learned an expensive lesson, losing his subsequent $12,500 investment.

How Decimalization Changed the Markets Before 2001, prices were quoted in sixteenths of a dollar. Suppose you wanted to buy 1,000 shares of a stock whose bid price was $10.1875 and

16

VALUE IN TIME

whose ask price was $10.25. If there were liquidity at $10.25, you could either (1) get your shares at $10.25 (for a total amount of $10,250) or (2) place a bid at $10.1875 (for a total amount of $10,187.50) and wait for your bid to get filled, hoping that nobody would bid higher than you, buy all the available shares at the ask, and consequently push the price up. The spread cost—the difference between the ask (the best price offered by sellers) and the bid (the best price offered by buyers)—was rather high at $0.0625; for 1,000 shares, the difference between placing the order at the ask and placing it at the bid was $62.50. This high cost would have pushed buyers to place their orders at the bid and sellers to place their orders at the ask. Because of the time precedence rule that prioritizes the execution of orders, traders would place their orders early enough to be executed first. As a consequence, you could have market visibility and guess what large players wanted to do. Indeed, Table 1.1 shows the order size for the bid and the ask before decimalization. Another consequence was that the price did not change much, since it took quite a large volume to move the price up or down by one tick (the smallest level of price change between the bid and the ask). Before decimalization, a tick was one-sixteenth of a dollar, or 6.25 cents. We see in Table 1.2 a similarly sized order book after decimalization. It shows 20,000 shares on the bid, but distributed between $10.19 and $10.13. It also shows 22,000 shares on the ask, distributed between $10.20 and $10.25. As a trader, suppose that you want to order shares at the bid. In Table 1.1, you are competing against 20,000 shares. If you wait longer, the bid could increase, and your chances to get shares at $10.1875 could diminish. Therefore, you will be inclined to rush your order in. However, if you want to buy shares at the bid in Table 1.2, you are competing against only 500 shares. You now have less motivation to place your order at the bid, since competition is not showing up. You prefer to keep your hand closed like a good poker player. If you are lucky enough, someone will sell 600 shares at market price and the bid will be lowered one cent. This may allow you to get your shares at a cheaper price.

TABLE 1.1

Book of Orders (before Decimalization)

Buyers

Sellers

Bid Volume

Bid

Ask

Ask Volume

20,000

$10.1875

$10.25

22,000

17

Effective Volume

TABLE 1.2 Book of Orders (after Decimalization) Buyers

Sellers

Bid Volume

Bid

Ask

Ask Volume

500 3,000 5,500 1,000 5,000 4,000 1,000

$10.19 $10.18 $10.17 $10.16 $10.15 $10.14 $10.13

$10.20 $10.21 $10.22 $10.23 $10.24 $10.25

10,000 3,000 7,000 300 700 1,000

The book of orders lists the prices at which buyers and sellers are ready to trade as well as what volume they want to trade. The spread is the difference between the best bid price (here, $10.19) and the best ask price (here, $10.20). In this example, the spread is $0.01. A buyer wanting to buy 100 shares may buy them at the bid for $10.19 per share and wait in line until the existing bid order of 500 shares is first executed, or may pay one cent more and have the order executed instantaneously at the ask for $10.20.

Furthermore, if you need to buy 12,000 shares in Table 1.2 and you place an order at the ask, the price will move up one tick to $10.21. This may be undesirable, especially since you would still like to buy another 100,000 shares at a good price. If you put large buy orders at the bid, you will show your hand to the market and attract other buyers. The cheapest course of action would be to buy 9,500 shares at the ask, then sell 600 shares at the bid. The ask price would stay unchanged, but the bid price would fall one cent to $10.18. This would eventually cause sellers to lower the ask to $10.19, allowing you to buy your next set of shares at a lower price. It is a legitimate price manipulation that funds need to use in order to accumulate or distribute shares during sideways trading ranges. In Chapter 4, we will see if this manipulation is common practice. In addition to this, program trading would automatically use these tactics to dispose of or purchase large blocks during sideways trading ranges, making sure that the price stays in the trading range until the strategic move is finalized. In conclusion, we have seen that decimalization killed market visibility while favoring price manipulations. Fortunately, the Effective Volume tool that I present in Chapter 1 provides a way to see through tactical moves by large players. It allows all other traders to analyze the repetitive market tactics and show the underlying strategic decisions of large players. The Effective Volume tool does not imply that these strategic decisions are correct and that you need to trade in the same direction.

18

VALUE IN TIME

However, knowing what large players are doing is key to helping your trading decisions, whether you are a retail player or a competing institutional player—especially if you have a correct measure of value. The Active Boundaries tool that I present in Chapter 2 will allow you to obtain a very accurate measure of value. Since I began using the Effective Volume tool, one of my trading principles has been not to trade against large players. This is not to say that I always trade with the large players, but I am not fool enough to trade against them.

How Large Funds Adapted to Decimalization A large fund has the dual advantage of size and power, but it also has limitations. Funds provide liquidity to the markets, and the system is designed to allow them some flexibility. The decimalization rule served them this flexibility on a silver platter. This rule was initially conceived as a way to attract private investors and lower the spread costs, but it was in fact an implicit authorization for large funds to manipulate markets. Indeed, before decimalization, if a fund wanted to lower the price of a stock, it had to sell at the bid enough shares to take out all the outstanding buy orders. Since the spread cost was high, all players entered their order in advance (first come, first served), and it was easy to see what large players wanted to do. Market manipulation at that time was quite costly. (For example, you had to sell perhaps 10,000 shares at a spread cost of $0.0625. This meant that if in fact you wanted to lower the price in order to buy a larger quantity at a lower price, you would have to purchase these 10,000 shares back $0.0625 higher. The manipulation would have cost you $625.) Decimalization, though, lowered the spread cost, and therefore freed large players from disclosing their orders. This greatly reduced the size of the order book, allowing anybody with just a few hundred shares to increase or decrease the stock price. Typically, these days, if you have 1,000 shares, you can easily push the price down by one cent, at a cost of 1,000 × $0.01 = $10, which is 60 times less than before decimalization. I have no proof that markets are constantly manipulated. However, if a service suddenly costs 60 times less than it did the day before, you can be sure that this service will be used more often.

New Tools Are Necessary Further on in this chapter I make detailed comparisons of the different tools that are used to study the price/volume relationship, but what I want to stress here is that a tool is an instrument that you are using to take a

Effective Volume

19

measurement. That measurement gives you some clue about the underlying reality. I like to compare trading a stock through technical analysis to the actions of an engineer who is in charge of a petroleum extraction rig. This engineer is responsible for digging a deep hole and eventually hitting a target. When drilling, the crew will encounter different types of ground texture; resistance and friction will increase. They will also encounter changing heat and pressure conditions. The engineer knows by experience that they will need more than one instrument to understand what is happening to the rig deep down in the hole. Similarly, traders need to use different tools when analyzing a stock. The market is very complex. It is, of course, different from what it was 100 years ago, but it is also more complex than even 15 years ago. Just look at three key changes that have happened since then: 1. Communication speed has resulted in very quick price adjustments

to news. Markets are becoming more efficient, but also crowded with many retail investors enjoying online communication. 2. Decimalization has changed the tactics of large players. 3. Hedge funds are bringing liquidity but also volatility (very large swings of price and volume). A trader needs tools that can handle these changes. Such tools therefore need the following characteristics:

r The tools need to catch the strategic moves using an analysis of the accumulation/distribution tactics. Usually, institutional investors fragment one large order into many small orders that can then be sent undetected—this is called order fragmentation. Each small order will then be executed in one or more market transactions. Although data related to each transaction and each fragmented order is available, the tools need to “reconstruct” the fragmented orders, using minute data so that traders can have a better understanding of what institutional players are up to. r The tools must be able to filter the noise out of the important signal. (We will see later in the chapter that only 25 percent of the total exchanged volume is responsible for 75 percent of the price changes. You’d better know the direction of the 25 percent and not move against the direction of these changes.) r The tools must tell you if the moves of the large players are significant enough to induce a price change or to make or break a trend. r The tools must allow you to make volatility adjustments between price and volume, which carry very different levels of volatility.

20

VALUE IN TIME

r The tools must show you what the position and expectation of other active traders are, because you will need to buy cheap—and a cheap price will be harder to find if everybody else expects the share price to decrease and is ready to sell. A cheap price is found when the last seller has finished selling and new buyers come in with a high expectation for the price to increase. Finally, wouldn’t it be nice if your computer scanned hundreds of stocks, applying all these new tools and giving you buy and sell signals?

VOLUME THAT MOVES THE MARKETS When I started this work, I was almost completely convinced that large players were mainly responsible for stock price movements, because of the large size of their trades. Therefore, monitoring the movements of large players seemed to be the best way to monitor the whole market. My concern was to find out when institutional investors were moving in or out of stocks. The analogy with Dr. Ilya Prigogine’s work was telling me that I needed to do three things: 1. Measure the impact of volume changes to price changes at a level that

was as close as possible to the transactional level. 2. Separate large from small volume. 3. See the evolution of such volume.

Therefore, I needed to be able to compare this evolution between fixed periods of time. I was looking for a tool that could do these things. Because I am lazy, I tried to find an already existing tool, one that I could use right away. I found two categories of tools: the “tick volume” tools and the “end of day” tools. As we will see later in this chapter, both types of tools have their own limitations and therefore neither could meet my needs. Still wondering why nobody had found an answer to an obvious question (“What are large players doing?”), I started to develop my own tools. In trying to answer that question, I realized that there are not that many different types of data to work with: on the time interval basis (1-minute, 5-minute, 10-minute), you can play with the open, high, low, and close of the price data, and add to it the volume data. On the transactional level, you have the order size, the execution time, the execution size, the price of execution, and some other minor information.

21

Effective Volume

TABLE 1.3 One-Minute Data

9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06

14:27 14:26 14:25 14:24 14:23 14:22 14:21 14:20 14:19 14:18 14:17 14:16 14:15 14:14 14:13

Open

High

Low

Close

Volume

$11.07 $11.06 $11.06 $11.05 $11.04 $11.03 $11.02 $11.02 $11.01 $11.02 $11.04 $11.05 $11.06 $11.07 $11.07

$11.07 $11.06 $11.07 $11.06 $11.06 $11.04 $11.04 $11.02 $11.02 $11.02 $11.04 $11.06 $11.06 $11.08 $11.08

$11.06 $11.06 $11.06 $11.05 $11.03 $11.03 $11.02 $11.02 $11.01 $11.02 $11.01 $11.04 $11.05 $11.06 $11.07

$11.06 $11.06 $11.06 $11.06 $11.05 $11.04 $11.04 $11.02 $11.02 $11.02 $11.02 $11.04 $11.06 $11.06 $11.07

5,889 200 28,335 18,131 33,188 3,298 29,658 17,825 11,351 40,889 14,015 13,802 32,536 16,399 20,041

Typical one-minute data format, including the minute opening price (open), the high of the minute (high), the low price of the minute (low), and the closing price of the minute (close). The last column represents the volume exchanged during that trading minute.

I started with the raw one-minute data such as that displayed in Table 1.3. Because the trading day is 6.5 hours long, there is a maximum of 390 trading minutes. Table 1.3 shows a typical data set, each line representing one minute. More recent data are shown at the top of the table. As shown in Table 1.3, each trading minute has an opening price, a high price, a low price, and a closing price. This means that during a trading minute, traders have been buying and selling shares. The price variations between the low and the high indicate that such activity existed. These price variations are usually done tick by tick. Conventionally, upticks and downticks are tiny price movements that move the price up or down by one tick (usually one cent). Within one trading time interval of one minute, there can be several upticks and downticks. From time to time, a volume spike takes the price up or down several ticks at a time. I call price inflections the small price changes that occur between one trading minute and the next. Let’s call the volume that is responsible for a price inflection the Effective Volume. We will see later how to calculate it. Please note that for me a price inflection of one tick has the same weight as a price inflection of two or more ticks, at least to measure the up and down buying and selling movements.

22

VALUE IN TIME

A price inflection indicates that the equilibrium between the bid and the ask was broken because of underlying market activity (traders pushing the price down instead of traders pushing the price up or vice versa). During one trading minute, the equilibrium can be broken on one side and then suddenly reverse to the other side. This usually happens when buyers and sellers trade similar volume sizes with a similar determination. In the course of trading, suppose that an institution intends to place a large buy order. Either this large buy order will go as a block directly between institutions or that institution will have to buy from the market. Large orders placed at market usually push the price up. In order to go unnoticed, a large order has to be fractioned into tiny orders that will be brought to the market on a systematic basis. This must be done without triggering a new uptrend before the whole lot has been bought. This requires a careful tactical execution that involves a mix of order sizes and timing variations. Institutions either use special order-placing algorithms or obtain the assistance of a market maker.

How to Detect Such Movements Only a fraction of the orders that reach the market are executed. Executed orders create transactions between a buyer and a seller. The buyer’s and the seller’s respective buy and sell orders are mutually filled. If you only study the transactions, it is very difficult to see the direction of these transactions. Because for each transaction there is a buyer and a seller, it is impossible to tell if sellers are stronger than buyers. The direction of the trade is indicated by the small price change that occurs on the transaction: If the price increased on the transaction, the buyer was stronger (pushed the price up). Otherwise, the seller was stronger. Because institutions split up large orders into numerous small orders, studying the transaction size does not help in figuring out whether the transaction was generated by an institution. What we therefore need to do is study all the aggregated transactions within regular time intervals of one minute. The idea is to reconstruct the size of the original order by adding up all the transactions that occurred within one minute and to compare that number to the price variation. Let’s study this idea in one example: a large buyer. Let us suppose that an institution wants to buy 100,000 shares on the market of a stock that is trading 500,000 shares per day. The institution will probably have to use one of the following four tactics: 1. Place a large buy order at the bid. 2. Place regular small buy orders at the bid.

Effective Volume

23

3. Place regular small buy orders at the ask. 4. Place a large buy order at the ask.

Let’s study the consequences of such moves. 1. Place a large buy order at the bid. This is a passive strategy, since the

institution has to wait for sellers to come to it. But, since regular buyers are still active, these players would need to bid the price up to get their shares. As a result, the institution will also have to raise its large bid, with the risk of starting a new uptrend. This method is ineffective for accumulating shares. It is easily detectable, since the large bid signals to the market that a large player is accumulating. 2. Place regular small buy orders at the bid. This is also a passive strategy, but the institution will not be easily detected. As we will see in Chapter 4, this strategy does not allow large players to take a significant position, and is therefore probably not often used. 3. Place regular small buy orders at the ask. This strategy is more active, because the institution is actively buying shares. This method requires that the institution have patience in its accumulation, to avoid price spikes that could trigger a new uptrend. However, because the buying is regular during a short period of time, the supply of shares will momentarily dry up and the price will momentarily increase. The institution needs to monitor these small price increases. If the small price increases trigger a change of key technical patterns, they could attract more buyers while the institution has not met its targeted number of shares. On small price increases, the institution must therefore either (1) wait for new sellers to come and push the price back down or (2) push the price back down by itself with a small sell order. Because of the statistical significance of the repetitive buying pattern, even distanced buy orders placed at the ask form a pattern using the Effective Volume, which I explain in the next section. The visible pattern is that large volume will be more often linked to price increases than to price decreases. 4. Place a large buy order at the ask. This very active strategy is used only when several institutions are competing to get shares, or when an institution wants to trigger a price increase by signaling to the market that it is buying shares. This is easily detectable through the monitoring of the price trend. It should now be clear to the reader that what we need to analyze is not the situation at one point in time, but the regularity of the pattern during a set of consecutive, identical time intervals.

24

VALUE IN TIME

Traditional Way to Calculate Shares Accumulation Larry Williams created a now widely used formula to calculate the accumulation/distribution (A/D) balance on daily charts. Figure 1.2 shows such a calculation. The principle is to weight the total volume exchanged during the day by the price gain/loss, divided by the price spread during that day.

r Share accumulation means buying. r Share distribution means selling. The simple idea behind this is to say that if shares are exchanged during the day and the closing price is higher than the opening price, for example, the total result is considered positive: Buyers are stronger than sellers. This means that on average, there is share accumulation during the day. However, if the price spread during the day is very large compared to the gain, it means that traders have been fighting during the day. Therefore, the strength in accumulation of shares should be proportional to the extent of the fight. In Figure 1.2:

r The gain = the outcome of the fight = $10.4 − $10.2 = $0.2. r The spread = the extent of the fight = $10.5 − $9.8 = $0.7. High $10.5 Close $10.4 Gain/Loss Spread Open $10.2

Low $9.8

FIGURE 1.2 Larry Williams accumulation/distribution example #1. Accumulation =

Gain × Volume Spread $10.4 − $10.2 × 100,000 shares = 28,571 shares $10.5 − $9.8

25

Effective Volume

r The number of shares shown by the accumulation calculation is therefore: 100,000 ×

$0.2 = 28,571 shares $0.7

However, if the opening price had been $10.4 and the closing price $10.2, there would have been a loss of −$0.2, and we would have seen the same number of shares (28,571 shares), but on the distribution side. We can note two potential problems with this formula: The first problem is shown in Figure 1.3. Because the spread and the gain shown in Figure 1.3 are identical to the spread and the gain shown in Figure 1.2, the result of the accumulation/distribution calculation is identical: 28,571 shares in both cases. However, some traders will tell you that the close of Figure 1.2 is stronger than the close of Figure 1.3, because the price in Figure 1.2 closed higher. Therefore, the share accumulation shown in Figure 1.2 is maybe more important than the share accumulation shown in Figure 1.3. This is why traders who calculate the accumulation/distribution of shares on the

FIGURE 1.3 Larry Williams accumulation/distribution example #2. Accumulation =

Gain × Volume Spread $10.1 − $9.9 × 100,000 shares = 28.571 shares $10.5 − $9.8

Accumulation as defined by Larry Williams is independent of the relative position of the gain within the high-low range.

26

VALUE IN TIME

basis of price spread during a day will also look at where the close ended compared to the price spread. If the close ended near the high, they would conclude that the buying side had been stronger than the selling side. The second potential problem is a lesser-known one: manipulation of the opening price sometimes exists. A strong opening price may attract buyers, while a strong closing price may attract sellers. A fund that wants to sell a large number of shares could therefore try to set a positive tone by forcing a strong opening. In case of a different opening price, the Larry Williams accumulation/distribution formula yields quite a different result. Figure 1.4 shows that if the opening price had been $10.5 instead of the open of $10.2 shown in Figure 1.2, the Larry Williams formula would have resulted in a distribution of 14,286 shares instead of the accumulation of 28,571 shares calculated in Figure 1.2. My message here is simply that in some cases, the opening price might be less valid as a parameter than the closing price. In general, methods that use end-of-day data could be more vulnerable to price manipulations, since they rely on fewer data points. The comments relative to Figures 1.2–1.4 have not been backed by any research data. The interested reader should refer to Larry Williams’ book Long-Term Secrets to Short-Term Trading.

FIGURE 1.4 Larry Williams accumulation/distribution example #3. This method is very sensitive to opening price manipulations. In this case, an opening price at $10.5 instead of $10.2 results in a distribution of 14,286 shares instead of an accumulation of 28,571 shares. Distribution =

Loss × Volume Spread $10.2 − $10.5 × 100,000 shares = −14.286 shares $10.5 − $9.8

27

Effective Volume

Do Not Trade Like My Grandmother The classic technical tools that use the daily price changes and the daily volume are based on two assumptions regarding volume: The first assumption is the price repartition of volume. These tools suppose that the volume is regularly distributed at every tick between the low and the high prices of the day. Let us take the example of the company Tellabs on September 20, 2006. On that day, during trading hours, about 11 million shares changed hands. The Larry Williams accumulation formula gives: Opening price: $10.35 High price: $10.41 Low price: $10.05 Closing price: $10.29

Distribution =

$10.29 − $10.35 × 11,000,000 $10.41 − $10.05

Distribution = −1,833,333 shares Based on this example, this means that at every tick between $10.05 and $10.41, 297,000 shares were exchanged (see Figure 1.5a) (11,000,000 divided by the difference between $10.05 and $10.41 plus 1 tick, since the subtraction eliminates one of the two ticks at the extremities—the tick of

FIGURE 1.5a The linear volume repartition by price level. Linear volume repartition by price level is the first simplification implied by technical tools that use daily price variations and volume.

28

VALUE IN TIME

FIGURE 1.5b The real volume repartition by price level. The real volume repartition by price level is very different from the linear repartition.

the minute low price or the tick of the minute high price; between $10.05 and $10.41, there are therefore 37 ticks and not 36 as a normal subtraction would show). In reality, the volume exchanged forms an irregular pattern, as shown in Figure 1.5b. The second assumption is the time repartition of volume. The traditional tools suppose that the volume is regularly distributed every minute between the open and the close of the trading day. In our example, this means that at every minute between 9:30 A . M . and 4:00 P . M ., 11,000,000 ÷ 390 = 28,205 shares have been exchanged (see Figure 1.6a). In reality, the daily buying and selling pattern clearly shows that volume came in spikes, and that a large proportion of the transactions occurred at the end of the trading day (see Figure 1.6b). The two assumptions made by traditional tools that use end-of-day data to calculate the accumulation/distribution of shares are so drastic that as a trader, I have little confidence in using these tools, although some traders may find them reliable. Indeed, one big characteristic of volume is that it comes in spikes. Typically, you would see many transactions of 100 shares and then suddenly a single transaction for 10,000 shares, or a set of transactions that would fill many small orders. In short, volume has a very high volatility on a minuteby-minute level. On the day-by-day level, too, volume could jump 100 percent from one day to the next. This volatility is well known to traders whose mantra is “A price increase on a strong volume day is more valid than a price increase on a weak volume day.” This is experience talking, similar to my

29

Effective Volume

Number of Shares

Tellabs: Linear Volume Historam by Time, 09/20/2006

FIGURE 1.6a The linear volume repartition. The linear volume repartition by time level is the second simplification implied by traditional tools that use daily price variations and volume.

grandmother’s advice when making jam: “If you close the pot when the jam is hot rather than when it’s cold, the jam will keep longer.” She was talking about what she knew from experience; she didn’t have to be knowledgeable about the microbiological phenomenon. Most of today’s traders still act in the markets like my grandmother did in the kitchen. They understand little about the trading mechanism, and few really are aware of what their trading tools are calculating or what their limitations are.

Number of Shares

Tellabs: Real Volume Historam by Time, 09/20/2006

FIGURE 1.6b The real volume repartition by time level. The real volume repartition by time level is very different from the linear volume repartition.

30

VALUE IN TIME

Do not trade like my grandmother made jam. You need to understand what is going on in the market that you trade. There are two ways to gain knowledge: 1. Invest time to study how the market works. I advise you to participate

in one of the seminars that Dr. Alexander Elder gives, or even one of his weeklong trading camps. Not only do these courses give you a working structure, but they also help you to feel how the market is moving. You will gain knowledge and confidence. 2. Use modern tools that will tell you what is happening.

EFFECTIVE VOLUME To define the Effective Volume tool, I applied three modifications to Larry Williams’ method. Let’s look at Figure 1.7, where we see the evolution of the price during one trading minute. We can see that the price evolved among five ticks: $10.00, $10.01, $10.02, $10.03, and $10.04. If we suppose that 5,000 shares were traded during that trading minute, Larry Williams’ formula would tell us that the share accumulation is: Accumulation =

$10.03 − $10.01 × 5,000 = 2,500 shares $10.04 − $10.00

1. The first modification is to replace the open of the actual trading

minute with the close of the previous trading minute. This modification looks at the volume that has a real impact on the price from one trading minute to the next. If the price increased, the Effective Volume will be positive. Otherwise it will be negative.

FIGURE 1.7 Larry Williams accumulation/distribution.

31

Effective Volume

2. If the close from the previous minute is lower than the low of the actual

minute, the second modification is to replace the low of the current minute with the close of the previous minute. Similarly, if the close from the previous minute is higher than the high of the actual minute, then the Effective Volume method requires replacing the high of the current minute with the close of the previous minute. In our example, if we suppose that the previous close was at $9.99, we then would need to use that previous close in our calculation instead of the open of the minute (see Figure 1.8). This modification would give us the following number of shares being accumulated: Accumulation =

$10.03 − $9.99 × 5,000 = 4,000 shares $10.04 − $9.99

3. A last small adjustment still needs to be done: When applying the mod-

ified Larry Williams formula on small time intervals, it is necessary to add 0.01 to the top and to the bottom of the formula. The reason for this is that when shares are distributed between, for example, a low of $9.99 and a high of $10.04, it means that the shares traded at $9.99, $10.00, $10.01, $10.02, $10.03, and $10.04—six ticks instead of five (as would have been the case with the Larry Williams formula that simply subtracts $9.99 from $10.04, which would equal $0.05 or only five ticks). Applying the three small modifications to our example, the Effective Volume calculation gives the following results: Accumulation =

$10.03 − $9.99 + $0.01 × 5,000 = 4,167 shares $10.04 − $9.99 + $0.01

FIGURE 1.8 Modified Larry Williams accumulation/distribution.

32

VALUE IN TIME

Effective Volume Formula The Effective Volume is calculated by using the following formula, which is a modified version of the Larry Williams accumulation/distribution (A/D) formula: (Closei−1 − Closei ) + PI × Volumei Highi − Lowi + PI where

Closei−1 = Closing price corresponding to time interval (i − 1): TIi−1 Closei = Closing price corresponding to time interval i: TIi Highi = Max (Highi , Closei−1 ) Lowi = Min (Lowi , Closei−1 ) PI = Price interval (usually US $0.01)

As you can see, the Larry Williams formula was changed in three ways: 1. I replaced the open of the time interval with the close of the previous

time interval. 2. I adapted the high and the low of the current time interval to the value

of the close of the previous time interval. 3. I added the PI number, usually 0.01, to use the exact number of ticks between Closei−1 – Closei and between Highi – Lowi , and not just the mathematical difference between these values. The last column of Table 1.4 shows the calculated values for Effective Volume using the preceding definition. A simpler definition of the Effective Volume, as presented in the Foreword written by Dr. Elder, consists of considering as “selling volume” the volume that pushed the price down from one minute to the next and as “buying volume” the volume that pushed the price up. Both definitions would produce similar results, because the main influential element of both definitions is to consider only the volume that is linked to a price change from one minute to the next, while the price variations within one trading minute carry a relatively small importance. The rightmost column shows the Effective Volume figures calculated for every trading minute. It is their cumulative value that gives the Effective Volume flow shown in Figure 1.9.

What Is the Effective Volume Flow? The Effective Volume flow is the total value of cumulated Effective Volume values from one minute to the next. It is interpreted similarly to any

33

Effective Volume

TABLE 1.4 Effective Volume Example

9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06 9/25/06

14:27 14:26 14:25 14:24 14:23 14:22 14:21 14:20 14:19 14:18 14:17 14:16 14:15 14:14 14:13

Open

High

Low

Close

Volume

Effective Volume

$11.07 $11.06 $11.06 $11.05 $11.04 $11.03 $11.02 $11.02 $11.01 $11.02 $11.04 $11.05 $11.06 $11.07 $11.07

$11.07 $11.06 $11.07 $11.06 $11.06 $11.04 $11.04 $11.02 $11.02 $11.02 $11.04 $11.06 $11.06 $11.08 $11.08

$11.06 $11.06 $11.06 $11.05 $11.03 $11.03 $11.02 $11.02 $11.01 $11.02 $11.01 $11.04 $11.05 $11.06 $11.07

$11.06 $11.06 $11.06 $11.06 $11.05 $11.04 $11.04 $11.02 $11.02 $11.02 $11.02 $11.04 $11.06 $11.06 $11.07

5,889 200 28,335 18,131 33,188 3,298 29,658 17,825 11,351 40,889 14,015 13,802 32,536 16,399 20,041

0 0 0 18,131 16,594 0 29,658 0 0 0 −10,511 −13,802 0 −10,933 0

other indicator that gives a general view of the accumulation or distribution of shares (see Figure 1.9). Its interpretation is straightforward: Effective Volume that is trending up means accumulation (buying); if it is trending down, that means distribution (selling). Effective Volume sometimes precedes price and sometimes follows price. We will study later how to interpret its movements compared to the price movements.

Why Does the Effective Volume Not Use the First Minute of Trading? The Effective Volume method requires ignoring the first minute of a trading day, because the volume exchanged during the first minute of trading relates to overnight news and overnight orders. Traders who place their trades before the opening are usually not professional traders. Large players such as institutions do not react on news, but instead carefully plan their entries and exits. I therefore believe that since the main objective of Effective Volume is to monitor large players, it is better to avoid the first minute of trading.

Why Take the Close of the Previous Minute? It is also important to take the close of the previous minute instead of the open of the actual minute. The reason is that if the open is different, it means that the last transaction that ended the previous trading minute

34

VALUE IN TIME

Tellabs: Total Effective Volume Flow, 09/20/2006 (in ,000 shares)

Tellabs Share Price, 09/20/2006

FIGURE 1.9 The Effective Volume flow and the price pattern. The Effective Volume flow shows the accumulation/distribution trend. It is computed by cumulating the Effective Volume calculated at every trading minute. We can see in this example that the Effective Volume trend is consistent with the price trend during the trading day of September 20, 2006.

exhausted either the bid or the ask, forcing the next transaction to open the next trading minute at a different price. This indicates the movement of the supply/demand balance. Indeed, if we start a new trading minute at a higher price, we know that the previous transaction took out the ask, and that the next transaction is a buying transaction that could not meet a seller at the previous ask, and therefore forced the ask up.

How Can We Monitor the Movements of Large Players? In order to follow the tactical moves of large players, we need to separate the Effective Volume into two groups: the large players and the small players. A separation volume separates these groups. This means that for every trading minute we will examine the size of the Effective Volume that has been exchanged during that time interval. If the size of the exchanged

35

Effective Volume

Effective Volume is higher than the separation volume, then we put this Effective Volume into the group of the Large Effective Volume. Otherwise, we put it into the group of the Small Effective Volume (I explain later in the chapter how to define the separation volume). Obviously, the total Effective Volume is the sum of the Large Effective Volume and the Small Effective Volume. The Large Effective Volume separation is easier to understand with graphs. Figure 1.10 shows the price evolution during the last trading hour for the company Tellabs on September 20, 2006. Figure 1.11 shows the corresponding volume during the last trading hour. The first step is to calculate the Effective Volume. Figure 1.12 shows the total volume from which only the volume that corresponds to price inflections was kept. Figure 1.13 shows the total Effective Volume. You can already visually notice that the Effective Volume represents only about half of the total volume exchanged. The second step was to then separate the Large Effective Volume (Figure 1.14a) from the Small Effective Volume (Figure 1.14b). Now, look closely at these two figures. If you count the number of bars in Figure 1.14a, you will notice that there are only 17 bars. Figure 1.14b, by contrast, shows a total of 26 vertical bars, but these 26 bars are shorter than the bars in Figure 1.14a. The bars in Figure 1.14b represent the volume exchanged by small players, while the bars in Figure 1.14a represent the volume exchanged by large players. Now, if you take a pair of scissors and cut out the vertical bars of Figure 1.14a, adding them end to end, you will create a very high vertical bar.

Tellabs: Price of Last Trading Hour, 09/20/2006

FIGURE 1.10 Tellabs closing price of last hour of trading on September 20, 2006. The chart shows an example of the minute-by-minute closing price bars for the last trading hour.

36

VALUE IN TIME

Tellabs: Total Volume of Last Trading Hour, 09/20/2006

FIGURE 1.11 Tellabs total volume of last hour of trading. The chart shows an example of the minute-by-minute volume bars for the last trading hour for the company Tellabs on September 20, 2006.

The height of this vertical bar represents the sum of all the volume exchanged by large players during the last hour of trading, and that is responsible for a price inflection. If you do the same with the small players, you will create a second vertical bar by adding up all the bars of Figure 1.14b. The height of this second vertical bar represents the sum of all the volume exchanged by small

Tellabs: Total Volume Corresponding to Price Inflections of Last Trading Hour, 09/20/2006

FIGURE 1.12 Total volume corresponding to price inflections. The first step is to consider only those trading minutes for which there was a price change between the previous minute and the actual minute.

Effective Volume

Tellabs: Total Effective Volume of Last Trading Hour, 09/20/2006

FIGURE 1.13 Total Effective Volume. Tellabs: Large Effective Volume of Last Trading Hour, 09/20/2006

FIGURE 1.14a Large Effective Volume. Tellabs: Small Efffective Volume of Last Trading Hour, 09/20/2006

FIGURE 1.14b Small Effective Volume.

37

38

VALUE IN TIME

players. You will then notice that this second vertical bar has the same height as the first one. The separation volume between small and large players during a fixed analysis period (one trading day, for example) is defined as the one-minute volume that divides all the players into two groups (large and small) in such a way that the volume of shares exchanged during that fixed period becomes equal (or as equal as possible) between the two groups. Since the Effective Volume method builds two groups of players with the same number of shares, we can say that each group has the same purchasing or selling power as the other. The difference between the two groups is the determination or the overall trend that they are able to force on the price. Let’s have a look at Figures 1.15a and 1.15b. Which is easier to see: the trend of arrows in Figure 1.15a or in 1.15b? The difference between the two figures is that in Figure 1.15b I have erased all the smaller arrows. Without the noise generated by smaller arrows, the general trend is easier to grasp at a glance.

Playground Analogy To better understand the Effective Volume concept and its analytical power, let’s imagine schoolchildren aged six through 15 playing on a

FIGURE 1.15a Large and small arrows. It is difficult to see the overall trend because of the noise generated by the smaller arrows.

Effective Volume

39

FIGURE 1.15b Large arrows only. The overall trend direction is clearer after the smaller arrows are eliminated.

playground in winter. These kids are instructed to throw snowballs at two Chinese gongs that are placed in the center of the playground. The first gong is called “Long” and the second is called “Short.” The goal of the kids is to make as much noise as possible. The two gongs make distinct sounds, and the stronger the throw, the louder the sound. Now, the goal of the teachers is to find out which gong produces the louder sound. To complicate the game, the school principal has two bags of magic powder that he throws every quarter of an hour on the playground. These two types of powder are called “positive earning surprise” and “negative earning surprise.” When the “positive earning surprise” powder is thrown, the “Long” gong gets bigger and the “Short” gong gets smaller. The opposite happens when the “negative earning surprise” powder is thrown. Obviously, the children aiming at the bigger gong will hit it more often. Therefore, it pays to study the size modification of the gong, and to study how many balls are hitting the gong. This is called the standard technical analysis. Some professors prefer to study what the principal is doing, to see if he takes from one bag more often than the other, to see if he brought more of this or that powder, and so on. This is called fundamental analysis. Obviously, if no child throws a snowball, no sound will be heard. Therefore, it is fair to say that the sounds originate from the children’s activity, just as price moves are a consequence of buying and selling activity. On the playground, a six-year-old makes fewer snowballs, strikes with less power, and misses more often than a 15-year-old. It pays to look at the older kids’

40

VALUE IN TIME

FIGURE 1.16 How volume is distributed.

behavior to guess the strength of the coming sound. Because we cannot watch all the children (there could be 2,000 or 20,000), what we end up doing is looking at all of the coming snowballs and listening to the collection of small and loud sounds made by each snowball. What we need is recording equipment that can both separate the coming balls by strength (size and velocity) and analyze only the sounds of the stronger snowballs. My studies have shown that on a regular basis, half of the volume exchanged in one day has no impact on price movements (price inflections). These are the snowballs that miss the gong. The other half, which does have an impact on price changes, is called the Effective Volume. The Effective Volume represents the balls that hit the gong. My studies have also revealed that if you separate the Effective Volume further in two groups of identical size in terms of total number of shares, the group with larger volume will be responsible for most of the price movements (see Figure 1.16). As a rule of thumb, I can say that 25 percent of the volume involved in stock trading is responsible for 75 percent of the price movements. I call this the Large Effective Volume, which is roughly 25 percent of the total volume. Knowing whether the players responsible for this 25 percent are buying or selling is critical for successful trading.

PRACTICAL EXAMPLES OF EFFECTIVE VOLUME CALCULATIONS The Effective Volume method can be used in three instances: 1. A flat or sideways trading range will probably break in the direction of

the Large Effective Volume flow (a trading range is formed when the price stays for some time at about the same value).

Effective Volume

41

2. A price uptrend with a negative Large Effective Volume flow indicates

that problems may lie ahead. 3. A price downtrend with a positive Large Effective Volume flow indi-

cates that the downtrend might not be sustainable.

Follow the Large Players Accumulating in a Trading Range The most effective way of using the Effective Volume tool is to look for accumulation by large players during a trading range. When I see such an accumulation taking place during a few consecutive days, I purchase shares. I then place a stop below that trading range and wait for the stock price to rise. If large players stop buying before the price rises, I review the situation. Unfortunately, large players are not always right. They do not always act on privileged information or after running sophisticated analysis. However, if you have to buy a stock, you are certainly better off buying a stock that is experiencing strong accumulation by large players, especially if your traditional tools indicate that the time is right to buy. Here is a good example with the company Federated Investors Inc. (FII). The weekly traditional analysis chart shows in Figure 1.17 that at point A the stock price is back to a one-year-old support level. Since it is hitting the line of support, the probability for a reversal is high. The Relative Strength Index (RSI) shows an oversold signal, indicating a possible cheap value compared to past prices. The daily graph (Figure 1.18) shows that at point A we are in a trading range, but it is difficult to evaluate the best timing to purchase the stock: Is it better to buy at point A or at point B? Please note that during a trading range, both RSI and moving average convergence/divergence (MACD) are of little use. The Effective Volume analysis clearly shows that during the trading range, one or more large players have been heavily accumulating (see Figure 1.19). You can see that the difference between the buying and the selling pressure by large players was greater than 1,000,000 shares during the last 20 trading days leading to point B, or 50,000 shares per day. Knowing that 800,000 shares on average are exchanged every day, the imbalance between buyers and sellers was 50,000/800,000 = 6.25%, which is by experience very important. A good question would be: How do we know that we need to buy at point B instead of buying at point A? The answer is: We do not know! There is no way to know when the large players will be satisfied with the accumulated shares, and when (if

42

VALUE IN TIME

FIGURE 1.17 Federated Investors weekly price graph. Source: Chart courtesy of StockCharts.com.

ever) the price will move up (we may suppose that the price will move in the direction of the Effective Volume accumulation, which is often the case). My own rule of thumb is to buy during a trading range when the accumulation by large players is constantly above 5 percent of the daily volume for a minimum of three consecutive trading days. Figure 1.20 shows an increase in price from point B, which was not triggered by any news. We may speculate that the large players decided that they had bought enough shares and that it was now time to push the price up, attracting new buyers. (Please note the dating convention used in the graphs: 07/10/06 means July 10, 2006. The same convention is used throughout the book.) Another question that you may ask is “When do I need to sell?” This is a good question, but Effective Volume alone will not give a satisfactory answer. Indeed, finding a good selling point is much more difficult than finding a good entry. You could sell for a few different reasons:

r Your target has been reached. r Large players have stopped buying.

Effective Volume

43

FIGURE 1.18 Federated Investors daily price graph. Source: Chart courtesy of StockCharts.com.

r The price has not moved for some time. r There has been bad news. r The price has become expensive compared to the underlying value of the equity. We will analyze the selling decision process in Chapters 5 and 6.

Follow the Insiders Insiders have many reasons to sell, but only one reason to buy: to make a profit. We are going to see how to try to catch insiders’ moves. I do not consider company officers here, but rather indirect insiders, the ones who by chance got access to restricted information (although they are not restricted from buying and selling shares). What characterizes the difference between an insider and a large player is the time span they use to buy or sell. A fund will need quite a

44

VALUE IN TIME

FII: Effective Volume by Size 20 days (in ,000 shares) 1,200

FII Share Price, 20 days

FIGURE 1.19 Federated Investors Effective Volume analysis.

long time to accumulate or distribute shares, while a typical insider may be satisfied with only a one-time purchase of a few thousand shares. Also, an insider plays like an option investor: The option will expire at some time. Similarly, the insider’s advantage will expire at the publication of the news. The closer we are to the release date of the news, the smaller the insider’s advantage is, because of the higher probability that other insiders will get the information and will move ahead of him. There are several types of news:

r News that is linked to the day-to-day business: the discovery of oil for a petroleum company, a new patent for a high-tech company, an approval from the Food and Drug Administration for a pharmaceutical company, a very large contract, and the like.

Effective Volume

45

FIGURE 1.20 Federated Investors Effective Volume analysis.

r Exceptional news: a Securities and Exchange Commission (SEC) inquiry for any listed company, the purchase of another company, and so forth. r The regular news: the good/bad quarterly earnings releases. Day-to-Day Business Insider Let’s have a look at an example of an insider move for PetroQuest Energy, Inc., a natural gas exploration company. Looking at Figure 1.21, we can see two uptrends in Large Effective Volume: uptrend A and uptrend B. Uptrend A is normal; large players buy the stock, pushing the share price up. However, uptrend B is more difficult to explain, since the price is decreasing during the same period. Also, uptrend B in Large Effective Volume is stronger than uptrend A; during uptrend A, there was a net difference of 120,000 shares being bought. This difference

46

VALUE IN TIME

PetroQuest: Effective Volume by Size 10 days (in ,000 shares)

PetroQuest Share Price, 10 days

FIGURE 1.21 PetroQuest: Effective Volume analysis.

pushed the price up from $9.3 to $9.9, or 6.5 percent. How can we explain that during uptrend B the price declined by 3 percent while the net buying by large players was about 170,000 shares? The price should have proportionally increased by another 9 percent. The B arrow of Figure 1.21 shows that some lucky investor bought 170,000 shares at an average price of $9.9, to see the price increase to over $12 in a matter of two days, as shown in Figure 1.22. This is a no-risk profit of more than $357,000. This is not a hedge fund or an institutional investor—just a standard information leak. The PetroQuest 8-K filing on January 27, 2006, stated: On January 27, 2006, PetroQuest Energy, Inc. (the ‘Company’) issued a press release announcing production and estimated proven reserves results for the year ended December 31, 2005. In

47

Effective Volume

PetroQuest Share Price, 10 days

FIGURE 1.22 PetroQuest: price spike.

addition, the Company provided 2006 production guidance, an update of hedge transactions, and an overview of recent acquisition and drilling activities. The main reason I developed the Effective Volume analysis was that I was fed up with being the last to know when some news was coming into the market. When I returned home after a good working day, it often was too late to profit from good news or to avoid losses if the news had been bad. I lost money more often than necessary on bad news that was already known by a few. I also sold too quickly before good news hit the market. Today, I still miss some big moves, but with the Effective Volume tool, I gain more of an insider view. Most important, I now have some time to act before the news hits the market rather than just react to the news. Earnings Leaks Let’s have a look at an earnings-related insider move for Ariba, Inc. (ARBA). Figure 1.23 shows normal behavior by large players who are pushing the price up during the A uptrend. By contrast, we can admire the share accumulation that took place during the B price downtrend, just before the earnings release that triggered the price jump shown in Figure 1.24. Of course I cannot say with 100 percent certainty that this is an example of insider trading, but those 100,000 shares that were purchased between $7.5 and $8 (between January 19 and January 23) saw their value jump to $9.5 overnight—a gain of about $175,000. Another lucky investor! A Mistaken Signal Signals before earnings can be dangerous, however, especially signals on the downside. Let’s have a look at the company

48

FIGURE 1.23 Ariba: Effective Volume analysis.

FIGURE 1.24 Ariba: price spike.

VALUE IN TIME

Effective Volume

49

Cognizant Technology Solutions. Cognizant is an interesting case study. I lost money on it by following the large volumes. You can see from Figure 1.25 that large players had been selling before May 2, 2006, while the price during the last trading day seemed to hold fast. At that time, I was looking for a short play, and thought I had found one. (A short play means betting that the price of the stock will go down, by borrowing shares and selling them on the market. The profit is made by repurchasing the stock later on from the market at a lower price and pocketing the difference.) I placed my short order on May 2, just before the market closed. I have rarely lost money so quickly for not doing my homework (see Figure 1.26). The reason for the misinterpretation was that we were very close to the day on which earnings would be announced. At such a time, some funds would prefer to be out of the stock instead of taking a possible hit because of a bad earnings report. Very few large funds would increase their risk just

FIGURE 1.25 Cognizant: Effective Volume analysis.

50

VALUE IN TIME

FIGURE 1.26 Cognizant: price spike.

one day before earnings are announced. In this case, the decrease of Large Effective Volume just prior to the earnings release had risk reduction as its motive, and I should not have interpreted it as selling due to superior information. Had we seen large players increasing their positions just before the earnings release, it would have been a positive sign for the stock. Why? Because funds do not increase risk without reason, especially just before a major earnings release; therefore, we could have concluded that the funds had superior information indicating positive earnings. Since that time, I have avoided placing shorts ahead of an earnings release. Standard Technical Tools This box explains a few standard technical tools, which are sometimes referred to in the later chapters of this book. The experienced trader may just want to skip to the next section. All these are very well explained in Trading for a Living and Come into My Trading Room, both by Dr. Alexander Elder. If you do not know these indicators, I strongly advise you to study them. I will give some brief comments about these indicators. Moving Averages: A moving average is a measure of the average price that people have been ready to pay for a stock during a period of time. It is the consensus value of the stock over a short-term or longer-term period (depending on whether you are averaging on 20, 50, or 200 days). The price may stay above or below the average for a considerable time, drawing an uptrend or a downtrend. If you are in an uptrend, it pays to buy when the price is back down to its value line. If you are in a downtrend, it pays

Effective Volume

51

to sell when the price is back up to its value line. This means that trading against the price trend is very difficult, because you have to fight against the consensus opinion of the herd. Unless you are superhuman, it is frankly not a good idea to try to change the direction of a moving boat all by yourself. A moving average will give you an indication of value compared to past prices within a trend. (See Come into My Trading Room, by Dr. Alexander Elder.) Relative Strength Index (RSI): J. Welles Wilder developed this momentum oscillator. This indicator compares the recent price gains to recent price losses and converts the result into a number between 0 and 100. A number below 30 indicates that the stock is oversold. Typical long downtrends can keep an RSI signal below 30 for many days. The buy signal comes when the RSI moves back up over the 30 line, indicating that the stock is probably changing its momentum to a new buying trend. Another buy signal is generated when the stock reaches a new bottom; a stronger RSI, however, indicates that the new push-down in price did not increase the average loss, and the RSI indicates a higher number for the gain/loss comparison. My experience shows that the RSI is widely followed. I suspect that since it is easy to program, many automatic trading methods use this signal to enter or exit stocks. At the start of a new trend, the RSI will give you an indication of value compared to past prices. (See Trading for a Living, by Dr. Alexander Elder.) Moving Average Convergence/Divergence Histogram (MACDH): This momentum oscillator was developed by Gerald Appel. It compares a fast and a slow moving average in order to detect whether the price change is quicker or slower than before. It compares the acceleration (rate of change) of the fast and the slow moving averages. If the acceleration of the fast moving average is higher than the acceleration of the slow moving average, this indicates a positive momentum in the price. (See Technical Analysis: Power Tools for Active Investors, by Gerald Appel.) Support/Resistance Lines: These lines are important. They indicate price congestions, or the price levels where many buy/sell decisions are taken. (See Trading for a Living, by Dr. Alexander Elder.)

TECHNICAL SECTION: HOW TO CALCULATE THE SEPARATION VOLUME This section is for readers who want to know more about some of the technical details of the Effective Volume method.

52

VALUE IN TIME

Fixed Separation Method There are many ways to separate Large from Small Effective Volume. You could, for example, pick a number and categorize every volume that is above this number as large volume. This separation method does not work well, though. Indeed, since volatility is one of the main characteristics of volume, if you fix the separation volume at a specific number, on some trading days you will end up having mainly Large Effective Volume, and on other trading days, mainly Small Effective Volume. The reason for this is found in the behavior of large players. The buying by large players is usually executed during trading ranges, when nobody else is really paying attention. During that time, it is difficult to buy more shares than the supply side can support without raising the price. This means that large players will adapt their buying and selling order size to what the market can give or take from them. This situation changes every day. Therefore, on some days, a large player who wants to accumulate will buy more shares than on some other days. The consequence of this fact is that we need to recalculate our separation volume every day.

Average Separation Method The most obvious way to separate Large from Small Effective Volume is to calculate the per-minute average Effective Volume exchanged for all the minutes of the day where a price inflection was found. The volume above that average is called Large Effective Volume and the volume below it is called Small Effective Volume. Such a separation is represented in Figure 1.27; I have used the Effective Volume of one trading day only (September 20, 2006) for the company Tellabs. For Figure 1.27, the separation between the Large and Small Effective Volume is calculated as 25,951 shares, which is the average Effective Volume. As a reference, in Figure 1.27 there are 260 vertical bars. (Each bar represents a trading minute with valid Effective Volume, regardless of whether negative or positive. Please note that since I am interested only in the size of each bar, I turned positive all the negative Effective Volume in Figure 1.27.) The total number of Effective Volume shares was about 6,750,000 shares, compared to the total number of exchanged shares—about 11,000,000. We can see that the total Effective Volume represents only about 50 percent of the total number of shares traded. If we plot the Large and Small Effective Volume flow using the average Effective Volume as separation volume, then we can see that the price pattern is following the large players’ pattern (see Figure 1.28). In general, we may say that large players are the ones moving the price.

53

Number of Shares

Effective Volume

FIGURE 1.27 Tellabs one-day Effective Volume. On a sequential view of the vertical bars that correspond to the Effective Volume, it is difficult to see if the average Effective Volume is a good separation between the Large and the Small Effective Volume.

Figure 1.28 represents the Large and Small Effective Volume flow for September 20, 2006, using the average Effective Volume as separation volume. Let’s analyze this separation volume in more detail. If we come back to Figure 1.27 and sort the 260 vertical bars from the highest to the lowest, we obtain the very interesting results shown in Figure 1.29. What stands out immediately is that the total surface covered by the Large Effective Volume bars looks much bigger than the total surface covered by the Small Effective Volume bars. An analogy will help us to understand the importance of that discovery. Suppose that the Department of Transportation wants to assess whether the new speed limit regulation is well respected on a particular road. One junior engineer places a very precise radar system on the side of the road. After one week of measurements, he notices that the radar registers an average speed of 14 mph, compared to the 40 mph limit. He concludes that the speed limit is well respected. However, looking at the data, we notice that among the vehicles that passed on that road, there were 90 bicycles and 10 cars. The bicycles’ average speed was 10 mph, but the cars’ average speed was 50 mph. Now we see that the problem lies within the data: The cars’ speed is much higher than the bicycles’ speed, and there are many more bicycles than cars. It is quite similar to the stock market: We need to cope with volume volatility. The Internet is bringing more retail players to the market (more bicycles), and a growing number of funds are taking sizable positions (faster cars). In real life, it would make sense to put the cars on the highway and the bicycles on a cycling path. In the stock market, however,

54

VALUE IN TIME

FIGURE 1.28 One-day Effective Volume separated by average size. The separation of Effective Volume into Large and Small Volume, when done using the average volume by vertical bars, produces a pattern where the Large Effective Volume is closely following the price pattern.

bicycles and cars compete for the same shares. This is where volatility was born. If you add to this volatility the fact that the recent decimalization killed market visibility, it is no wonder that the general public now believes that markets are manipulated. Later in the book I will provide more details on volatility (Chapter 3) and possible market manipulation (Chapter 4).

Equi-Power Separation Method Remember that I define the Effective Volume as the volume responsible for a price change from one minute to the next. This means that it makes no difference whether the price increased by $0.01 or $0.02. In both cases, I consider the Effective Volume as buying volume. Therefore, every share that enters into the definition of Effective Volume has the same ability to

Effective Volume

55

FIGURE 1.29 Repartition of Effective Volume by size. When rearranging all the bars of Figure 1.27 by size, we notice that the Effective Volume separation between the Large and the Small Effective Volume using the average separation method leaves too many shares within the pool of the Large Effective Volume. In this example, 70 percent of the shares belong to the large players group.

move the price up or down as any other share. If we separate all the Effective Volume shares into two groups that include the same number of shares, we have in theory two groups with the same intrinsic power to move the price. In our example in Figure 1.29, we know that the total number of Effective Volume shares was 6,750,000 shares. Let’s count the shares from the left of Figure 1.29 to the right until we reach 50 percent of 6,750,000 shares (i.e., 3,375,000). The calculation shows that we reach the 50 percent midpoint with a separation number of 42,500 shares. This means that all the Effective Volume sizes that are larger than 42,500 shares are labeled as large, and the rest are labeled as small. Figure 1.30 shows this new separation. Please also notice in Figure 1.30 that I have labeled as Large Volume the volume corresponding to only 48 one-minute time intervals. This is out of the 260 one-minute time intervals that constitute the total number of time intervals for which we had price inflections. In other words, these 48 one-minute time intervals theoretically have the same power to move the market as the remaining 212 one-minute time intervals, because these 48 minutes include the same number of shares as the remaining 212 minutes.

56

VALUE IN TIME

FIGURE 1.30 Equi-Power Effective Volume separation method. The Equi-Power Effective Volume separation method labels an equivalent number of shares as Large and Small Effective Volume. It allows for a more balanced separation where, in theory, each group of shares has the same power to move the price.

However, the difference between the two groups is that small players are scattered, and therefore they deliver their power in a very diffuse way; it will have only a limited influence on the price. In contrast, the volume corresponding to the large players will carry the colluding will of a few large holders, which will have much more influence on the price pattern. Large holders will have a general tendency to deliver their purchasing or selling power in a dedicated way that will determine the price direction. If we now come back to Figure 1.28 and adapt it using our new definition of the separation volume, we can see in Figure 1.31 quite a different pattern of behavior. In Figure 1.28, which uses an average separation method, we see that large players have more influence than small players. This is normal, since the large players group includes 70 percent of the Effective Volume shares. However, in Figure 1.31, which uses the Equi-Power separation method (each group has 50 percent of the number of Effective Volume shares), we notice a much more balanced influence between large and small players. The interpretation of the graph issued using the Equi-Power separation method is as follows: If both large and small players show a well-balanced pattern, it means that no institution was active during the period of analysis. (The large players group is therefore the group of large retail players,

Effective Volume

57

FIGURE 1.31 One-day Effective Volume separated by midpoint size. Compared to Figure 1.28, Figure 1.31 shows that the separation of Effective Volume into Large and Small Effective Volume using the Equi-Power method produces a more balanced pattern for the repartition of influence of each group on the price evolution.

while the small players group is the group of small retail players.) However, if the large players’ pattern is very different from the small players’ pattern, this shows how institutions have been moving.

IMPROVE YOUR TRADING: DECIDE ON THE BIG PICTURE The stock market is getting quite complex, with many different players using a large variety of analytical tools and trading instruments. As a trader (retail or professional), you need to know what other traders are doing and when they are doing it.

58

VALUE IN TIME

The great majority of traders are momentum players or trend followers. This means that they are moving in and out of the market by following others’ decisions. We may call it herd behavior. Herd analysis is a concept that is easy to understand and easy to model through price-based technical indicators (RSI, MACDH, trend indicators, etc.). However, indicators that are based only on price usually give you information after the fact. Tools that are based on the price/volume relationship tend to catch the buy/sell decisions somewhat earlier, when these decisions are being spread to the herd. These tools are more powerful than tools based only on price, because they combine unrelated data (volume and price are believed to be unrelated) to strengthen the analysis. In the following section, I briefly study the tools that use a price/volume spread analysis, as well as the tick volume analysis tools. I find these tools useful, but fuzzy in the sense that you can’t be sure whether the indicator tells what really happens in the market. These indicators may be generally correct; otherwise, why would they be used extensively? The objective of the new tools that I present in this book (see Figure 1.32) is to study the decisions of other traders before these decisions are spread to the herd level. To move a herd of traders, two things must happen: 1. The herd needs to be ready to be moved. It is impossible to move a

herd that is not in a position to move. You therefore need to study the position of the herd. This is the purpose of the Active Boundaries indicator presented in the next chapter. 2. You need trendsetters. These are a few key people who provoke change. You cannot see their move easily, except if you go down to the tactical level. This is what the Effective Volume method and its associated Effective Ratio and divergence analysis methods are for.

FIGURE 1.32 The evolution of the technical tools.

Effective Volume

59

A COMPARISON WITH TRADITIONAL TOOLS This section is very theoretical and reviews other well-known volumebased tools. You can skip it without compromising your understanding of the rest of the book.

Price-Based Indicators When I started trading, I tried all the technical indicators that were available on the market, with different time frames and settings. I discovered that it was quite easy to find an indicator that would justify any possible trading decision. I also quickly found indicators that worked well for me: moving averages, RSI, MACDH, and support/resistance lines. These have been mentioned previously in this chapter.

Price/Volume-Based Indicators The importance of a price/volume combination analysis has been understood for many years, so I will cite here only the work of a great pioneer of that analysis. Richard Wyckoff worked extensively on the price/volume relationship more than 80 years ago. A trader from the 1920s, Wyckoff wrote several books on the market, and eventually set up the Stock Market Institute in Phoenix, Arizona. At its core, Wyckoff’s work is based on the analysis of trading ranges, determining when stocks are in basing, markdown, distribution, or markup phases. Incorporated into these phases are the ongoing shifts between “weak hands” (public ownership) and “composite operators,” now commonly known as smart money. There are several ways to use the combination of volume and price end-of-day data:

r Weight the volume to the price spread for the day. r Weight the volume to the price change from the previous day. r Compare the day price/volume relationship to the previous day’s price/volume relationship.

r Use volume to weight other price-based indicators such as the RSI or the MACDH. These different variations follow a similar purpose: to determine the behavior of traders at the decisional level. They are used to assess whether more traders decided to buy or to sell that day, and to give an indication for the next trading day. These tools help you to figure how the demand/supply balance is working out at the global level. The idea is that all the actions and all the opinions of traders are in the volume, while the price will indicate the direction of the movement.

60

VALUE IN TIME

What You Really Need to Know I often read in the literature that this or that indicator measures the supply/demand equilibrium or the buy/sell equilibrium. Some will even use a still less precise term called bear/bull equilibrium. These terms are not helpful, because they are difficult to define and therefore to measure. Let’s look at a few general statements:

r When bulls are present or are getting stronger, the price should increase. When bears are getting stronger, the price should decrease. When bulls are getting weaker, the price should decrease. When bears are getting weaker, the price should increase. When the price trends up and if you recognize that the supply/demand equilibrium is moving in the direction of the bears, it means that the smart money is moving out. r The key point is to find where the herd is moving by studying the strength of the buyers when the price reaches a top or the weakness of sellers when the price reaches a bottom.

r r r r

I think that most of these statements are meaningless and thus useless. You can find them all over many trading books because they make things easy to understand, but the underlying concepts are so imprecise that the only way to explain them is by using a general market psychology type of explanation or supply/demand equilibrium. Measuring the supply/demand equilibrium is an objective that is publicized by the majority of technical analysis books and tools. They would typically say that this or that divergence between the signal and the price indicates that bears or bulls are getting weak or strong. The great majority of these authors transmit their trading experience in terms of patterns (price, price/volume, etc.). They will explain the pattern in terms of equilibrium or in terms of force or weakness. Very few come out with a mathematical formula to model this or that pattern or to catch the buyer/seller equilibrium. The reason is that formulas are imperfect; on many occasions they do not work well. Even general patterns are not sure bets (the head-and-shoulders pattern, the cup-and-handle pattern, etc.). Everybody agrees that markets are not perfect; this is the reason you need a risk-management policy such as stop loss. The problem of unpredictability is not in the markets, but it is in how we interpret and measure the markets. As a trader, you need to better understand how the market works and have tools to better measure market movements. Then your trading will progress. You will gain confidence in

Effective Volume

61

what you are doing and will place stops only when you are physically away from the markets. Do not trade like a grandmother, who follows only the well-known patterns. Grandmothers are, of course, usually right, but you will greatly increase your confidence after you understand and can measure the forces that are behind the formation of patterns. Let’s revisit some more specific tools. Force Index Force Index, which was invented by Dr. Alexander Elder, is best defined by taking Dr. Elder’s own words in his book Come into My Trading Room: Force Index helps identify turning points in any market by tying together three essential pieces of information—the direction of price movement, its extent, and volume. Price represents the consensus of value among market participants. Volume reflects their level of commitment, financial as well as emotional. Price reflects what people think, and volume what they feel. Force Index links mass opinion with mass emotion by asking three questions: Is the price going up or down? How big is the change? How much volume did it take to move the price? It is very useful to measure the force of a move because strong moves are more likely to continue than weak ones. Divergences between peaks and bottoms of prices of Force Index help nail important turning points. Spikes of Force Index identify zones of mass hysteria, where trends become exhausted. Here is the Force Index formula: Force Index = (Close today − Close yesterday) × Volume today Then Dr. Elder continues with an explanation of how to use Force Index with price divergence: Trend reversals do not have to come as a surprise; divergences between Force Index and price usually precede them. If the market is trying to rally, but the peaks in Force Index are becoming lower, it is a sign of weakness among the bulls. If a stock or a future is trying to decline, but the bottoms in Force Index are becoming more shallow, it is a sign of weakness among the bears. As I do not take anything for granted, I was wondering how correctly the multiplication of two different variables (price and volume) could lead to a meaningful result. Usually in physics, you would compare variables by

62

VALUE IN TIME

dividing them, in order to see the influence of a change in the first variable on a change in the second variable. For example, if you drive 100 miles for two hours, your speed would be the distance divided by the time, or 50 miles per hour. Then why can we not adjust the Force Index formula and call it the Weakness Index? (We know that the objective of both the Force Index and the Weakness Index is to measure the bulls/bears equilibrium.) Weakness Index =

Volume today Close today − Close yesterday

Let’s take an example. If on day one, the price change is +10 cents on a volume of 100,000 shares, which is twice the average volume, but on the following day, the price change is only +5 cents on a volume of 400,000 shares, what could we conclude? The Force Index formula would conclude that the market is going to move higher, because the Force Index is increasing. In fact, the strength is doubling: more buyers came in, and as a consequence the price continued to move up. Day one Force Index = +10 cents × 100,000 = 1,000,000 Day two Force Index = +5 cents × 400,000 = 2,000,000 The Weakness Index formula would conclude that the market is going to move lower, because the Weakness Index is increasing. In fact, it says that to move the price up by one cent on day two, we needed eight times more shares than on day one, indicating an increasing weakness. Between day one and day two, it looks like the supply of shares is increasing and that the market will soon reverse down. (This calculation is commonly known as the measure of price elasticity to volume.) Day one Weakness Index = 100,000/10 cents = 10,000 shares/cent Day two Weakness Index = 400,000/5 cents = 80,000 shares/cent Both analyses are correct, because neither the Force Index nor the Weakness Index indicates if the price moved on the strength of demand or on the weakness of supply, which are very different. However, we saw before in this chapter that the buying/selling strengths are stronger than the strength of the demand/supply, because of the will difference. Therefore, I am inclined to say that the Force Index is a better indicator than the Weakness Index.

Effective Volume

63

Another difficulty in interpreting the Force Index lies in the mere multiplication of price by volume. Indeed, because of its name, we may assume that the Force Index measures the buyers’ strength in an increasing price trend and the sellers’ strength in a decreasing price trend. However, it also includes the influence of the sellers’ strength in an increasing price trend and the buyers’ strength in a decreasing price trend. The problem is that since it multiplies price by volume, we have no certainty as to which element of the buy/sell balance is stronger. We may, however, suspect that it is the price direction that dictates the side to which the buy/sell balance will tilt. Volume Weighted by the Price Spread Other methods try to measure the buy/sell balance during the day by weighting the volume with the balance between (Close price − Open price) and (High price − Open price). The idea is that this balance in prices is a good representation of the balance between buyers and sellers (volume balance). Unfortunately, this is a very incorrect assumption because of the large intraday volume volatility—we all know how strong the volume is at the beginning and at the end of the day. Let’s have a look at professional players at the close of the trading day. Since many technical tools follow the closing price, professional players will put a lot of energy (volume) into a close that will favor their positions. End-of-day indicators are therefore easily manipulated by large players, and may be giving an incorrect view on the buyers/sellers equilibrium. Let’s take one example that will show you the limitations of this type of indicator. Suppose that a share price moves down 50 cents on 400,000 shares that were exchanged during the trading day up until 30 minutes before the close. During the last 30 minutes of trading, however, a large fund appears and pushes the price up 70 cents on only 100,000 shares. The push is sudden and the supply of shares dries up, which results in a sudden price increase. It happens so quickly that sellers do not have time to offer new shares for sale, and as a result, the share price ends the day strongly at +20 cents. All the end-of-day methods will consider that the price moved up 20 cents on volume of 500,000 shares, while in fact the price was mostly down during the day. A strong close would indeed normally attract buyers on the next trading day, which is probably a requirement for the large fund to sell its shares at a good price. In this case, we see that a tactical move by a large fund, because it was taken at a key period of time (at the closing of the day), may induce other traders to make strategic decisions that will not be to their advantage. You now understand why there is a need for a tool that allows you to see tactical moves as well as the strategic ones.

64

VALUE IN TIME

I believe that these methods that combine price and volume work fine when used with end-of-day data, but that they lose efficacy when you shorten the analysis period to 1 hour, 30 minutes, 15 minutes, 5 minutes, and finally 1 minute. These methods were indeed developed before decimalization. Before decimalization, if you looked at intervals of one minute, you would hardly see any significant change in price. You would have to look at longer time intervals such as 10 or 15 minutes to try to analyze the volume/price relationship. Typically, you would use on the 10-minute analysis period one of the methods used for the end-of-day analysis, such as the on-balance volume method or the volume spread analysis method. I believe that a 10-minute analysis period is about the shortest time frame for which you can use these methods. On very small time intervals of one minute, these methods do not work, because they have been designed to model accumulation/distribution and supply/demand at the decisional level of all the players, including large funds. The one-minute time range is more suited for analyzing tactical moves than decisional moves. When a large fund is accumulating, this accumulation takes place during many days. Therefore, a repeating pattern of accumulation using these traditional methods on end-of-day data has a good chance of catching the strategic buying decisions of a large fund. This is especially true knowing that accumulation is better executed by active buying (placing buy orders at the market price) than by passive buying (placing limit buy orders, which wait for sellers to come in). Indeed, large limit orders are more visible to the market than market buy orders, and will invariably push other buyers to buy higher and sellers to retrieve their orders, waiting for better prices. As a position trader, you do not need to know more to make a successful trading decision. However, if you go down very close to the transactional level, you will quickly notice that these traditional methods no longer work well. The main reason is that close to the transactions, you are facing tactical decisions and not strategic decisions. Tools that are used to model strategic decisions perform very poorly for tactical decisions. A typical tactical decision would be, for example, to send small or midsize buy orders on a regular basis, as long as the price stays within a given range, and then send a quick sell order to attract more sellers and bring the price back into the required range. You would continue this tactic until the supply of shares dries up. Traditional tools are very poor at capturing such tactical behaviors, because they encounter two important mathematical issues: 1. On very small time intervals of one minute, the price spread is also very

small. It is so tiny that interpretation becomes extremely hazardous.

Effective Volume

65

Any type of volume can easily move the price up or down by one tick. At the level of the transaction, the market looks very erratic. Volume spikes would come on the bid or on the ask at what looks like stochastic, unpredictable moments. 2. At the level of the minute bar, price and volume exhibit a very different behavior: For example, you can imagine on the first bar that 5,000 shares have been traded, pushing the price up two cents. Then, on the following bar, imagine that 100,000 shares pushed the price down one cent, and then one minute bar later, that 200 shares pushed the price up again two cents. The problem that must be tackled is the incredible difference in volatility between price and volume. Price is fundamentally nonvolatile at the one-minute bar level, while volume is extremely volatile. Therefore, multiplying price by volume at the end-of-day level could still have some meaning, whereas multiplying price by volume at the level of the trading minute is mathematical nonsense. However, even if you can solve these issues, there is still the danger of focusing on the tactical movements instead of trying to grab the strategic moves. What you then need to do is view these tactical moves in a daily or weekly flow and then compare the flow to past flows to understand the strategic moves. It is like assembling a big puzzle with pieces coming from different bags. You first need to sort the pieces before assembling them to get the complete image. Tick Volume Analysis It is now clear that going down to the transactional level gives us a better chance of catching tactical moves by large players. To use transactional data, a first tool was developed by Don Worden but later on extensively used and publicized by Laszlo Biriny. This tool, which Biriny called the “Money Flow Index,” compares the amount of trading on small price upticks with the amount on small downticks, with the hypothesis that large-volume single trades are more important than small-volume trades. This was true in the past, but less so today. According to research from the consulting firm Aite Group, at the end of 2006 the share of automated computer trading following predefined algorithms approached one-third of total U.S. equities trading volume; this number is expected to rise to 53 percent by the end of 2010. These trades use algorithms that slice orders to intentionally hide the order size, drip releasing orders to the exchange, and as a consequence increase order fragmentation. Therefore, the hypothesis on which this Money Flow Index tool is based is quickly challenged by market reality.

66

VALUE IN TIME

A second set of tools counts orders executed at the ask as buying orders and orders executed at the bid as selling orders. The balance indicates the equilibrium between buyers and sellers. These tools are used in the same way that the end-of-day accumulation/distribution methods are used: They give a broad understanding of the market direction (decisional level). When working with transaction data, however, the problem is that transactions occur randomly. Therefore, the only way to start measuring and comparing them is to link them to a specific time interval: You need to add transactions into cumulative minute data. You then arrive at the minute analysis and the Effective Volume method.

WHAT WE LEARNED REGARDING EFFECTIVE VOLUME We started this chapter with a simple question: What are large players doing? We saw that the monitoring of the volume involved in small price changes from one trading minute to the next, which I defined as the Effective Volume, is a very good tool for detecting tactical moves by insiders, institutional investors, and other large players. Effective Volume often allows us to detect future price changes. The Effective Volume tool is excellent for detecting trendsetters, but we will see in the next chapter that in order to really monitor trends, the Effective Volume tool must be used in conjunction with the Active Boundaries tool.

CHAPTER 2

Price and Value The Active Boundaries Indicator

his chapter is about trends: why trends exist, how they are created, and how we can monitor them. Before studying the general theme of trends, it is first necessary to examine a central question: Why are shares bought or sold?

T

BUY LOW Every investor knows that it is easier to make money when you buy at the bottom than when you buy at or near the peak. As a consequence, in order for you to know if a stock is cheap, you need to determine what the value of that stock is. There are as many ways to define the value of a stock as there are investment strategies:

r One very common way for selecting stocks with good value is to choose stocks that show a low price-earnings (P/E) ratio. Since investors look at the future, you may elect to look at a forecast of next year’s P/E ratio. Typically, some industries maintain a P/E ratio of 10, while high-growth industries command a higher P/E ratio. r A second way to find stocks with good value is to read the future earnings estimates issued by analysts. r Still another way is to study the earnings surprise. This means that you compare the real earnings published by the company to the analysts’ 67

68

VALUE IN TIME

FIGURE 2.1 Traditional methods to detect value: RSI, trend line, moving average, and support line. Source: Chart courtesy of StockCharts.com.

average estimate. In case of positive surprise, you may decide to buy the stock, since you could also expect positive surprises in following quarters. Before I came up with my own tools, I used to follow four types of technical signals that indicate when a stock is becoming cheap (see Figure 2.1): 1. Support line. The support level is indicated by a horizontal line that

connects price bottoms. It is said that if the price goes below its support, it has the potential to go much lower, but if support holds, price will bounce up. This is why buying at support is a strategy followed by many traders, who place a stop-loss limit order just below the support line. 2. Trend line. A price trend line is generally a line that is drawn by connecting tops or bottoms on charts. The trend line is the traders’ consensus value. If the price hovers above its trend line, we say that price is above value; when it is hovering below, we say that it is below value. 3. Relative Strength Index (RSI). This indicator compares the recent price gains to recent price losses and converts the result to a number between 0 and 100. A value lower than 30 indicates an oversold

Price and Value

69

position, while a value higher than 70 indicates an overbought position. In fact, you can also say that the price has increased a lot and the stock is becoming overbought or the price has decreased a lot and it is becoming oversold. Traders would simply buy when the price is in oversold territory (I suspect that quite a few automatic trading systems use this indicator). 4. Moving average. A moving average (MA) of a price is used to smooth out the daily price variations in order to focus on the trend itself. For example, the 50-day simple moving average shown in Figure 2.1 is the mean of the previous 50 days’ closing prices. At each new day, the most recent price is added to calculate the mean, while the price data of the oldest of the previous 50 days is dropped out. Linear or exponential weighted moving averages use a mean of price data that have been weighted to give more importance to more recent data. The interpretation of the moving average is similar to the interpretation of the trend line. However, many traders use a combination of moving averages on different time scales to assess the potential future value of a stock. For example, some short-term traders would buy stocks whose 20 MA crosses above the 50 MA, while long-term investors would look at the 50 MA crossing above the 200 MA.

TRADITIONAL MEASURE OF “CHEAP” For an investor with a long-term view, value is best measured in terms of growth, earnings, market strength, and the like. However, a trader who typically has a shorter-term view will prefer to follow the price-based technical indicators such as those discussed earlier. The striking difference between the two approaches is that long-term investors look at the future earnings growth, while traders look at past price levels. Who is right? My gut feeling is that future earnings growth expresses real value. However, past price levels express the market’s interpretation at that time of the future earnings growth. Past price levels only represent a speculation of what future earnings will be. Therefore, for traders to look at past price levels in order to predict future prices is highly speculative. The trader who does this is just saying, “Yesterday, the stock was more expensive. Therefore, it is now comparatively cheap.” Obviously, this measure of value does not give a clue as to whether tomorrow’s price will be even cheaper. We are not here to judge who is correct and who is not. An individual is sometimes an investor, sometimes a trader, and sometimes a speculator. But what an active trader, a speculator, and an investor have in common is

70

VALUE IN TIME

that they all want to make a profit. They expect to make a profit. But is this expectation shared by others? What most people tend to forget is that buying and selling a stock mainly means dealing with other people. When you buy a stock, it means that you consider it cheap. It also means that someone else (a seller) considers it expensive. This means that the concepts of “cheap” and “expensive” had better be measured against the group of traders who are actively following that specific stock, instead of against some other measure. Why? It’s simple: In the stock market we are not trading reality, but rather the perception of the reality. Therefore, instead of talking about “cheap” or “expensive,” we should talk about “expectation.”

The Expectation Concept Indeed, we don’t really buy a stock because it is cheap or because it is a good stock. These reasons are rationalizations of the buying act. Sometimes a trader buys a stock to cover an existing position. Even in this case, the trader would not buy the stock unless the share price is expected to increase. Later on the trader expects to sell the position at a profit. The stock market is not like a supermarket. In a supermarket you buy bread because you want to eat that bread, not because you want to resell it to someone else at a higher price. If the price of the bread increases, for your next purchase you will most likely turn to some other food or some cheaper store. In the stock market, you buy a stock because you want to resell it at a higher price. Your feeling about the expensiveness of the stock (its value) is intimately linked to your expectation to sell the stock higher. To meet that expectation, the price will need to increase, which itself depends on other traders’ expectations (you will need to find a buyer). What is this expectation concept, then, and how can we measure it? Let’s consider an example. Suppose that the market for stock XYZ is composed of 500 shares, divided equally among five investors: Mary, John, Stewart, Ely, and Kris. These five investors have purchased their 100 shares at different prices (see Table 2.1). Now, suppose that two other investors, Lilly and Edward, also want to buy 100 shares each. If the market price for the XYZ share is at $10.5, Lilly and Edward simply have to find someone who is willing to sell their 100 shares at that price. Let’s suppose that both Lilly and Edward know the purchase price of the actual shareholders. With whom would they start bargaining? My guess is that Mary and John have no specific reasons to sell at the present price, since neither of them is showing a significant profit. However, any of the other three are good candidates, since they might be happy to take their profit. Now, let’s look at Table 2.2, which shows a different situation in which Ely and Kris purchased their stocks at different prices from

71

Price and Value

TABLE 2.1 Shares Distribution Resulting in a High Average Profit

Mary John Ely Stewart Kris Average profit

Shares

Buy Price

Profit

100 100 100 100 100

$10.8 $10.0 $7.0 $8.5 $9.0

−2.78% 5.00% 50.00% 23.53% 16.67% 18.48%

If you take a group of shareholders, it is probable that those who are running the highest gains will sell their shares first (we suppose here a market price of $10.5). Here, Ely, Stewart, and Kris will be more likely to sell than Mary and John.

Table 2.1. I kept the same name for each investor, but we have to view this as a completely new situation. We see that Ely is in real trouble. He has incurred such a large paper loss that most likely he will not be willing to part with his shares and realize that loss. We say that such an investor is locked in. The other change concerns Kris, who is now turning an 8.7 percent loss, maybe close to his stop-loss limit. Kris is not locked in, because his loss is relatively small. Everyone knows that it is easier to take a small loss than a large one. Kris is probably in the right state of mind to be a seller now. Neither Mary nor John is ready to sell, because their profit or loss is still small. Stewart might be willing to sell to take his profit. The conclusion is that there are now only two candidates (Kris and Stewart) who are ready to sell their shares to the two buyers. Suppose now that Stewart did not buy at $8.5, but instead at $18. If you look at Table 2.3, it is clear that Stewart has also become locked in due to

TABLE 2.2 Shares Distribution Resulting in a Small Average Loss

Mary John Ely Stewart Kris Average profit

Shares

Buy Price

Profit

100 100 100 100 100

$10.8 $10.0 $15.0 $8.5 $11.5

−2.78% 5.00% −30.00% 23.53% −8.70% −2.59%

We can see here that Ely became locked in because of his large paper loss, and Kris is ready to sell in order to avoid Ely’s fate and be locked in by a larger loss.

72

VALUE IN TIME

TABLE 2.3 Shares Distribution Resulting in a High Average Loss

Mary John Ely Stewart Kris Average profit

Shares

Buy Price

Profit

100 100 100 100 100

$10.8 $10.0 $15.0 $18.0 $11.5

−2.78% 5.00% −30.00% −41.87% −8.70% −15.63%

Only Kris is now ready to sell his shares. The others are either locked in or too close to their purchase price.

a large paper loss. The problem is that now the only possible selling candidate is Kris, but we have two parties willing to buy: Lilly and Edward. One of the two parties will have to raise the offered price in order to increase the paper gain of the other shareholders. Table 2.3 shows a supply problem: There are not enough shares potentially for sale at the $10.5 price. There are two ways to attract sellers: 1. One of the investors could increase the offering price until, for exam-

ple, John has a high enough profit to decide to sell. 2. Someone pushes the price down (by shorting and/or by selling small quantities of stocks at the right time) and therefore pushes Mary to her stop-loss limit, forcing her to sell. When reading this example, you might say that the market does not work like this, because there are always some people ready to sell their shares. If I have to raise the price by one cent, who cares? I will still get my shares. This is true if you are a retail player, but funds think differently. If you are a billion-dollar fund (a midsize fund) and you want to invest a maximum of 1 percent of your portfolio in a single stock, this would mean investing up to $10 million in a single stock. If you want to buy $10 million worth of shares of a $500 million capitalized company, you would end up owning 2 percent of the total issued shares of the company. For funds, shares availability is a critical issue. What did we learn from this example? Three things: 1. If you know the exact profit situation of all the shareholders, you are

in a much better position to properly time your trading activities (buy or sell your shares). 2. If you compare the average profits of Tables 2.1, 2.2, and 2.3, it seems that the higher the average profits (Table 2.1), the easier it will be

Price and Value

73

to find a large supply of shares to buy. The lower the average profits (Table 2.3), the more difficult it will be to find cheap shares. 3. Shareholders who have just bought shares still have a high expectation

of generating a profit; therefore, they will not sell until their expectation is met or until they reach their maximum endurable loss under their risk management policy—if they have one. We will come back in Chapter 4 to point 3, which tends to give us some insight on how to measure the supply of shares. Let’s examine points 1 and 2 again. It seems from these points that if we monitor the average profits of all the shareholders, we will have more chances to improve our timing for our trading activities, even if we do not know the exact profit position of each shareholder. Let’s come back to our XYZ company. This time, suppose that 20 traders are involved. Suppose now that none of them is at present invested in the XYZ company, whose stock price is sitting at $10. If XYZ is a good company with good prospects, we may believe that the great majority of these 20 traders will decide to buy the stock. We may rightfully believe that these traders will have positive expectations of the stock price appreciating. Now, suppose that one of the traders already bought XYZ at $5 one month ago. Do you believe that, today, this trader’s expectation regarding future price appreciation is the same as the expectations of the other 19 traders? Probably not! That trader has been invested for a month. It is clear that, because he is sitting on a 100 percent profit, his expectations have already been met. His expectations of a further price gain would be significantly lower than the expectations of the other 19 traders. In other words, if you look at the average expectation of the 20 traders, that trader’s expectation brings the average down. Now, suppose that this group of 20 traders is the only active trader group that is authorized to trade the XYZ shares, but that, as a specialist, you are also authorized to trade XYZ, and you have access to all the transactions that the group of 20 traders performs. Couldn’t you make an excellent profit on the stock by just tracking the average expectation of that trader group, buying when expectation is high and selling when it is low?

Return on Investment as a Measure of Expectation This means that it pays to track expectation. We saw that the early trader’s expectation was lower than that of the other traders. Just how low? We do not know, but certainly lower.

74

VALUE IN TIME

We can say that the higher the profits traders are enjoying, the lower their expectations will be of getting further profits from their shares. The propensity to sell their shares will increase as their profits increase. In mathematical terms, we say that a trader’s expectation at time t of a further price increase is inversely proportional to the profit that this trader is enjoying at time t. This means that, for example, if you already have a 50 percent profit, your expectation for a further profit increase is lower now than at the time you bought the stock, when you had 0 percent profit. Please note that from now on, I will use the term return on investment (ROI) to mean the profit or loss (expressed as a percentage) compared to the initial purchasing cost. Float ROI Forgetting our 20 traders for the time being, suppose that trading of XYZ is opened to the public. However, let’s now suppose that each trader is authorized to own only a single share of our XYZ company. Let’s also suppose that XYZ has 100 million shares outstanding and ready to trade. This means that 100 million traders play that XYZ stock with their only share. They will buy their single share, resell it later, and so on. Let’s take the hypothesis that ROI is reflecting expectations quite well! We can then easily calculate at time t the average ROI of all 100 million shares that were traded in the past. To do this, we calculate the ROI of each share, take the sum of those ROIs, and divide the sum by the total number of shares (100 million). If you repeat the calculation at time t + 1, t + 2, and so on, you will end up with a visual picture of the evolution over time of the ROI of the 100 million shares (see Figure 2.2a and b). I call this concept the Float ROI. It is the average ROI of all the shares that are available for trading.

FIGURE 2.2a Google share price.

Price and Value

75

FIGURE 2.2b Google: Active Boundaries for 219 million shares. The Upper Boundary sets the limit at which price tends to reverse downward. The Lower Boundary sets the limit at which price tends to reverse upward.

Let’s analyze Figure 2.2a and b. Figure 2.2a shows the share price of the company Google during the nine-month period prior to September 1, 2006. Notice the swings in the share price. These swings are rather well captured in Figure 2.2b, which shows, for the total of the 219 million issued shares, the average ROI calculated at each trading minute. (You can imagine the number of calculations involved: It took 872 million mathematical operations to build Figure 2.2b. Using mathematical optimizations, this type of calculation can now be executed in a few seconds, but this was very burdensome to calculate with the slow computers that existed just a few years ago.) If you look at Figure 2.2b, you will see that I drew two horizontal lines. The upper line, which I call the Upper Boundary, joins as much as possible consecutive maxima in average ROI. The lower line, which I call the Lower Boundary, joins as much as possible minima in average ROI. What Do These Lines Say? The Upper Boundary says that at this high level of collective ROI, the average expectation for a further price increase is very low, and we may expect more profit taking to occur. The Lower Boundary says that at this low level of collective ROI, the average expectation for a further price decrease is very low, and we may expect more bargain seekers. I selected Google because of the high turnover of its total float: The 219 million shares exchange hands every 34 trading days on average. This is very fast, compared to the standard turnover of around 90 to 180 trading days. Google is a very actively traded company. This is the reason the average ROI on the float captures the price movements so well. On less

76

VALUE IN TIME

actively traded companies, the average ROI on the float would not give such good results, because the number of shares belonging to active traders is much lower than the total number of issued shares. Let’s examine what I mean by active traders. Look for the Actively Traded Shares The most common measure of a company’s profitability is its earnings per share (EPS). The EPS is then compared to the stock price using the price-earnings (P/E) ratio. To calculate the profit per share, it is best to use the fully diluted number of shares, including the shares that must be issued in the future when employee and management options are exercised, or shares that are part of a convertible debenture. However, to calculate the liquidity, people use the available float, which is the number of shares that are issued and available for trading. Since insiders are restricted from selling their shares, and since institutional investors move rather slowly, the available float is smaller (sometimes much smaller) than the number of issued shares. When large buying/selling volume shows up, the price of very illiquid shares tends to move more quickly than the price of very liquid shares. In reality, I have noticed that the number of shares that are available for trading at a certain point is much lower than what can be calculated in theory. Many shares are locked in by long-term investors or by funds. Lockedin shares will be sold either (1) on exceptional circumstances (such as news that completely changes the long-term outlook for the company) or (2) slowly over time, at a pace that is much slower than the price swings. This is why I use still another calculation, which is the Active Float. The Active Float is defined as the number of shares that are actively traded. The Active Float is smaller than the available float. This number of actively traded shares is not a publicly available figure, and it must be readjusted on a yearly basis. My method for approximating the Active Float is described next. How Can We Evaluate the Active Float of a Company? First, we have to recognize that there are many types of shareholders, all working in different time spans. Let us separate them into two groups: investors and traders. Usually, investors have a much longer investment time span than traders, who come in and out of stocks rather quickly. Now, there is no clear distinction between the two groups, but separating them in our minds allows us to focus only on the traders. Let us take a few hypotheses:

r The first hypothesis is that for most companies, the pool of active traders is very stable over time. After studying a company, active

Price and Value

77

traders will trade it on multiple occasions. Indeed, since it took a lot of time for them to read the company’s financial reports, to study competitors, and to analyze the market and growth perspectives, they will naturally prefer to stay with a company that they feel they know very well. In general, active traders follow the stock of the same company for years, while the price trend itself changes in shorter periods of time such as weeks or months. At a minimum, we can say that the pool of active traders for a given stock changes much more slowly than the price cycle of the stock. r The second hypothesis is that each single trader does not change his trading strategy overnight. This means that all things being equal, traders will perform the same analysis and will react the same way as they did before. Also, their endurance for loss or their joy when running profits reflect their own personalities, which will not change overnight. This means that the Active Float ROI will show very stable reversal points in time, because it is an inverted image of the average expectation of the active shareholders. r The third hypothesis is that computers automatically generate a large part of the trading that occurs in the stock market. Program trading requires a long development time, with very slow adjustment cycles compared to market swings. This means that for the same stock, an algorithm will make its consecutive automatic trading decision by following the same programmed logic. The Active Float ROI will thus nicely catch this repetitive pattern generated by program trading. Basing my reasoning on the last two hypotheses, I define the Active Float as the number of shares that are necessary to give to the Active Float ROI pattern the most regular set of reversal points. This means that if the total float of a stock is 100 million shares, I first plot the Float ROI graph based on these 100 million shares. I then plot another graph of a Float ROI based on 90 million shares, another one on 80 million shares, and so on down to 10 million shares. I then compare the graphs and take the one that gives the most regular reversal pattern. What the hypotheses say is that this pattern will continue in the future as long as no major event hits the company. If a major event occurs that completely changes shareholders’ expectations regarding the company’s future earnings, this pattern will no longer be valid. Therefore, a new pattern will develop that will reflect the new expectation of the active pool of shareholders. Because the term ROI could be confused with the concept of return on investment that is most known for a company investing in assets, and because the Active Float terminology could be confused with the available float, from now on I will use the term Active Boundaries instead of the Active Float ROI terminology.

78

VALUE IN TIME

FIGURE 2.3 Becton, Dickinson: share price evolution.

Figure 2.3 shows medical technology company Becton, Dickinson and Company’s share price evolution in the 12 months ending September 1, 2006. We see the stock price evolving in five phases: 1. Phase A: three months; slow price slide resulting in a 10 percent

decrease. 2. Phase B: 4.5 months; uptrend for a 30 percent gain. 3. Phase C: 2.5 months; 8 percent retracement. 4. Phase D: two months; trading range. 5. Phase E: three months; uptrend for a 20 percent gain.

As you can see, Figure 2.4 captures the stock price movements rather nicely. Let’s look at this in detail (I will show later how to calculate the Active Boundaries):

r Notice that the general Active Boundaries pattern moves between a −5 percent Lower Boundary and a 10 percent Upper Boundary. Because the 10 percent average profit is twice as high in amplitude as the −5 percent average loss, this indicates that the stock price is in a long-term uptrend. r The phase A downtrend is well captured between the Lower Boundary and an intermediate Upper Boundary (Int. UB) that is located at a level of 0 percent. Please note that if the Upper Boundary is close to or lower than 0 percent, this indicates a downtrend, since below 0 percent the average profit is negative. Also note that at the beginning of November 2005, at point 1, there was a large price gap upward due to a good earnings release. This news changed the way that shareholders regarded

Price and Value

79

FIGURE 2.4 Becton, Dickinson: Active Boundaries for an Active Float of 80 million shares. The 80 million shares that constitute the Active Float for Becton, Dickinson capture the price reversals very well, allowing for a good evaluation of the value of the stock.

the company’s potential future profits. Phase A was broken at point 1, when the signal crossed over the intermediate Upper Boundary. r Phase B forms a strong uptrend. The price gap upward is followed by a small retracement until point 2. This retracement, which leaves us well inside the positive average profit zone, indicates that a new uptrend is probably emerging. We can draw an intermediate Lower Boundary (Int. LB) starting at point 2, which will monitor the uptrend. r Phase B was broken at point 3, where the new downtrend of phase C starts. This C downtrend continues until, at the start of phase D, we again reach the long-term Lower Boundary. r At the right of Figure 2.4, the last phase (phase E) has now reached the Upper Boundary, where active traders have historically judged the stock price as expensive. We may expect the price to reverse down from that point. Figure 2.5 shows that taking 40 million shares to calculate the Active Boundaries does not allow us to nicely catch the different phases the same way as was possible in Figure 2.4. The whole process of selecting the right number of shares takes a few trials and requires past data for at least a six-month period. At key turning points, such as when the Active Boundaries signal hits the Upper or the Lower Boundary, the Effective Volume analysis, which we studied in Chapter 1, will often show the direction of the next move. For example, in the case of Figure 2.6, the end of downtrend A is signaled by the accumulation in shares preceding October 31, 2005.

80

VALUE IN TIME

FIGURE 2.5 Becton, Dickinson: Active Boundaries for an Active Float of 40 million shares.

FIGURE 2.6 Becton, Dickinson: Effective Volume analysis up to October 31, 2005. The combination of the Active Boundaries signal and the Effective Volume analysis often shows the future direction of the price. In this case, the divergence of the large players’ shares accumulation during the downtrend indicates that the downtrend is coming to an end.

Price and Value

81

WHY DO TRENDS EXIST? Trends exist because members of the group of active shareholders have concurrent levels of long-term expectation regarding the stocks’ future price movements, but divergent short-term expectations. It is the average of their long-term expectation that helps to form a price trend, and it is the average of their short-term expectations that moves the stock price within the trend’s boundaries. Active Boundaries, since they are measuring a range of expectations, are an excellent tool for monitoring a trend. Let’s take the example of Tellabs (Figure 2.7a and b). Tellabs is a large telecommunications equipment provider. Between July 2005 and May 2006, the share price gained 100 percent (Figure 2.7a). It is interesting to follow it step by step, from point 1 to point 14. In Figure 2.7b, points 1 and 2 define the Active Boundaries. Point 3 confirms the Upper Boundary. You will notice that the price at point 3 is higher than the price at point 1, even if they both command the same level of Active Boundaries. In both cases, active traders consider that the Tellabs stock price is high and that in the short term they do not expect it to increase further.

Active Boundaries to Monitor Uptrends Let’s examine points 6 and 7. These are continuous hits of the Upper Boundary without pulling back to the Lower Boundary. This is typical of a strong uptrend. At point 6, the Active Boundaries indicator does not tell

FIGURE 2.7a Tellabs: price evolution.

82

VALUE IN TIME

FIGURE 2.7b Tellabs: Active Boundaries. Active Boundaries monitor active traders’ expectations. Using Active Boundaries is an excellent way to monitor trends, because trends are created by a combination of the long- and short-term expectations of active traders.

whether the price will reverse down or will continue to move up. It just says that it is hitting the expensive Upper Boundary limit, a limit where traders usually reverse their positions. Since they did not reverse position, we may conclude that the trend is strong. At this stage, it is better to observe the Large Effective Volume in order to assess the behavior of large players. If the Large Effective Volume indicates profit taking, or even a lack of buying by large players, we can conclude that the continuous price increase will soon slow, stop, and possibly reverse down. Points 9 to 11 also reflect the same strong uptrend characterized by consecutive hits of the Active Boundaries signal to the Upper Boundary limit. As you can see, the Active Boundaries indicator does not give buy and sell signals. It simply indicates when the expectation of active traders is high or low. When the average profit of the active traders is high, this means that their expectation for a further price increase is low. This indicates that the probability for the price uptrend to end is high. However, this does not guarantee that the price will decrease. In a strong uptrend such as the one between points 9 and 11, my strategy is to sell only when the large players are no longer buying. Indeed, if large players are not buying (since the price uptrend will attract profit takers), it means that retail players are the only ones fueling the uptrend. In Figure 2.8a and b, I represent the Effective Volume analysis, the price evolution, and the Active Boundaries of Tellabs over the 80 trading

Price and Value

FIGURE 2.8a Tellabs: Effective Volume and price evolution.

FIGURE 2.8b Tellabs: Active Boundaries.

83

84

VALUE IN TIME

days ending March 6, 2006. This allows a much better interpretation of what is going on for the stock:

r From point 6, even though we are at the Upper Boundary, the Large Effective Volume flow shows a strong accumulation, indicating that selling is not warranted. r From point 7, although the price has started to decline, the healthy accumulation is continuing. In such a case, you can either sell to protect your profits or decide to go on since large players are continuing their accumulation. If you sell, the probability that you will buy back later at point 8 is very limited, because at point 8, the signal slightly and quickly touches the Lower Boundary before surging sharply due to a price gap. The rapidity of the move makes it unlikely that you will be able to catch the opportunity. (Disclosure: I sold at point 7 and took my profit, but had decided to buy at point 8 after the market closed. Unfortunately, the price gapped up the next day and continued up without me.) r From point 9, the accumulation is increasing. r From point 10, even if the price continues its uptrend, the accumulation by large players has stopped. You don’t have to be a rocket scientist to know that without fuel, the rocket will stop its ascent and eventually fall. I have to clearly state that I do not believe that large players are wiser than retail investors. I even find that many retail investors perform deeper research on a company before and after they invest than large players do. They do that because they are risking their hard-earned money and because they know how strong the competition is. The only advantage they can get over the competition is by doing research and using good risk management techniques. Let’s come back to large players for a moment. Even if they are not necessarily wiser than retail investors, large players must be followed very closely, because they are the ones providing market liquidity. We already saw in the previous chapter that the Large Effective Volume flow detects the movements of large players, who often are responsible for starting trends. It should also be clear that uptrends couldn’t be sustained if large players were not involved. In an uptrend, a pullback can be monitored by following the behavior of large players. If the Active Boundaries signal hits the Upper Boundary while continuous accumulation by large players is taking place (as shown by the Large Effective Volume pattern), it indicates that the uptrend will continue. As you see, it is the combination of the Active Boundaries and the Effective Volume that allows us to take advantage of pullbacks during

Price and Value

85

uptrends. The Active Boundaries indicator allows you to see the positions of the active players. It is a static view of the market. The Effective Volume gives you a view of the tactical decisions of large players. It offers a dynamic view. What we will see in the next chapter is how to detect strategic decisions.

Is It Better to Trade Bounces and Pullbacks or Long-Term Trends? When a stock is trading at $8.5 in September 2005 and at $17 in May 2006, it is evident in hindsight that it would have been wise to buy at $8.5 and sell at the top (a 100 percent gain in about nine months). The problem is that at point 2 of Figure 2.7a and b, it is impossible to know that the stock has a 100 percent gain potential. The Active Boundaries do not help in any way to give you a clue about the breadth of the trend. Active Boundaries do, however, help in two ways: 1. They enable you to detect value inside of a trend, especially in combi-

nation with the Effective Volume. 2. They enable you to detect when a long-lasting trend is broken. In Chapter 6, I revisit in detail the different trading strategies that could be planned using the various indicators presented throughout the book. However, if we try to play the bounces and pullbacks in Figure 2.7a and b we could have the following scenario:

r We decide to invest at around $9.5, at point 4. Indeed, the previous

r

r r

r

points 1, 2, and 3 are utilized to set the Active Boundaries for the actual trend. At point 5, there is a first unexpected pullback. Suppose that we are lucky enough for our stop loss not to be hit. The pullback to $8.8 could have very well hit a stop loss, since it was a decrease of 7.4 percent from our buying price. At point 6, we hit the Upper Boundary and sell at around $10.8 for a 13.7 percent profit. We watch with great sorrow (since we just sold off) as the stock price rises to point 7, but then it dips to the Lower Boundary of point 8. This is a very critical point, because if we miss the point 8 buy opportunity, we miss the trend from point 8 to point 11. Point 8 is very difficult to catch, because the Active Boundaries signal only touches the Lower Boundary once and then jumps out. Suppose that we catch the buy signal at point 8. Without the Effective Volume signal to tell us to keep the stock for the uptrend from point 9

86

VALUE IN TIME

to point 11, we would have sold at point 9 and missed the subsequent uptrend that would have given us a 17.6 percent gain: from $12.5 to $14.7. That is correct if we could have exactly guessed the point 11 sell, which is impossible to guess if we rely only on the Active Boundaries indicator. As a matter of fact, point 11 is really detectable only through the divergence analysis tool that is explained in Chapter 3, even if the Effective Volume analysis is already showing that we are reaching the end of the uptrend at point 11. A Comment on Shorting You may have already noticed that I do not talk about shorting. If the tool is so good as to detect when a stock is turning expensive, why not short it? We could! Trading is a two-way street, and shorting is a respectable strategy. At this point in the book, I believe that it is too early to introduce shorting through modern volume analysis. Chapter 7 will show shorting examples, but I first need to walk you through other subjects, such as the price/volume divergence analysis and the supply/demand equilibrium.

If we add to the previous gain the last 21 percent profit that was generated by the uptrend starting at point 12 up to point 13, the whole process of timely buying and selling would have produced a 77.7 percent profit (without reinvesting the gains from the previous trades). Compared to the 100 percent profit of the long-term holding, it is clear that long-term value investors would have handsomely beat active traders, if they had guessed value and initial timing correctly.

What Are the Advantages and Disadvantages of Swing Trading? A swing trader typically catches short-term price trends that range from a few days to a few weeks. There are several advantages and disadvantages to swing trading. Some of the disadvantages are:

r Swing trading involves more buy/sell decisions, which implies the possibility of more mistakes, and also the increased costs of frequent movements (commission and slippage). It is unrealistic to believe that even with the best tools, we will be able to obtain the 77.7 percent profit stated earlier. I consider it a good job when I grab 30 percent of a long-term trend, and an exceptional job when I get 50 percent. r Swing trading involves a lot more work and stress than long-term investing does.

Price and Value

87

Some of the advantages of swing trading are:

r When you swing trade, you are invested in a specific stock only during the time of the swing. Therefore, if you calculate the profit you generate for each day during which your money is invested, you will see that most often the return will be greater for swing trading than for longterm investment. In our example, we were invested only one-third of the time with swing trading compared to the length of time of a longterm investment strategy, but we produced more than three-quarters of the profit generated by that long-term strategy. If during the idle periods we can invest our money in other positive swing trades, we could be better off than with long-term investing in a single stock. r The second, lesser-known benefit lies in the lower risk implied by swing trading: Swing trading allows using risk management techniques that give good results, because the risk is fractioned into multiple trades that are well controlled during the course of their lifetimes. Long-term investors typically allow for much larger intermediate losses, because they know very well that they have bought value and they are confident that, in the long term, value always prevails. These investors typically hold losing positions for years, and some even die with their losing positions. I believe that I will one day inherit from my father real paper stock certificates of now-unknown African mining companies that he bought in the 1960s, companies that have long since been out of business. Active Boundaries may be used for any investment style. What is important to understand is that you need to use the Active Boundaries tool differently depending on your investment style. Investing is a very stressful game, because you are risking your hardearned money to get future profits. When investors enter the market, they know that they may lose money. However, it is only when they actually start losing that things turn bad: “What could I have done with the money that I lost? How many hours, days, or weeks will I have to work to gain that money back? Maybe I should risk more money to make up the loss.” The same types of emotions arise when you start making money: “Maybe I’ll be able to move into a nicer place. Maybe I’ll get the new car I wanted to buy.” Needless to say, this is not how you make rational investment decisions. The market does not care about your new car or even about the trading position that you just took. To be more specific, all the other players are trying their best to grab your money, and believe me, they want to use their profit for a new car, too. One of the main advantages of the Active Boundaries indicator is that it will help you dissociate your personal objectives from what the market

88

VALUE IN TIME

is telling you. This will allow you to make decisions based only on market signals. However, you must use it in a way that matches your investment style.

Active Boundaries and Price Gaps The gap down to point 12 in Figure 2.7a and b is interesting to point out, because it covers almost exactly the distance between the Upper and the Lower Boundaries. It is comparable to the gap up from point 8 that also covers the distance between the Upper and Lower Boundaries. When will a gap in price push or not push the Active Boundaries signal beyond the Upper and Lower Boundaries? The answer relies on the definition of the Upper and Lower Boundaries and what they represent: They represent the limits of the value of the company in the eyes of the active traders. The stock price will tend to evolve between these limits. A price gap is the result of news that came after the market closed for the day. If the news is within the normal business development of the company (such as an earnings surprise), the active traders’ expectation will not change, and the Active Boundaries signal will continue to fluctuate within the actual Upper and Lower Boundaries. However, if the news is so strong (SEC inquiry for options back-dating, earnings restatement, etc.) that it changes the company’s future itself, then the expectation of the active traders will widely change, and the Upper and Lower Boundaries themselves are likely to evolve. Let’s take a look at a real-life example: IMAX Corporation (Figure 2.9a and b). IMAX is well known for its trademark large-screen movie theaters. Figure 2.9a shows how the stock price reacted to negative news that was published by the company on August 8, 2006. IMAX had been putting

FIGURE 2.9a IMAX: price evolution.

Price and Value

89

FIGURE 2.9b IMAX: Active Boundaries. A significant price gap may result from overnight news. If the news is important enough, active traders could change their opinion about the company, which will result in a new set of Active Boundaries. This new set will correspond to the new expectation of the active traders.

itself up for sale and just announced that it could not find any buyer at its target price, which was probably higher than $11 per share. The company concurrently announced an informal SEC inquiry related to the way revenue had been reported in the past. The speculative high prices were cut by 40 percent, and the Active Boundaries signal broke its Lower Boundary. Shareholders’ expectation of a potential share price appreciation had changed so strongly overnight as to form a new set of Active Boundaries. In other words, it is as if there had been an old IMAX and a new IMAX. This change of expectation was due more to a very rapid rotation of shareholders on the bad news than to a change in the expectation of the previous shareholders. As we can see in Figure 2.10, about 20 million shares changed hands during the three days following the August 8 news. This volume is close to the number of active traders as computed by the Active Float method. This means that the total pool of active traders changed within three days. In conclusion, we can say that if a price gap pushes the Active Boundaries signal far beyond the Upper Boundary or the Lower Boundary, the company is likely going through fundamental changes that force us to reconsider its real value.

Active Boundaries and Trend Reversals Active Boundaries do not help much in detecting price trend reversals, because we can detect reversals only when these reversals are strong enough to pierce through the Lower Boundary. As can be seen in Figure 2.11a,

90

VALUE IN TIME

FIGURE 2.10 IMAX: volume surge on price gap downward. Very heavy volume happening in conjunction with a dramatic price change usually indicates that the company is going through a fundamental change. Source: Chart courtesy of StockCharts.com.

Tellabs reversed its trend from point 14, but it is only at point B that it was noticed, when the signal pierced through the Lower Boundary (see Figure 2.11b). At point B, the strength of the downward movement was too strong for it to be considered a mere correction. Do not forget that in the uptrend, we would already have sold at point 13, as explained with Figure 2.7a and b.

FIGURE 2.11a Tellabs: price evolution.

Price and Value

91

FIGURE 2.11b Tellabs: Active Boundaries. The Active Boundaries signal is late at pointing out trend reversals. These reversals are shown when one of the boundaries is broken, usually on heavy volume.

Point B was followed by two consecutive lower highs (H1 and H2) that showed how weak buyers had become. On the Active Boundaries signal, if the signal reverses upward when it hits the Lower Boundary but afterward reverses downward again before reaching the Upper Boundary, this indicates a strong selling pressure. Point C is interesting to analyze: We experience a steep drop in price from $11 to $9, followed by a small trading range around $9 and then a slow climb back to $10. However, the Active Boundaries signal shows a steep drop from −10 percent to −25 percent followed by a quick reversal during the trading range, and then a quick rise to +10 percent (the Upper Boundary). This is very peculiar to the Active Boundaries indicator: It will mainly have short-term moves due to price variations and long-term moves due to the much slower volume changes.

Price and Volume Changes Impact the Active Boundaries The Active Boundaries movements are influenced by price and volume changes:

r r

Price changes are responsible for the short-term Active Boundaries signal variations. Volume changes are responsible for the long-term Active Boundaries signal variations.

92

VALUE IN TIME

In the case of Tellabs, however, from the C gap downward and during the following five days, more than 70 million shares were exchanged (Figure 2.12), which is about half of the Active Float. (The Active Float is the number of shares that are used to calculate the Active Boundaries indicator. For Tellabs, the Active Float was 150 million shares.) In this case, since the exchanged volume was very important, the Active Boundaries signal showed a steep change due to the combination of steep price and volume changes.

Active Boundaries in Downtrends Using the Active Boundaries indicator, downtrends can be monitored the same way as we monitor uptrends. Meridian Resource Corporation (TMR) is a natural gas exploration company that in 2006 was experiencing a set of dry holes and hence declining reserves. Figure 2.13a shows the stock price downtrend, while in Figure 2.13b both the Upper and the Lower Boundaries indicate the most probable reversal points in the downtrend. Once again, these reversal points are not pure buy or sell signals. They still need to be combined with the Effective Volume signal, which indicates what large players are doing at these critical points. An easy rule of thumb that you can use to determine if the stock is in a long-term downtrend or in a long-term uptrend is to add the value of

FIGURE 2.12 Tellabs: volume surges on price gaps downward. Source: Chart courtesy of StockCharts.com.

Price and Value

93

FIGURE 2.13a Meridian Resource: price evolution.

FIGURE 2.13b Meridian Resource: Active Boundaries. Active Boundaries are very useful for the monitoring of downtrends. They must be used in conjunction with the Effective Volume tool in order to detect reversals within a trend.

the Upper Boundary to the value of the Lower Boundary. A positive result means that, on average, active traders have positive returns, which is typical of uptrends. If the result is negative, the opposite is true. In Figure 2.13b, the Upper Boundary is set at 10 percent, and the Lower Boundary is set at −22 percent. The sum is −12 percent, indicating a downtrend. The midpoint is −6 percent (midpoint = 10% − [(10% + 22%) ÷ 2]). This midpoint figure is important, because it tells you the force you are up against. Indeed, if you decide to go long at the Lower Boundary level (points 1, 3, 7, and 12 in Figure 2.13b), you know that you have to fight a −6 percent downtrend that is working against you. Needless to say, you

94

VALUE IN TIME

must be very confident about what you are doing. I would never purchase a stock that is in a downtrend, even at the Lower Boundary, except if large players are signaling a coming trend change.

Trading on Active Boundaries The price trend is up when (Upper Boundary + Lower Boundary) > 0. The price trend is down when (Upper Boundary + Lower Boundary) < 0. From this we can deduct two simple trading rules:

1. Go long when:

r r r

The long-term price trend is up, The Active Boundaries signal is reaching the Lower Boundary, and Large players are buying.

2. Sell when:

r r

The Active Boundaries signal is reaching the Upper Boundary, and Large players are selling.

How to Measure Your Profit Potential Using Active Boundaries Active Boundaries catch the upswings and downswings and give good indications about the potential trading profit that you can make by buying at the Lower Boundary and selling at the Upper Boundary. This profit can be roughly evaluated by looking at the spread between the Upper and the Lower Boundaries. For example, Figure 2.13b shows that the spread is 32 percent. You might therefore think that when buying at the Lower Boundary, if the signal later reaches the Upper Boundary, you will make a 32 percent profit. This is unfortunately not always the case, but in reality it is still a good approximation. The profit could be higher or lower than 32 percent. Table 2.4 lists the trading profits that could have been made by buying at the Lower Boundary and selling at the Upper Boundary of the Active Boundaries indicator for Meridian Resource, as shown in Figure 2.13b. You can see that buying at point 1 and selling at point 2 generated only 11 percent profit. The reason for this difference is that between point 1 and 2, the price was generally trending down, even if at point 2 the price is slightly higher than what it was at point 1. In general it is easier to make a profit if

95

Price and Value

TABLE 2.4 Meridian Resource: Trading Profits from Long Trades Buying Point

Selling Point

Buying Price

Selling Price

Real Profit

1 3 7

2 4 8

$5.7 $3.9 $3.5

$6.3 $5.4 $4.5

11% 38% 29%

you are long in an uptrend rather than short, because you do not have to fight against the trend. Table 2.5 shows the profits made by shorting at the Upper Boundary and covering at the Lower Boundary.

The Influence of the Active Float on the Spread between the Upper and the Lower Boundaries It is interesting to note that the size of the Active Float has a direct impact on the spread between the Upper and the Lower Boundaries. Indeed, the smaller the Active Float, the smaller the spread between the Upper and the Lower Boundaries. However, the smaller the Active Float, the more reversals you will have on the Active Boundaries signal, pushing you into overtrading. This is illustrated in Figure 2.14, which shows the Active Boundaries signal for an Active Float that is 6.5 times smaller than the Active Float used in Figure 2.13b. We can see that the spread between Upper and Lower Boundaries fell to 20 percent from the 32 percent seen in Figure 2.13b. We will also notice that there are many more reversals in Figure 2.14. However, what I find the most striking is that when we select a small number of shares for the Active Float (compared to the total number of issued shares), the Active Boundaries signal is then influenced only by short-term price changes and completely ignores the long-term trend. TABLE 2.5 Meridian Resource: Trading Profits from Short Trades Buying Point

Selling Point

Buying Price

Selling Price

Real Profit

2 4 8

3 7 12

$6.3 $5.4 $4.5

$4.0 $3.5 $3.3

37% 35% 28%

Source: It is easier to make profit if you are long in an uptrend or short in a downtrend than the opposite. In the case of Meridian Resource, Table 2.5 shows a slightly higher profit than Table 2.4.

96

VALUE IN TIME

FIGURE 2.14 Meridian Active Boundaries for an Active Float of 10 Million Shares. Source: Trend direction = Upper Boundary + Lower Boundary = 10% − 10% = 0%.

This can clearly be seen in both Figures 2.14 and 2.15, where the very small number of shares used for the Active Float gives a general trend of 0 percent (the midpoint between the Upper Boundary and the Lower Boundary), compared to a real long-term downtrend of −6 percent shown in Figure 2.13b. If we use a small number of shares for the Active Float, the Active Boundaries signal will miss the long-term trend direction. Table 2.6 summarizes for the company Meridian the influence of the size of the Active Float on the other parameters such as the profit potential and the average number of days between trend reversals.

FIGURE 2.15 Meridian Active Boundaries for an Active Float of 5 Million Shares. Source: Trend direction = Upper Boundary + Lower Boundary = 6% − 6% = 0%.

97

Price and Value

TABLE 2.6

Summary of the Influence of the Active Float Size on the Profit Potential and the Trend Reversal Cycle

Active Float

Active Float Turnover

Profit Potential

Trend Reversal

65 Million shares 30 Million shares 10 Million shares 5 Million shares

99 45 15 8

32% 28% 20% 12%

60 56 20 12

days days days days

days days days days

For example, if we take the last line of Table 2.6, we can see that an Active Float of five million shares will turn over in eight days. This means that in eight days, five million shares of Meridian Resource are exchanged. In other words, after eight days, new active shareholders have replaced previous active shareholders. The average profit of these new active shareholders who own the last five million exchanged shares will therefore rapidly fluctuate around 0 percent. Indeed, these new active shareholders will have only eight days to make a profit or a loss, and within eight days the price trend has no time to fully develop. Therefore, as shown in Table 2.6, the lower the Active Float, the shorter the trend reversals and the smaller your profit potential will be—although this lower profit can be repeated on more trades. However, do not forget that in the case of Meridian, buying at the Lower Boundary and selling at the Upper Boundary forces you to fight against the −6 percent downtrend. Table 2.6 shows that an Active Float of five million shares allows you a maximum profit of only 12 percent. Once again, it is much safer to trade in the direction of the price trend!

How to Set a Profit Target Using Active Boundaries Let’s examine how to set a profit target. Openwave Systems Inc. is a software provider for wireless communications. I traded the company on several occasions before its incredible drop that started in April 2006, and Active Boundaries were very helpful in indicating the profit potential and setting price targets for my trades. First, on the general price movement, we can clearly see in Figure 2.16a that during the analysis period, the stock evolved within three different trends, which were captured by three sets of Active Boundaries (Figure 2.16b): 1. The first trend (A) was a slow 2.75 percent uptrend that lasted over

a long period of about 10 months. The 2.75 percent is the midpoint

98

VALUE IN TIME

FIGURE 2.16a Openwave Systems: price evolution.

between the Upper Boundary value (12 percent) and the Lower Boundary value (−6.5 percent). 2. The downtrend (B) was very strong (−7.5 percent) and lasted only

three months. 3. The new uptrend (C) that can be seen at the right of the graph is a 6 percent uptrend. Let’s look at Figure 2.16b. At point 1, on December 5, 2005, if you decide to invest, you have to give yourself a profit target within the range of the Upper Boundary value. At that time, the Lower Boundary (A) was set at −6.5 percent. The stock price for point 1 corresponding to this Lower

FIGURE 2.16b Openwave Systems: Active Boundaries. Active Boundaries allow you to evaluate a realistic price target when investing in a swing trade.

99

Price and Value

TABLE 2.7

Openwave Systems Target on December 5, 2005

Target to Upper Boundary Percentage increase to Upper Boundary

$19.11 19.8%

Boundary was $15.95. This means that, on average, at point 1 the active shareholders were losing 6.5 percent on their investment, and that the average buying price of the active shareholders was 6.5 percent higher than the current price of $15.95: Average buying price = $15.95 ÷ (1 − 0.065) = $17.06 At point 1, for the average profit to reach the Upper Boundary, we will have to increase the average buying price by 12 percent: Target price for the average price to reach the Upper Boundary = $17.06 × (1 + 0.12) = $19.11 = Target when buying at $15.95 As shown in Table 2.7, if you can enter a trade at $15.95, this $19.11 target price, if achieved, would give you a 19.8 percent gain. Of course, the real price when we reach the Upper Boundary depends on the number of shares exchanged between point 1 and point 2, but this target calculation gives you a good evaluation of your potential trading profit. This potential profit must be compared to the potential maximum trading loss. For example, if you fix a stop loss at −8 percent, the potential gain is 2.4 times the potential loss: 19.8% ÷ 8% = 2.4 At point 2, on January 9, 2006, if you decide to short the stock, you have to give yourself a profit target within the range of the Lower Boundary value. At that time, the Upper Boundary (A) was set at 12 percent. The stock price for point 2 corresponding to this Upper Boundary was $19.73. This means that, on average, at point 2 the active shareholders were earning 12 percent on their investment, and that the average buying price of the active shareholders was 12 percent lower than the current price of $19.73: Average buying price = $19.73 × (1 − 0.12) = $17.36 At point 2, for the average profit to reach the Lower Boundary, we will have to decrease the average buying price by 6.5 percent: Target price for the average price to reach the Lower Boundary = $17.36 × (1 − 6.5%) = $16.23 = Target when shorting at $19.73

100

VALUE IN TIME

TABLE 2.8

Openwave Systems Target on January 12, 2006

Target to Lower Boundary Percentage decrease to Lower Boundary

$16.23 −16.7%

As shown in Table 2.8, if you can short at $19.73, this $16.23 target price will give you a 16.7 percent gain. This potential profit must be compared to the potential maximum trading loss. For example, if you fix a stop loss at 8 percent, the potential gain is about twice the potential loss: 16.23% ÷ 8% = 2.03 You probably have noticed in this small example that investing in the direction of the trend usually gives you a higher profit potential than investing against the trend. In our example, going long in the uptrend offered us a 19.8 percent profit potential, while shorting the reversal on the same uptrend offered us only a 16.7 percent profit potential, because of the 2.75 percent uptrend (A) that we needed to fight against when shorting.

How to Set the Active Boundaries The Active Boundaries indicator is an oscillator whose natural tendency is to return to zero. (An oscillator is a type of indicator that will oscillate around a specific value, which is usually zero.) Two forces trigger the movements of the Active Boundaries indicator: price and volume. Price is responsible for the quick movements: If the price increases one cent, it affects the profit/loss linked to all the shares that were previously bought. Therefore, the average profit/loss will move accordingly. However, it is volume that produces the slow-moving reversal pattern. Therefore, if the volume exchanged every day is high compared to the Active Float, then the volume change will have an important influence on the Active Boundaries signal reversals. This is why Figures 2.14 and 2.15 show a very jagged Active Boundaries signal: The Active Float used in Figure 2.15 was five million shares, which is only about eight times the daily number of traded shares. It is obvious that the float number used for the calculation is important, especially compared to the daily turnover (the number of shares exchanged per day). For example, if the Active Float is set at 50 million shares but the turnover is only 100,000 shares per day, it will take 500 trading days to trade a number of shares that is identical to the Active

Price and Value

101

Float. As a rule of thumb, the Active Float should take more than 30 trading days but less than 90 trading days to turn over. These limits are set by experience. The 90-day limit implies that you will need stocks that trade a good part of their Active Float each day. Since 1 ÷ 90 = 1.1%, it means that the stock you select must see 1.1 percent of its Active Float exchanged every day. For the company Openwave Systems, you therefore need a minimum of 444,444 shares exchanged every day: 40 million ÷ 90 = 444,444 shares. The 30-day limit implies that you need an Active Float that is large enough to incur stock movement cycles at least every 30 trading days. By experience, the more quickly the Active Float turns over, the more difficult it will be to use this Active Float signal for trading. The reason is that the possible trend change must be confirmed by the Large Effective Volume flow. However, in very actively traded stocks, the activity of large players is partly masked by the traders’ activity, which makes the signals more difficult to read. When shares are trading very actively so that the turnover is very fast, this means one of two things: 1. The number of issued shares is low compared to the demand or the

supply. This means that the fast equilibrium moves between demand and supply will easily push the price up or down. A low number of issued shares is often the cause of high price volatility. 2. The stock is in the hands of day traders. For example, Finisar Corporation, a maker of fiber-optic systems for telecommunications networks, is trading its entire 308,000,000-share float in about 20 trading days, probably indicating a very active group of day traders. Once an Active Float is fixed, it must stay fixed for a very long period of time. It is only after one year, for example, that we adapt the Active Float to reflect, for instance, stock splits, stock dilution, and the like.

GRANDMOTHERS ARE ALWAYS RIGHT! Many expressions that traders use come from their long experience in the market. I call this the “grandmother experience,” because grandmothers have been right for centuries without always analyzing why. I will go through a few of these traditional grandmotherly sayings and try to explain them through the lens of the Active Boundaries indicator.

102

VALUE IN TIME

Stocks That Need to Rest Stock message boards are full of stocks that “need to rest”:

r “After a climb, a stock needs to rest before climbing again.” r “After a downslide, a stock needs to rest before continuing the decline.” These statements can easily be explained by the Active Boundaries behavior:

r After a climb, we reach the Upper Boundary, and the stock cannot move up again in price. The upward movement will continue if the Active Boundaries indicator comes down, which naturally occurs during a price trading range, since the indicator tends to move back to 0. During this phase, new buyers come in with higher expectations than shareholders who sold off. These new buyers expect the share price to continue to rise. As soon as the stock price rises again out of the trading range, this attracts still a new set of buyers who want to ride the trend until it again reaches the Upper Boundary. r After a down movement, we reach the Lower Boundary, and the stock cannot continue its down movement. The downward movement will continue if the Active Boundaries indicator comes up, which naturally occurs during a price trading range, since the indicator tends to move back to 0. During this phase, new buyers come in with higher expectations than shareholders who sold off. The Active Boundaries indicator increases until a new wave of selling starts. This new wave forces many recent new buyers to cut their losses before they are locked in with deeper losses. These new sellers fuel the price downtrend until we again reach the Lower Boundary. Let’s take the example of Chico’s FAS, Inc. Chico’s FAS operates 799 retail stores specializing in women’s clothing and accessories. In September 2006, the company had a market capitalization of $3.5 billion with net annual profits around $200 million. The company enjoyed healthy growth from late 2004 until early 2006, increasing the number of its stores. In late February 2006, the company announced that its gross margin would slightly decline over the year due to marketing expenses. In August 2006, the company reported its first negative same-store sales in more than a decade. This explains the price movements shown in Figure 2.17 and Figure 2.18a. Notice the very long uptrend (A) and the long downtrend (B).

Price and Value

103

FIGURE 2.17 Chico’s FAS: volume surge on price gap downward. Source: Chart courtesy of StockCharts.com.

You can see in Figure 2.18b that these two trends have been captured by a set of Active Boundaries: set A for the uptrend A and set B for the downtrend B. The strength of uptrend A equals 6.5 percent, which is the midpoint between the A Upper Boundary (23 percent) and the A Lower Boundary (−10 percent).

FIGURE 2.18a Chico’s FAS: price evolution. Trading ranges allow the pool of active shareholders to change. New shareholders come in with a higher expectation than departing shareholders with regard to the future share price increase.

104

VALUE IN TIME

FIGURE 2.18b Chico’s FAS: Active Boundaries. During a trading range, the Active Boundaries signal will trend back to the 0 percent equilibrium, making space for the next move to the Upper or Lower Boundary.

The strength of downtrend B equals −11 percent, which is the midpoint between the B Upper Boundary (−2 percent) and the B Lower Boundary (−20 percent). Note that downtrend B is much steeper than uptrend A, and is unsustainable. Hence, the Active Boundaries will soon need to be adapted to reflect a coming sideways trading or a price reversal. The purpose of this example is to study in Figure 2.18a and b two trading ranges, or periods of sideways trading: trading range A (TR-A), a threeand-a-half-month trading range from late November 2005 to early February 2006, and a two-month trading range B (TR-B) during May and June 2006. Trading range A (TR-A) happened after a very strong price surge that saw the price increase 50 percent, from $30 to $45. This very strong move brought the Active Boundaries signal close to the Upper Boundary, indicating that the traders’ expectation for a further price increase was at its lowest. At this position, you would expect a reversal to happen, with traders taking their profits. However, the duration of the trading range had as a consequence a turnover in the pool of active traders. Indeed, during that period, more than 100 million shares changed hands. Since the Active Float for Chico’s is set at 120 million shares, we may assume that the majority of the active shares changed hands and went into the hands of traders with a higher expectation for the price to continue to increase. We can see in Figure 2.18b the consequence of the trading range: the new traders brought the average profit down close to 0 percent (see point 1), the point from which the expectation is neutral and from which a new “leg-up” can start. A similar move appeared during trading range B (TR-B), in May and June 2006. After a gap down of about 20 percent, the price stabilized for

Price and Value

105

an entire month. During that period, new shareholders were looking for a bargain entry price into a stock that seemed much cheaper than before. These new shareholders slowly replaced old shareholders. This sideways movement alone brought the Active Boundaries signal close to the Upper Boundary, where expectation was very low, considering the fight against the downtrend. At the next negative news, the share price resumed its descent.

Panic Selling (the Falling Knife) I often read the following types of sayings on message boards of stocks that have experienced a very steep price decline: “This is panic selling; the stock will come back,” “It is the specialist getting cheap shares; do not fall into the trap,” and “I will not give my shares away.” These are emotional comments from traders who are usually locked in and are trying to justify their positions. When can we say that heavy selling is panic selling or is legitimate selling? Let’s have a look again at the three large selling movements on Figures 2.17 and 2.18a and b. These three movements (2, 4, and 6) show quick selling moves that stop at the Lower Boundary. The Active Boundaries signal does not extend past the Lower Boundary, meaning that professional traders are looking for value. This is legitimate selling, because these professionals interrupt the selling trend once they recognize value at the Lower Boundary. Reliant Energy (Figure 2.19a and b) shows us a good example of panic selling (sometimes called capitulation) that was quickly followed by a long new price uptrend. As you can see, the stock was experiencing a

FIGURE 2.19a Reliant Energy: price evolution in a panic selling move.

106

VALUE IN TIME

FIGURE 2.19b Reliant Energy: Active Boundaries in a panic selling move.

healthy uptrend (A) until the price reversed down to the Lower Boundary, which was already low at −14 percent. From that point, the stock gapped down to −30 percent, which is unsustainable, except if investors believe that the company will soon be bankrupt. The professional investors quickly recognized the price bargain and pushed the stock into a new uptrend (B). Some people have asked me if it is worth it to buy stocks when there is heavy selling. Grandmothers tell you “not to catch a falling knife.” The decision to buy into heavy selling depends on two factors: who is trading the stock and what the level of the value zone you are entering into is. Of course, if the company’s existence is in doubt, do not step in. You may consider buying stocks that fall heavily through their Lower Boundary and reach average profit levels that are historically at their lowest, but do so only if the following three conditions are met: 1. The ratio of institutional investors to the total number of sharehold-

ers should be greater than 1:2, because institutional investors have a good perception of the intrinsic value of the company and will heavily move into the stock once they notice that the price is not reflecting the company value. 2. The drop must have locked in so many traders that the supply of shares will have gone down to less than 10 percent of the total float (we will see in Chapter 4 how to measure the change of supply of shares). 3. Large Effective Volume must clearly indicate that large players are moving in.

Price and Value

107

The Dead Cat Bounce Have you ever heard of the dead cat bounce? I will not explain the origin of this saying, but in trading terms, it means that if a long downtrend is being reversed, some traders will announce that “it is just a dead cat bounce.” What they mean is that the future for the company is bleak, and the reversal in price will be short-lived. Bargain hunters are the ones who triggered the bounce, but the selling pressure on the price will continue and the downtrend will regain its momentum. If you look again at Figures 2.17 and 2.18a and b, you can see that there are three gaps down in price (2, 4, and 6) that are preceded by three small bounces (1, 3, and 5). The first price bounce (1) does not qualify as a dead cat bounce since the downtrend is not yet established. At that point, we may think that the market is going into a new up-leg. The second price bounce (3) sends us to the newly redefined Upper Boundary, where the price could reverse down. This bounce can be named a dead cat bounce since it is a false upside reversal that will fail and prices will continue to fall further down to the Lower Boundary (4). The third price bounce (5) also copies the pattern of a dead cat bounce since it is followed by a lower low in price. You may note here that the last gap down (6) is so strong both in price and in volume (see the very high volume bar in the lower part of Figure 2.17) that it almost completely exhausted the supply of shares to less than 1 percent of the total float.

Exuberance It is normal for a group of traders to change their collective expectation on a stock. For example, after stellar earnings and a few upgrades, the collective expectation will usually change. The stock price will gap up and new Upper and Lower Boundaries will form. However, a sudden price increase could happen as a result of gradeB news, an article in a popular magazine, or simply pure rumor. During extreme price increases, retail traders are overwhelmed by joy and congratulate each other on stock message boards. By experience I know that exuberance attracts correction and that correction on exuberance is usually violent. I therefore always find it more useful to measure the price increase using the Active Boundaries signal than by comparing the price to past price levels. Active Boundaries are very good for discovering points of traders’ exuberance. Stocks that are heavily followed by retail players could escape common sense. An example of such a move is shown in Figure 2.20a and b. Envoy Communications Group, a Canadian company listed on the NASDAQ, specializes in retail branding in Europe and North America. Much could be

108

FIGURE 2.20a Envoy

VALUE IN TIME

Communications

Group:

price

evolution

during

exuberance.

written about this mismanaged company, whose share price lost 80 percent of its value between February 2004 and December 2005 (Figure 2.20a). Let’s concentrate on the technical side. We can see in Figure 2.20b that the Active Boundaries captured the downtrend very well. The spread between the Upper and the Lower Boundaries was about 40 percent, giving good trading opportunities. The downtrend was very strong (−8 percent). You can notice two points of exuberance, where the stock price increased so much that the Active Boundaries signal passed over the Upper Boundary, indicating that traders did not expect the price to go much higher. You will notice that the second exuberance run (point 8) was

FIGURE 2.20b Envoy Communications Group: Active Boundaries during exuberance. Exuberance is mainly found in stocks that are traded by retail traders. Individuals indeed have no good reference regarding stock value, and can find themselves quickly transfixed in their hopes for even higher gains.

109

Price and Value

extraordinarily strong, pushing to an average profit of 50 percent. This is impossible to sustain; hence the following price decrease. This is typical of stocks that are mainly traded by retail players who do not have a good grasp of how to ascertain value.

FOR MATH LOVERS: HOW TO CALCULATE THE ACTIVE BOUNDARIES Let us consider that the total Active Float of a given stock is set (for example) at 80 million shares. The Active Boundaries indicator is simply the average profit/loss of the last 80 million shares that have been traded. Since the minute-by-minute data stream tallies all the stock transactions that took place during one minute, for each of these transactions we can evaluate the profit/loss at any given time. Indeed, we can calculate at every trading minute the average profit/loss on the last fixed number of shares that were traded. If there is a 1 percent price jump from one minute to the next, for example, the average profit of all shares that are part of the Active Float will be largely affected—if the average profit was 5 percent, it will jump to 6 percent. However, if during that new trading minute only 1 percent of the Active Float was traded, this amount of shares will have very little influence on the average profit/loss of the total number of active shareholders. If the share price is stable within a trading range, this indicator will trend back to 0 percent. For example, suppose that today the price is $4 with an average profit of −10 percent. This means that on average the shares traded at $4.44 ($4.44 − $0.44 = $4.00). Now, suppose that there is a news story that moves the price to $6 overnight. That will bring the average profit to +35 percent for the 80 million shares: ($6 − $4.44) ÷ $4.44 = 35% However, if you have 10 million shares that change hands at $6, the average profit for the last 80 million shares will come down as indicated by the following formula: New profit (80 million shares) = 35% × 7/8 + 0% × 1/8 = 30.6% The last purchased 10 million shares have a profit of 0 percent, which means that the average profit will decrease from 35 percent to 30.6 percent, even if the price stays at $6.

110

VALUE IN TIME

Average Profit on a Company’s Shares Bought at Different Prices If an investor buys shares of the same company X over two different periods of time, that investor makes buying/selling decisions about X based on the average profit on X. The procedure for average profit calculation is to first take the average purchasing price and then calculate the profit. For example, if an investor bought 100 shares at $3 and 1,000 shares at $4, and if the current share price is $5, the average profit is calculated as: Step 1: Average buying price = [(100 × $3) + (1,000 × $4)] ÷ 1,100 = $3.909 Step 2: Average profit = ($5 − $3.909) ÷ $3.909 = 27.9% However, traders make their buy/sell decisions independently of other traders. Therefore, there is no meaning in calculating an average price of a portfolio of all the traders together. What we need to calculate is the profit for each share and then calculate the average profit per share of the company X. In this case, that gives: Step 1: Profit on 100 shares bought at $3: ($5 − $3) ÷ $3 = 66.7% Step 2: Profit on 1,000 shares bought at $4: ($5 − $4) ÷ $4 = 25% Step 3: Average profit: (66.7% × 100/1,100) + (25% × 1,000/1,100) = 28.8% This average profit reflects the sum of all the individual decisions based on each shareholder’s own profit, weighted by the number of shares owned by these shareholders. In this case, we suppose that each shareholder bought shares at one specific time, and did not average up or down. Because this is not always the case, the correct average profit is situated between the two methods of calculation shown above. Both these methods give similar results in terms of Active Boundaries definition.

WHAT WE LEARNED REGARDING ACTIVE BOUNDARIES Three hypotheses were made to define the Active Boundaries indicator:

1. The pool of active traders is stable. 2. Each single trader does not change his trading strategy or personality

overnight.

Price and Value

111

3. Program trading requires a long development time, with very slow ad-

justment cycles compared to market swings. The Active Boundaries tool indicates when a stock is cheap relative to traders’ expectations.

r It is a slow-moving oscillator on volume change. r It shows fast-moving variations due to price fluctuations. It is easy to say that we should “buy low and sell high,” and it is tempting to use the Active Boundaries as buy and sell indicators. However, we saw that, when following the Active Boundaries indicator,

r Position traders will only expect profits that are not larger than the distance between the Upper Boundary and the Lower Boundary, missing out on long-term trends. r Long-term investors will have to go through pullbacks in order to get the most out of the uptrend. The Active Boundaries indicator will not tell the long-term investor if the pullback is temporary or the start of a new downtrend. The Active Boundaries indicator will simply indicate whether, for all the traders, the share price is getting cheaper or more expensive. Here are some more remarks on position trading using only the Active Boundaries indicator:

r Active Boundaries produce cheap and expensive signals that are not as intuitive as the price level can be. We will sometimes have to buy a stock at a higher price than our previous selling price. r When we sell at the first expensive signal, we could end up either missing the next run or rushing to buy the stock again in the next run, because we have the feeling that we are missing something. The Active Boundaries indicator will not help the emotionally weak trader. Since any trader is at some point emotionally weak, it is unwise to trade on the basis of the Active Boundaries indicator only. r It is rather easy for me to say that we should buy at the Lower Boundary. However, what tells us that the price will not go lower than the Lower Boundary? Indeed, all the long uptrends have to come to an end. Likewise, long downtrends also have to end. This means that at some point, we will break through the Lower or the Upper Boundary. There is no way, using the Active Boundaries signal, to know whether we will rebound or go through. Different signals are necessary: Effective Volume, the Effective Ratio (see Chapter 3), and divergence analysis.

112

VALUE IN TIME

r It is more difficult to earn money by betting against the price trend. r Price targets are set by the calculation of the distance to the Upper Boundary or Lower Boundary, depending on whether you are long or short. Active Boundaries capture trends very well; to my mind, they provide the only rational method that explains why trends exist. The Active Boundaries tool is excellent for monitoring trends.

CHAPTER 3

When Volume Diverges from Price

n Chapter 1, “Effective Volume: An Open Window into the Market,” we learned how to detect when trendsetters and large players are coming into or getting out of a stock. Effective Volume is indeed very useful for detecting if a price trading range is going to break to the upside or to the downside. In Chapter 2, “Price and Value: The Active Boundaries Indicator,” we saw how to detect the market value of a stock. We also saw how to capture and monitor trends. We finally understood that both methods are complementary and that each contributes in its own way to helping you make sound trading decisions. In this chapter, we are going to measure the balance between the buying and the selling forces. This will enable us to point out levels above which the change of equilibrium will have a high probability of impacting the price. We will therefore be able to make decisions before price changes occur. This is what this whole chapter is about: the introduction of a new tool for monitoring trends that is complementary to the Active Boundaries tool. Indeed, long-term trends are often interrupted by small pullbacks before resuming and going on to new highs. In a pullback, you may be tempted to sell in order to protect your profits or to buy the stock because you believe that it will regain its upward momentum. By performing an analysis of relative strength between Effective Volume and price, you can get a very good image of the strength that underlies the price movements. Most of the time, this analysis will tell you what you need to do. Since I developed the divergence analysis tool, I have greatly relied on it to separate true

I

113

114

VALUE IN TIME

from false Effective Volume signals. The math is a little more complicated than what I used in previous chapters, but you will see that the results are amazing. In this chapter, you will learn:

r The importance of measuring the demand/supply equilibrium in order to assess future price movements, and how to measure this equilibrium. r How to separate false from correct buy/sell signals. r The dynamic of a market that involves both large funds and retail players. r The importance of volatility adjustments to produce the right signals. If you are a fund manager, one ancillary benefit to this chapter is that the Effective Ratio tool presented will allow you to determine the number of shares that you can possibly accumulate/distribute per day without having an impact on the price.

EFFECTIVE VOLUME: TWO ARROWS FROM ONE BOW Before starting with the Divergence Analysis, we need to come back to the Effective Volume concept. Remember that the Effective Volume is defined as the volume that is responsible for a small price change from one trading minute to the next. In Chapter 1, I called this price change a price inflection. I also showed in Chapter 1 that we can separate the Large Effective Volume from the Small Effective Volume, which is helpful in detecting trendsetters. The Effective Volume analysis therefore allows us to shoot two arrows with the same bow: We can look at the general long-term trend of the total Effective Volume, or we can look at only the evolution of the Large Effective Volume. Let’s first examine how to use the general trend of the total Effective Volume such as that seen in Figure 3.1, which represents the Effective Volume flow trend and the price trend for Meridian Resource Corporation. In Figure 3.1, I have identified four segments: 1. Segment A: The price is in a downtrend. This downtrend is confirmed

by the downtrend in total Effective Volume flow. 2. Segment B: The price is in a trading range. This range is confirmed by

the flat total Effective Volume flow.

When Volume Diverges from Price

115

FIGURE 3.1 Meridian Resource, Effective Volume and price evolution. 3. Segment C: The price is in an uptrend. That uptrend is confirmed by

the uptrend in total Effective Volume flow. However, the total Effective Volume flow uptrend is much steeper than the price uptrend. 4. Segment D: The price uptrend continues, but the total Effective Volume flow is in a downtrend. What can we conclude from this example?

r In trend C: If detected early, the difference in steepness between the Large Effective Volume flow trend and the price trend can give us a good warning of what is happening. Conclusion: We need to compare trend strength. r In trend D: A pure divergence between the Large Effective Volume flow trend and the price trend is important enough to get our attention. Conclusion: We need to compare trend direction.

116

VALUE IN TIME

What is not shown in Figure 3.1 is that we also have to know if the difference in trend strength between the Effective Volume flow trend and the price trend is strong enough by historical standards. If you remember, one of the hypotheses of the Active Boundaries theory is that the same traders and the same funds will play the same stock over and over again. The Active Boundaries pattern is formed because, all things being equal from one situation to the next, traders make the same decisions. I believe that this is also true when analyzing the price/volume balance. Indeed, the price will move up or down depending on how the demand/supply of shares is evolving. The role of the Effective Volume method is precisely to measure the level of accumulation and distribution of shares. If we look at past Effective Volume accumulation/distribution patterns and compare them to price patterns, we may eventually be able to say:

r The present accumulation/distribution trend is much stronger than past accumulation/distribution trends.

r In the past, when the accumulation/distribution reached such strength, the price started to increase/decrease.

r Therefore, we may conclude that the probability that the same price movements will occur in the near future is high. The importance of measuring the strength of a trend clearly appears in Figures 3.2 and 3.3, which represent the Effective Volume analysis for the natural gas producer Chesapeake Energy (CHK) during 10 days and 40 days, respectively. In Figure 3.2, we can see that the price trend is flat, but that the Large Effective Volume is trending up, indicating that the trading range will probably break to the upside. Figure 3.3, however, shows a different picture: The last 10 days of trading (trading range B) are showing something that looks like a pause in the uptrend A that occurred prior to the trading range. This means that we actually do not know at this point if trading range B will break to the upside. The Large Effective Volume uptrend B, since it is much weaker than uptrend A, does not seem to be strong enough to move the price up. If you look more closely at Figure 3.2, you will see that the accumulation on Effective Volume is showing that 800,000 shares were accumulated during the last 10 trading days. Is this significant or not? During the same period, a total of 68 million shares have been exchanged. The Effective Volume accumulation is just about 1.2 percent of that total. If you now compare the accumulation that took place during uptrend A in Figure 3.3, you can see that four million shares were accumulated over 40 days, when a total of 150 million shares were traded. This is approximately 2.7 percent. Is 2.7 percent significant? It is more than 1.2 percent, but does this represent significant accumulation?

When Volume Diverges from Price

117

FIGURE 3.2 Chesapeake Energy: 10-day Effective Volume analysis.

I would like to remind you of something important when you look at the upper panel of Figure 3.3. What you see is not just a graph that trends up and then goes flat. It really represents the buying and selling by large players. Do you notice that in the upper panel of Figure 3.3, the large players’ uptrend A starts a few days before the price starts its move up? What happened was that the fund simply started to accumulate stocks and changed the demand/supply balance that was characteristic of the trading range preceding uptrend A. In order to attract more sellers, the buyers had to increase the stock price. This is not insider trading. It is just funds buying a cheap stock. To see if this type of accumulation is abnormally significant, we need to compare it to past accumulations that occurred during past price trends. The historical analysis of the past behavior of large funds will give us a measure of what is a normal accumulation and what is not. We now understand that we need a tool that compares the Effective Volume trend strength to the price trend strength.

118

VALUE IN TIME

FIGURE 3.3 Chesapeake Energy: 40-day Effective Volume analysis. The Effective Volume analysis will often give different results depending on the time frame used. It is preferable to use longer time frames. This allows us to see if the short-term trend is significant compared to the long-term one.

PRICE AND EFFECTIVE VOLUME TRENDS Before starting with the comparative analysis of price and Effective Volume trends, we first need to better understand how each of them moves individually.

Price Trend Let’s work on the concrete case of Darden Restaurants Inc. (DRI). Capitalized at $9 billion, Darden Restaurants is a restaurant operator in the United States and Canada. The company operates restaurants under the names Red Lobster, Olive Garden, Bahama Breeze, and others.

When Volume Diverges from Price

119

Figure 3.4 shows both the company share price (upper panel), as well as the price change rate for the past 60 days (lower panel). Please note that from now on I will call the price change rate the “price rate of change” since the terminology “rate of change” (ROC) is well known in traditional technical analysis. To understand if a trend is increasing, you need to take two points in the trend and calculate their difference in terms of percentage. For example, if we take point A (which shows a price of $34.98) and point B (which shows a price of $35.58), the difference between them in terms of percentage is 1.7 percent. This is represented on the lower panel by point C. Then take two other points, A′ and B′ , whose distance is identical to the distance separating A and B, and calculate their difference as a percentage. The result is 9 percent and is shown as point C′ on the lower panel of Figure 3.4. Obviously, 9 percent is higher than 1.7 percent. You can therefore deduce that at point C′ , the trend slope is more significant than at

FIGURE 3.4 Darden Restaurants: price compared to the price rate of change.

120

VALUE IN TIME

point C. Indeed, you can see that the A′ B′ slope is much steeper than the AB slope. The distance between point A and point B is called the analysis window. In this case, I took 2,000 trading minutes as the analysis window, which is about 5.3 trading days. If you move that analysis window from left to right and calculate for each trading minute the difference in price between the beginning of the analysis window and the end of the window, you get the graph represented in the lower panel of Figure 3.4. As you may have noticed, there is a price gap between points A′ and ′ B . When we calculate the price rate of change, this price gap will influence the calculations twice: when it enters into the analysis window (when B′ reaches it) and when it exits the analysis window (when A′ reaches it). This is clearly shown by the double gaps that appear on the lower panel of Figure 3.4, which I have labeled “entry point of price gap into the analysis window” and “exit point of price gap from the analysis window.” We will see later how to handle such gaps. One advantage of the price rate of change is that it usually changes its trend before the price signal itself. Measuring the rate of change of the price of a stock is similar to measuring the change of altitude of a climbing rocket. Assume that the rocket climbs from 1,000 feet to 2,000 feet in x seconds, and then from 2,000 to 2,500 feet during the next x seconds. You can conclude that since its climbing altitude is reduced within the same period of time (called the climbing speed), the rocket is in trouble and may eventually come back to earth. The rate of change measures this change in climbing speed. For the rocket, this rate of change is therefore declining before the rocket itself starts falling down. In Figure 3.5, I have depicted an exponential average function that I applied on the lower panel of Figure 3.4 in order to smooth it. The exponential average is a standard formula that smoothes curves. Compared to a straight average, the exponential average gives more significance to recent data than to older data. For the exponential average I usually use the same number of days that have been used to define the analysis window. As we will see later, the length of the analysis window itself is adjusted to the difference in volatility between volume and price. The advantage of working with a smoothed curve is that with it we can more easily detect trend changes. The main inconvenience is that the smoothed signal is delayed compared to the original signal. However, this delay is often compensated by the fact that the price rate of change is moving ahead of the original price signal. The net effect is therefore usually neutral. Price averaging (50- or 200-day moving average) and the calculation of the price rate of change form the basic elements of price-based traditional technical analysis tools.

When Volume Diverges from Price

121

FIGURE 3.5 Darden Restaurants: price rate of change smoothed with an exponential average function.

Effective Volume Trend The same principle can be applied to the analysis of the rate of change of the Effective Volume flow. Let’s have a look at Figure 3.6, which represents the total Effective Volume flow of Darden Restaurants for the past 60 days, and compare it to Figure 3.7, which separates the total Effective Volume flow into the Large Effective Volume flow and the Small Effective Volume flow. The main noticeable element of the two figures is that the total Effective Volume flow in Figure 3.6 is almost a copy of the Large Effective Volume flow in Figure 3.7. Now, if we look at Figure 3.8, we instantly see that the A, B, C, and D price trends do not always follow the corresponding Large Effective Volume flow trends of Figure 3.7. We can indeed see that trends C and

FIGURE 3.6 Darden Restaurants: total Effective Volume flow.

122

VALUE IN TIME

FIGURE 3.7 Darden Restaurants: Effective Volume flow separated by size.

D of both figures diverge. The following explanation is possible: At the end of price trend B the price was high enough. Large players ceased being net buyers (trend C in Large Effective Volume is flat), and the price fell down (trend C in price) on its own weight due to lack of buying. Then, large players started buying again (trend D in Large Effective Volume is increasing again), which allowed the interruption of the price reversal and the formation of the price trading range D. As a reminder of Chapter 1, in which I revealed the procedure for calculating the Large and the Small Effective Volume flow, we can see in Table 3.1 that the Effective Volume is the volume responsible for price inflections. I have connected consecutive lines with either gray or black arrows. The gray arrows represent positive price inflections, while the black arrows represent negative price inflections. The last two columns of Table 3.1 show that the Effective Volume is then separated into Large and Small Effective Volume.

FIGURE 3.8 Darden Restaurants: stock price.

123

13:54 13:53 13:52 13:51 13:50 13:49 13:48 13:47 13:46 13:45 13:44 13:43 13:42 13:41 13:40 13:39 13:38 13:37 13:36 13:35

11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06 11/16/06

41.07 41.03 41.03 41.02 41.02 41.03 41.03 41.02 41.03 41.03 41.02 41.03 41.02 41.03 41.04 41.03 41.02 41.02 41.04 41.04

Open

41.08 41.08 41.03 41.03 41.02 41.03 41.03 41.03 41.03 41.03 41.03 41.03 41.03 41.04 41.04 41.04 41.03 41.04 41.04 41.04

High

41.06 41.03 41.02 41.02 41.02 41.02 41.02 41.02 41.02 41.02 41.02 41.01 41.01 41.02 41.02 41.03 41.01 41.02 41.02 41.04

Low

41.06 41.07 41.02 41.03 41.02 41.02 41.03 41.03 41.02 41.02 41.02 41.01 41.02 41.03 41.04 41.03 41.03 41.03 41.02 41.04

Close

Price Inflections

Price Inflections and Effective Volume Effective Volume

–1,133 –5,057 –14,800 2,500 0 –2,100 0 4,900 0 0 3,533 –400 − 3,133 − 3,733 2,533 0 0 1,267 −1,300 0

Volume

1,700 5,900 14,800 2,500 700 2,100 4,900 4,900 11,000 800 5,300 600 4,700 5,600 3,800 1,900 1,700 1,900 1,300 500

0 −5,057 −14,800 0 0 0 0 4,900 0 0 3,533 0 −3,133 −3,733 0 0 0 0 0 0

Large Effective Volume

−1,133 0 0 2,500 0 −2,100 0 0 0 0 0 −400 0 0 2,533 0 0 1,267 −1,300 0

Small Effective Volume

Price inflections are the small price changes between one trading minute and the next. The Effective Volume is the volume that is responsible for these price inflections.

Time

Date

TABLE 3.1

124

VALUE IN TIME

TABLE 3.2

Darden Restaurants: 60 Days, Repartition between Effective and Non-Effective Volume

Total Volume Effective Volume Non-Effective Volume

Shares

Percentage

80,529,000 49,996,040 30,532,959

100.0% 62.1% 37.9%

In Table 3.2, we can see that for Darden Restaurants, the total Effective Volume calculated during the past 60 trading days amounted to 62 percent of the total exchanged volume during that period. We can also see in Table 3.3 that the 49,996,040 shares that constitute the total number of Effective Volume shares have been about evenly distributed between Large and Small Effective Volume. As a matter of fact, the separation volume between the large and the small size has been selected so that both groups have about the same number of shares. This means that, in theory, each group has the same power to move the share price. However, the last line of Table 3.4 shows that small players were involved in 80 percent of the price inflections that occurred during the 60 trading days. Furthermore, Table 3.5 shows that the price inflections in which small players were involved provoked price changes of 30,120 cents (a total of $301.20), which is more than twice the number of cents changes generated by large players (13,423 cents, or $134.23). Let’s now compare this ratio to the results of Table 3.6, which shows that the average level of price changes between consecutive trading minutes is 3.91 cents for price inflections that are due to large players, and only 2.19 cents for price inflections due to small players. In military terms, you could say that small players have much stronger firepower than large players, but those large players have a stronger impact on each price change because they shoot bigger bullets. What is really key is not just the firepower itself, but more importantly the “aim to target.” Let’s have a look at Tables 3.7 through 3.10. We can first see from Table 3.7 that even if small players exchanged more than 24

TABLE 3.3

Darden Restaurants: 60 Days, Repartition between Large and Small Effective Volume

Total Effective Volume Large Effective Volume Small Effective Volume

Shares

Percentage

49,996,040 25,365,552 24,630,488

100.0% 50.7% 49.3%

125

When Volume Diverges from Price

TABLE 3.4

Darden Restaurants: 60 Days, Repartition of Price Inflections between large and small players

Total number of price inflections Price inflections due to large players Price inflections due to small players

TABLE 3.5

Price Inflections

Percentage

17,186 3,433 13,753

100.0% 20.0% 80.0%

Darden Restaurants: 60 Days, Repartition of Cents Changes during Price Inflections between Large and Small Players

Total number of cents changes Cents changes due to large players Cents changes due to small players

TABLE 3.6

Cents

Percentage

43,543 13,423 30,120

100.0% 30.8% 69.2%

Darden Restaurants: 60 Days, Average Number of Cents Changes by Price Inflection for Large and Small Players

Average Cents Change per Inflection

Cents

Due to large players Due to small players

3.91 2.19

TABLE 3.7

Small Small Small Small

Darden Restaurants: 60 Days, Separation of Small Effective Volume Shares into Positive and Negative Price Inflections

Effective Volume positive volume negative volume net volume

Shares

Percentage

24,630,488 12,362,157 12,268,331 93,825

100.0% 50.2% 49.8% 0.4%

million effective shares during the 60 trading days, the direction of these shares was not clear. Indeed, about 12 million effective shares were linked to negative price inflections, while about the same number were linked to positive price inflections. Small players are therefore directionless! This is usually the case when the stock is mainly traded by large funds.

126

VALUE IN TIME

TABLE 3.8

Large Large Large Large

Darden Restaurants: 60 Days, Separation of Large Effective Volume Shares into Positive and Negative Price Inflections

Effective Volume positive volume negative volume net volume

TABLE 3.9

Shares

Percentage

25,365,552 14,310,963 11,054,589 3,256,373

100.0% 56.4% 43.6% 12.8%

Darden Restaurants: 60 Days, Separation of Small Effective Volume Shares into Number of Positive and Negative Cents Changes

Total cents change due to small players Positive cents change Negative cents change Net cents change

Cents

Percentage

30,120 14,670 15,450 −780

100.0% 48.7% 51.3% −2.6%

Darden Restaurants: 60 Days, Separation of Large

TABLE 3.10 Effective Volume Shares into Number of Positive and Negative Cents Changes

Total cents change due to large players Positive cents change Negative cents change Net cents change

Cents

Percentage

13,423 7,403 6,020 1,383

100.0% 55.2% 44.8% 10.3%

By contrast, you can see from Table 3.8 that large players were net buyers for 3,256,373 shares, or more than 12.8 percent of the total volume exchanged by these large players. This is a clear picture of the “aim to target” or intent of large players. This difference between large and small players is also very apparent when you compare Tables 3.9 and 3.10. Table 3.9 shows that the 24 million shares traded by small players had a net price influence of −780 cents. However, Table 3.10 shows that the 25 million shares exchanged by large players resulted in moving the price up by 1,383 cents. In short, small players had four times as many chances to move the price up and down, because they were involved in four times as many

127

When Volume Diverges from Price

price inflections as large players (Table 3.4). However, during the 60 trading days, the stock price was up about $6. Large players were net positive price movers for $13.83, while small players were net negative price movers for −$7.80. We can conclude from this example that large players are the ones who really move the markets, while small players are mainly making noise, statistically speaking. What should be clear by now is that when large funds get involved in a specific stock, they may easily influence the price direction. I would not say that there is rampant manipulation, but when you know exactly the level of shares accumulation you can achieve without moving the price, you also know how to move the price up after you have finished accumulating.

Effective Ratio Before proceeding in the comparison between price and Effective Volume trends, let’s note that it is not entirely true that the Effective Volume is the only force responsible for price changes. Indeed, Effective Volume is by definition the calculation of the number of shares that is responsible for price inflections. In other words, Effective Volume measures the number of shares that dynamically pushes the price up or down. These dynamic movements are important, because they express a strong will of traders. What the Effective Volume has great difficulties measuring is the static supply of shares: It is the entire set of limit orders that slows down the dynamic movements of the strong-will traders. These limit orders are buy or sell orders entered in the order book at prices that are different from the market price. When the market price reaches these limit orders, their execution simply slows down the main direction movement; it then slows the momentum of strong-will traders. I show in Chapter 4 how to deal with the measure of the supply of shares. However, for the purpose of our divergence analysis, we first need to adjust the Effective Volume to include the influence of the static players. If we study a fixed trading period that I call the analysis window, we understand that the active buying pressure on a stock during that specific period is the difference between the Effective Volume flow at the end of the period and the Effective Volume flow at the beginning of the period. This is due to the fact that the Effective Volume flow has been built by adding all the Effective Volume responsible for positive price inflections and subtracting all the Effective Volume responsible for negative price inflections. Force of active players = Positive Effective Volume − Negative EffectiveVolume

128

VALUE IN TIME

This force is what fuels or may change a price trend. As we have seen before, we also expect that this force will be slowed down if the number of passive traders increases. Therefore, we need to weight that force by the sum of the number of dynamic players and passive players, which is simply the total number of shares exchanged during the analysis period. I therefore define the Effective Ratio of a specific analysis window as: Effective Ratio =

Force of active players (during the analysis window) Total number of shares (during the analysis window)

For example, let’s suppose that for the past 30 minutes of trading, the force of active players was a positive number of 50,000 shares (meaning that there were 50,000 more shares that moved the price up than there were shares that moved the price down). If we also suppose that during the same 30-minute period the total exchanged volume was 500,000 shares, the real buying force would be 10 percent of the total volume (50,000 shares divided by 500,000 shares). However, if during the following 30 minutes, the force of active players was still 50,000 shares but the total volume was only 250,000 shares, it would indicate a 20 percent total relative buying force during that period (50,000 shares divided by 250,000 shares). The reason for this increased buying force is the relative scarcity of shares available for sale during this second period. In other words, the Effective Ratio is the rate of change of the proportion of Effective Volume flow to the total number of shares sold during a specific trading period. For the purpose of divergence analysis, the reason I prefer to use the Effective Ratio instead of the Effective Volume is that when funds need to accumulate a large number of shares, the fund managers modulate their buying tactics according to the level of supply that comes into the market. If the supply of shares is very low, the fund managers will buy fewer shares; otherwise the price would be pressured up. Figure 3.9 (a, b, and c) shows for Darden Restaurants the different stages of calculation, from the Large Effective Volume flow to the smoothed Large Effective Ratio. A few comments on these different figures:

r The upper panel of Figure 3.9a shows the Large Effective Volume flow. r The lower panel of Figure 3.9a shows the rate of change of the Large Effective Volume. It is obtained by subtracting (on the first panel) the Large Effective Volume flow at the beginning of the analysis window from the Large Effective Volume flow at the end of the analysis window: C = B − A and C′ = B′ − A′ .

When Volume Diverges from Price

129

FIGURE 3.9a DRI: calculation of the rate of change (lower panel) of the Large Effective Volume flow (upper panel).

r Figure 3.9b shows the Large Effective Ratio. It is obtained by dividing the result from the second panel of Figure 3.9a by the total number of shares exchanged during the analysis period. As a matter of fact, this is the real measure of the balance between large buyers and large sellers in percentage of the total number of shares. r Figure 3.9c is obtained by smoothing Figure 3.9b with an exponential moving average. How to Use the Effective Ratio The Effective Ratio is a measure of the buying/selling pressure during a fixed period of time called the analysis window. We can calculate either the Large Effective Ratio, which corresponds to only the large players, or the total Effective Ratio, which corresponds to both small and large players.

130

VALUE IN TIME

FIGURE 3.9b DRI: Large Effective Ratio (a measure of the buying/selling strength among large players).

I use the Large Effective Ratio to discover entry points and the total Effective Ratio to find exit points. Indeed, large players are usually trendsetters, and large players’ accumulation usually indicates a possible future price move. However, I prefer to see the movements of all the players in order to find exit points. Indeed, selling waves could be triggered by some negative news, but they mainly come from the fact that a stock is overpriced, which triggers profit taking by both large and small players. It is when the majority of all the shareholders start selling that the price moves down. The Large Effective Ratio can be used in stand-alone mode, to find out if the current accumulation/distribution by large players is strong

FIGURE 3.9c DRI: Large Effective Ratio smoothed by an exponential moving average function. This set of figures (a, b, and c) shows the steps in calculating the buying/selling force.

When Volume Diverges from Price

131

FIGURE 3.10a Tellabs: Effective Volume analysis.

compared to past accumulation/distribution (Chapter 6 uses this tool in combination with the Active Boundaries tool to form a trading strategy). Figure 3.10 (a, b, and c) shows such a comparison. We can see in Figure 3.10a that for the company Tellabs, large players accumulated shares in two phases: A and B. Phase A (shown by arrow A) was rather strong; As a consequence, the Large Effective Ratio crossed over the average of the peaks of past Large Effective Ratio trends, shown by the dotted horizontal line of 4.1 percent in Figure 3.10b. However, we can see at the right of the same Figure 3.10b that large players are becoming increasingly weaker, because of the weak B trend of Figure 3.10a. As a reference, Figure 3.10c shows the stock price.

FIGURE 3.10b Tellabs: Large Effective Ratio on a 3.3-day period.

132

VALUE IN TIME

FIGURE 3.10c Tellabs: price evolution.

Comparing the Effective Ratio to the Chaikin Money Flow Indicator It is interesting to make a direct comparison between the Effective Ratio indicator and standard indicators using end-of-day data, such as the Chaikin money flow oscillator. The Chaikin money flow oscillator is calculated by adding the daily readings of the accumulation/distribution signal during a period of, for example, 21 days, and then dividing this number by the total volume exchanged during that period. The daily accumulation/distribution signal is based on the hypothesis that the buying/selling pressure on a stock is well measured by the location of the close relative to the high and low for the day. This accumulation/distribution signal simply weights the daily volume by the spread between the close and the low prices divided by the spread between the high and low prices of the day. In Figure 3.11, I analyzed the evolution of the Effective Ratio of Tellabs for a 90-day period that included two large price gaps (A and B). These price gaps were caused by positive and negative news. We can see that prior to the positive news at point A in the upper panel of Figure 3.11, the Effective Ratio had been signaling a strong accumulation (lower panel). Also, prior to the negative news at point B, the Effective Ratio had been signaling a strong distribution. Dashed lines labeled “strong accumulation limit” and “strong distribution limit” are, respectively, the average of the peaks and the average of the troughs of the Effective Ratio signal for the past (for example, for the prior six months to two years). When the accumulation/distribution exceeds the limit, we may consider such accumulation/distribution as stronger than in the past. As a comparison, the standard Chaikin money flow indicator shown in Figure 3.12 gave a positive but decreasing money flow indication leading to point A (not specifically inviting us to buy). From point A, the Chaikin

When Volume Diverges from Price

133

FIGURE 3.11 Tellabs: Large Effective Ratio.

money flow indicator was showing a negative reading, indicating distribution of shares, while the Large Effective Ratio indicator (Figure 3.11) was continuously above the 0 percent limit, indicating accumulation by large players. At point B, however, while the Large Effective Ratio was urging us to sell, the Chaikin money flow indicator showed a rather strong positive accumulation reading, urging us instead to buy. Why these differences? There are three reasons: 1. The Chaikin money flow indicator usually uses a wider analysis win-

dow, such as 21 days, while the Effective Ratio uses a shorter analysis window of about three to five days. This is due to the fact that the Effective Ratio tries to catch insiders’ and funds’ moves. We know that insiders get the news only a few days before its release to the public. Therefore, an analysis on a long period would not give a good signal on the movements of the critical most recent trading days.

134

VALUE IN TIME

FIGURE 3.12 Tellabs: the Chaikin money flow indicator. Source: Chart courtesy of StockCharts.com.

2. The Chaikin money flow indicator, like many other indicators, is based

on the closing price. We have seen in Chapter 1 that the close of the day is heavily traded, so that large funds with deep pockets could eventually tilt the close value in their favor. 3. You can see in the lower panel of Figure 3.11 that the buy limit is touched when the Large Effective Ratio crosses over the 4.1 percent line. This means that the imbalance between large buyers and large sellers is very small: Its average maximum compared to the total number of shares exchanged is only 4.1 percent. This type of small imbalance is well below the measurement error rate of end-of-day tools such as the Chaikin money flow indicator and can therefore be detected only by using more precise tools.

PRICE-VOLUME DIVERGENCE ANALYSIS The next step is to compare price changes to volume changes. What do we want to get out of that comparison? We want to have an early warning of what is going on behind the scenes. We already saw that large buying by funds at the bottom of a downtrend has the power to change the price trend direction. Therefore, the divergence analysis between the Large Effective Ratio and price helps to indicate good entry points.

When Volume Diverges from Price

135

If buying points are relatively easy to find by studying large players’ behavior, the detection of selling points requires studying the behavior of all the players, not just the large trendsetters. There are two reasons to justify such a different strategy between buying and selling. First, large funds are much more cautious when they sell than when they buy. When a large fund buys a considerable amount of shares, this influences the demand/supply, which could eventually push the price up. If the price moves up before the fund has finished accumulating, the result is that the fund will turn a paper profit on its previously acquired shares. No fund manager will be blamed for making a profit. However, if the fund manager sells those shares too quickly, the demand/supply equilibrium could change so much as to push the price down. The fund manager will then start to incur a loss on the leftover holding position. For this reason, selling usually takes place during a longer period of time than buying. Moreover, the fund manager could be tempted to place shares at the ask price, in order not to pressure the price down by taking off the bid. This passive strategy is more difficult to detect through volume analysis. Therefore, using the total Effective Volume to evaluate selling signals will more often give better results than just taking the Large Effective Volume. Second, shareholders usually sell to take their profits. This means that some start selling at a 10 percent profit, others at a 20 percent profit. When you analyze the selling pattern, it is true that selling could start even when the price is still increasing, because shareholders do not always wait for tops before selling; they will start to sell when their profit target has been reached. However, in the case of buying, the great majority buy when they detect that the stock is a good value play. This is why finding bottoms is easier than detecting tops.

Buying Pattern Analysis When we are looking to buy a position, we want to know if at a given time the accumulation of shares by large players during the analysis period can justify the price change that occurred during the same period. Two cases can be pointed out: 1. If the accumulation of shares by large players is positive but the price

is moving down or sideways, then we may conclude that accumulation is under way, without knowing yet if this accumulation will be strong enough to move the price up at some point. However, by comparing to past divergences this positive divergence strength between large players’ activity and the price, we may conclude that the actual divergence is much higher than historical divergences, and therefore that it is a good time to buy the stock.

136

VALUE IN TIME

2. If the accumulation of shares by large players is proportionally

stronger than the price increase during the same period, we may conclude that the uptrend will continue. Figure 3.13 shows a divergence analysis example for Darden Restaurants. In the lower panel of Figure 3.13 are represented both the Large Effective Ratio of Figure 3.11 and the price rate of change of Figure 3.5, with some scale adaptation. The divergence between these two signals is represented in the upper panel of Figure 3.13. For example, let’s consider the two divergence points D1 and D2. D1 is calculated by subtracting P1 from ER1. This D1 measures the difference between the evolution in large players’ activity and the price

FIGURE 3.13 Darden Restaurants: 60 days, divergence analysis, including price gaps.

When Volume Diverges from Price

137

evolution during the 5.3 days that led to point D1. On point D1, we can see that the Effective Ratio was in a downtrend, but still above 0 percent, indicating a positive total buying movement. However, the divergence between the Effective Ratio and the price had been increasing up to point D1 because the decrease of the price rate of change had been more important than the decrease of Effective Ratio. This indicated that the price drop would be difficult to sustain for a long period if the accumulation by large players was to continue. By contrast, at the D2 point, it is clear that accumulation (ER2) is very strong, even if the price rate of change (P2) is still close to 0 percent. This probably indicates that the price will later on catch up with the strong large players’ accumulation trend.

Price Gap Corrections You may have also noted in the price rate of change signal in Figure 3.13 one trough (T) followed by one peak (P). These sudden changes are only due to the large price gap that entered and exited the analysis window, as was explained with Figure 3.4. Price gaps represent a sudden reaction to news that occurred while the markets were closed. This news induced a change of sentiment or a change in valuation, which is suddenly expressed in a price change. However, this sudden price change does not appear on the Effective Volume side for the following three reasons: 1. The Effective Volume is calculated on a minute-by-minute basis. Since

we must compare the price of the actual trading minute to the price of the previous trading minute in order to know the direction of the Effective Volume, the first trading minute of the day is never taken into account. To take into account the first minute of trading, we should compare it to the last minute of trading of the previous day, which would be strange since so much time passed between these two trading minutes. Therefore, the Effective Volume for the first trading minute must be set to 0. 2. The second reason is linked to the separation between large players and small players. The volume exchanged in the first minute of trading is usually higher than the volume exchanged later, because a large number of individual investors would have placed their orders during the night, and many of these orders would be executed at the opening of the trading day. If we use this unusually high volume of the first minute of trading, it will often be categorized as “large players” volume in the large/small players separation calculation. This calculation does

138

VALUE IN TIME

not correspond to the reality, since the volume of the first trading minute is the aggregate of a large number of small players’ decisions that took place during the night while the markets were closed. Therefore, this first trading minute should not be mixed with other trading minutes. 3. The last reason is mathematical: Volume comes in bursts, which means that volume is very volatile in the short term (you could have 100,000 shares exchanged in one minute, then 100 shares in the next). However, price is nonvolatile in the short term: Between one minute and the next, price usually moves by only a few cents. The problem is that price gaps introduce a very large peak in price change that is sometimes hundreds of times larger than the usual price change that occurs from one minute to the next. This greatly disturbs the results of the mathematical calculation and could lead to misleading signals. Figure 3.14 shows for Darden Restaurants the price rate of change that was represented in Figure 3.5, but for which price gaps have now been eliminated. When applying this new price rate of change to the divergence analysis, we obtain Figure 3.15, which is very similar to Figure 3.13. The major difference is that because of the price gap, we can see that in Figure 3.13 a strong selling divergence had been generated, while this is no longer the case in Figure 3.15.

Historical Analysis of Buying Divergences The objective in performing a historical divergence analysis is to detect the general reversal pattern of the divergence signal in order to find out if the current value of the signal is strong compared to historical signals.

FIGURE 3.14 Darden Restaurants: 60 days, gaps corrected price rate of change smoothed with an exponential average function.

When Volume Diverges from Price

139

FIGURE 3.15 Darden Restaurants: 60 days, divergence analysis, excluding price gaps.

Figure 3.16 shows the starting point of the historical divergence analysis for Darden Restaurants. It is clear that buying around $34 and selling around $43 or $44 would have been a good trade. In the middle panel of Figure 3.17, I have highlighted nine points that represent peaks in divergence. This means that at these points, the divergence between the Large Effective Ratio and the price rate of change was at its maximum, indicating possible buying points. However, the really good buying points are those that show the strongest divergence among those nine points. I have represented with a dashed line the average level of the peaks of the divergence signals. Experience has shown me many times that it pays to buy a stock when the divergence is higher than this average of the historical peaks. I have pointed out in Figure 3.17 the buy zones as defined earlier. Note that at the right of the graph, we are again in a buy zone.

140

VALUE IN TIME

FIGURE 3.16 Darden Restaurants: price pattern. Source: Chart courtesy of StockCharts.com.

Selling Pattern Analysis When selling, we want to know if at a given time, the distribution of shares by all the players during the analysis period can justify the price change that occurred during the same period. Two cases can be pointed out: 1. If the distribution of shares by all the players is strong but the price

is moving up or sideways, then we may conclude that selling is under way, without knowing yet if this selling will be strong enough to move the price down at some point. However, the comparison of the present selling divergence to historical selling divergences can lead us to a correct conclusion regarding the timing of our own selling decision. 2. If the distribution of shares by all the players is proportionally stronger than the price decrease during the same period, we may conclude that the downtrend will continue. In Figure 3.18, I have highlighted nine points that represent troughs in selling divergence. This means that at these points, the divergence between the total Effective Ratio and the price rate of change was at its minimum, indicating possible selling points. However, among these points, we have to eliminate the troughs that are above 0 percent, since those indicate a positive divergence. Indeed, by definition, a positive divergence cannot lead us to sell the stock. A positive divergence between total Effective Ratio and the price indicates not only that the price has been moving down more quickly than the total Effective Ratio, but that the total Effective Ratio is quite strong; this could indicate a possible price reversal coming down the

When Volume Diverges from Price

141

FIGURE 3.17 Darden Restaurants: historical buy divergence analysis. Historical buy divergence analysis allows comparing the actual divergence signal between the Large Effective Ratio and the price rate of change to the average of the positive past peaks of the divergence signals. A higher signal than the past average indicates that the divergence is much stronger than usual, signaling a probable good buying opportunity.

142

VALUE IN TIME

FIGURE 3.18 Darden Restaurants: historical sell divergence analysis. Historical sell divergence analysis allows comparing the actual divergence signal between the total Effective Ratio and the price rate of change to the average of the past troughs of the divergence signals. A lower signal than the past average indicates that the divergence is much weaker than usual, signaling a probable good selling opportunity.

When Volume Diverges from Price

143

FIGURE 3.19 Darden Restaurants: historical sell divergence analysis (price gaps not corrected). When price gaps are not corrected, the divergence analysis can lead to wrong trading signals.

road. The troughs that are above 0 percent can be labeled as false selling signals. However, after eliminating these false selling signals, the really effective selling points are those that show the strongest divergence among the negative troughs selling divergence points. In this case, we can see only points 4 and 5. But, since the historical comparison is going back one year (points not represented on the graph), the real average level of the bottom selling divergence signals has been calculated at −1.2 percent. As can be seen in Figure 3.18, there is no bottom selling divergence signal that is lower than the average. This indicates that in the long term, the uptrend is still solid. In Figure 3.19, I have represented the divergence signal that would have been generated if the price gaps had not been corrected. In such a case, we can see that the average is slightly lower at −1.93 percent and that points 4 and 5 are real selling signals that were generated by the price gaps. From now on, I will use the following notation on the divergence figures:

r Buy zone limit—the average of the peaks of the past buying divergences.

r Strong buy zone limit—a limit that is 1.5 times higher than the average of the peaks of the past buying divergences.

r Sell zone limit—the average of the troughs of the past selling divergences.

144

VALUE IN TIME

r Strong sell zone limit—a limit that is 1.5 times lower than the average of the troughs of the past selling divergences. The strong buy and sell limits will allow us to sometimes consider only the strongest of the divergence signals.

EXAMPLES OF DIVERGENCE ANALYSIS This section is divided into three parts: 1. The first part includes four straightforward examples on how to use

divergence analysis as a stand-alone tool. 2. In the second part, I use the trading examples from the previous chapters to see what the divergence analysis tells us for each of them. These examples are somewhat more complex than the four straightforward examples and might be more suitable for a second reading of the book. 3. In the third part, I show a complete example involving both the divergence analysis and Active Boundaries indicators. In this example, I guide you step by step in making trading decisions only on the basis of these two indicators.

Straightforward Signals The basic objective of the divergence analysis tool is to give an automatic warning when something strange is going on behind the scenes. This warning is triggered only when what is happening is very unusual compared to past movements. Let’s look at four straightforward examples. Westlake Chemical (WLK) Westlake Chemical Corporation is a manufacturer of basic chemical components (vinyls, polymers, etc.). The company’s main costs are linked to oil costs. In fact, it is a pretty orthodox and stable business. Figure 3.20 shows the price trend for the past few months. The drop of February 20, 2007, was due to an earnings report that fell below expectations. The upper and middle panels of Figure 3.21 show that large players started to buy at the start of December 2006 (trend D in Large Effective Volume), when the price started to drop (trend A in price evolution). The lower panel shows this divergence. This lower panel gives us two types of information: 1. It indicates when the divergence is greater than the average past max-

imum divergences (indicating that we are entering into a buy zone).

When Volume Diverges from Price

145

FIGURE 3.20 Westlake Chemical: buying opportunities. Source: Chart courtesy of StockCharts.com.

2. It indicates when the divergence is at its greatest (points 1, 2, and 3).

I would typically buy the stock if the following conditions are met:

r The divergence signal must be in a strong buy zone, meaning that it must be at least 1.5 times stronger than past peaks of the signal. (In the case of Westlake, the signal must reach 9 percent.) r The Large Effective Volume is trending up (large players are buying). r The price moves up above the 9-day average. I am adding this last condition in order to avoid investing too early; a retail investor is better off waiting for the trend to start. This condition may not apply to large funds that need to invest well ahead of the new trend. I have pointed out in Figure 3.20 the two buying opportunities (A and B) that met these conditions. A naysayer might object at this point and state that by mid-March, the stock price was lower than the price of the A and B buying opportunities. This is true, but trading is also about selling at the right time. We can indeed see in Figure 3.22 that by February 15, the Active Boundaries indicator was reaching the Upper Boundary, a place of trend reversal. Furthermore, we can see at the right of the graph in Figure 3.23 that on February 15, large players stopped buying and the price started to move down. If a stock becomes expensive and funds are not buying any longer, the price is likely to fall back down. It is difficult to miss such a selling opportunity. Sierra Health Services, Inc. (SIE) Sierra Health Services, Inc. is a managed health care organization—in other words, another orthodox and

146

FIGURE 3.21 Westlake Chemical: buy divergence analysis.

VALUE IN TIME

When Volume Diverges from Price

FIGURE 3.22 Westlake Chemical: Active Boundaries.

FIGURE 3.23 Westlake Chemical: small selling divergence.

147

148

VALUE IN TIME

FIGURE 3.24 Sierra Health Services: buying opportunities. Source: Chart courtesy of StockCharts.com.

stable business. Figure 3.24 shows the price trend for Sierra Health Services, indicating three buying opportunities (A, B, and C), which will be studied next. The upper and middle panels of Figure 3.25 show that SIE has been displaying a very strong divergence between large players and price. You might wonder why. The divergence analysis pointed out two buying opportunities: A and B. The opportunity A was somewhat difficult to catch, since the stock price quickly moved back down below its 9-day moving average (see Figure 3.24). The B entry was easier to catch. Another buying opportunity occurred on January 18 and 19, 2007 (opportunity C), when the price moved above the 9-day average after the divergence analysis flashed a strong buying divergence signal (point X in Figure 3.26). Eventually, the company was bought on March 12, 2007. It looks like the negotiations lasted a good three months. Celgene Corporation (CELG) Celgene Corporation is a biopharmaceutical company that develops medicines to treat cancer and immune/inflammatory-related diseases. This is clearly a more exciting, higher-growth, and higher-risk business. Figure 3.27 shows the price trend for Celgene Corporation, indicating two buying opportunities (A and B) and a selling opportunity (C), which will be studied next. The upper and middle panels of Figure 3.28 show that Celgene Corporation was displaying some general divergence between the Large Effective Volume and the price. The lower panel shows that this divergence produced a buy signal at point A. In August 2006, I presented for one of the first times the Effective Volume concepts. My presentation was to a train-

When Volume Diverges from Price

149

FIGURE 3.25 Sierra Health Services: buy divergence analysis (September to December 2006).

150

VALUE IN TIME

FIGURE 3.26 Sierra Health Services: buy divergence analysis (December 2006 to January 2007).

When Volume Diverges from Price

151

FIGURE 3.27 Celgene Corporation: buying and selling opportunities. Source: Chart courtesy of StockCharts.com.

ing class that had been organized by Alexander Elder. (If you have never participated in Dr. Elder’s seminars, I invite you to attend one of them. The experience is always fun and enlightening.) I communicated this stock pick to the class, as well as the next stock pick, American Airlines, which was not yet ready for a purchase. For the sake of full disclosure, I bought at point A, but later had to sell because of a time-limit stop. A time-limit stop obliges a trader to sell if, after a fixed number of days, the stock does not move in the direction of the trade. In Chapter 5, I will fully show how this parameter allows you to manage both the return and risk factors of your trades. You will also notice in Figure 3.27 the possible buying opportunity at point B. I rejected this buying opportunity (to my later regret), however, because the divergence signal at that time had already sunk below the buy limit after having produced the buy alert X (see Figure 3.29). This example shows that in order to trade the divergence signal, you need to be in the buying divergence territory while the price is above the 9-day average. On December 8, 2006, to my surprise, I received an e-mail from Alexander Elder, who had followed my pick and had bought Celgene. He was considering selling and wanted to see what my charts were saying. As you can see from Figure 3.30, the divergence analysis indicated a sell signal at that time. It is interesting to note that from the same original pick, an experienced trader (Dr. Elder) netted a profit higher than 20 percent, while I turned a 2 percent loss. This was a good lesson for me, however. It forced me to study the optimum level of stop loss that gives the best risk/return balance for my trades. Chapter 5 is dedicated to the issue of the risk/return balance.

152

FIGURE 3.28 Celgene Corporation: first buy divergence analysis.

VALUE IN TIME

When Volume Diverges from Price

153

FIGURE 3.29 Celgene Corporation: second buy divergence analysis.

AMR Corporation (AMR) AMR Corporation is a well-known airline company (American Airlines). I also presented this possible trade during Dr. Elder’s seminar in mid-August 2006. Figure 3.31 shows the price trend for AMR Corporation, indicating two buying opportunities (A and B), which will be studied next. As can be seen in Figure 3.32, the company experienced very heavy buying from large players while the price was still in a trading range. Point A indicates a maximum divergence, well above the buy limit. This triggered an instantaneous buy signal, since the price was above its 9-day average. Note also the second buying opportunity (at point B) that occurred on a price breakout movement, which the divergence analysis indicated (see lower panel of Figure 3.33).

More Complex Examples for a Second Reading As this point in the chapter, I have introduced and shown examples of the basic concepts of the Effective Ratio and the Divergence Analysis tools. I believe that studying more complex examples will be better suited to a second reading of the book. I therefore strongly advise momentarily skipping not only this section but also the following three sections: “Combining Divergence and Active Boundaries,” “How to Set the Optimal Analysis Window,” and “Empty Trading Minutes.” Indeed, I think that it would be more enjoyable in a first reading to jump directly to the conclusion of this chapter, and then move on to Chapter 4, “Supply and Demand.” Federated Investors Inc. (FII): The Standard Divergence Play We saw in Chapter 1 (Figures 1.17 to 1.20) that Federated Investors was

154

FIGURE 3.30 Celgene Corporation: sell divergence analysis.

VALUE IN TIME

When Volume Diverges from Price

155

FIGURE 3.31 AMR Corporation: buying opportunities. Source: Chart courtesy of StockCharts.com.

showing a strong accumulation by large players before it jumped out of its trading range on August 3, 2006. Figure 3.34 shows the buy divergence analysis between May and September 2006. As can be seen, this analysis produced three buy zones: 1. Buy zone A starts while the price is still in a trading range, on June 8,

2006. If you have the patience to wait until the top of the divergence is reached during that buy zone, you can see that this maximum divergence is reached on June 12 while the price is in free fall. This forbids us to buy, because as a rule you may not buy or sell against the price trend. 2. Buy zone B is more interesting, because the price trading range is kept all through the buy zone. Therefore, buying at the maximum divergence on July 24 is a good decision. 3. Buy zone C is more problematic. If you did not buy during buy zone B and are now receiving a new buy signal C that is as strong as the B signal, it is still less tempting, because in the meantime the price has already increased from $31 to $34, and the starting uptrend has now been discovered by everyone. Ariba, Inc. (ARBA): Difficulty of Detecting Insiders through Divergence Analysis We studied Ariba, Inc. in Chapter 1 (Figures 1.23 and 1.24). We saw that during the three trading days leading to January 24, 2006, large players were net buyers, while the stock price was declining. Then the January 24 earnings release was positive and pushed the price up by 19 percent overnight.

156

FIGURE 3.32 AMR Corporation: first buy divergence analysis.

VALUE IN TIME

When Volume Diverges from Price

FIGURE 3.33 AMR Corporation: second buy divergence analysis.

157

158

FIGURE 3.34 Federated Investors Inc.: buy divergence analysis.

VALUE IN TIME

When Volume Diverges from Price

159

As can be seen in Figure 3.35, the two buy zones A and B were good buy zones for the stock, since the price considerably appreciated after the buys. However, buy zone C was not very good, since afterwards the price declined. As we can see in Figure 3.35, the three days leading to January 24 were not within any of the A or B buy zones. It is important to understand that the divergence analysis is very good at finding out accumulation by large funds, but is not reliable for finding insider activity such as the one that the Effective Volume analysis showed for Ariba on January 24, 2006. The reason is very simple: Insiders usually get their privileged information only a few days before the news hits the wires. Therefore, they cannot drastically influence the divergence pattern, since this pattern is formed through the analysis of shares during a period that is longer than the time during which insiders accumulated their shares based on their privileged information. Figure 3.36 represents the sell divergence signal. The sell divergence is calculated on the comparison between the price rate of change and the total Effective Ratio. As can be seen, the average minimum of the sell signals was at −2.6 percent. Whenever the divergence signal falls below this average, we need to sell the stock that was bought earlier. In our case, the shares that we could have bought at buying point C would have been sold on March 23, 2006, at the start of the sell zone. Since the total Effective Volume usually changes its trend before the price starts to decline, the sell divergence analysis is a very good protection against bad trades. It often triggers before stop-loss limits. I use stop losses only for protection against catastrophic situations: a sudden market crash or very bad news that endangers the viability of the company. Also note in Figure 3.36 the false buy signals. These are false signals because their tops do not reach the 0 percent limit. A negative divergence always means that the Effective Volume trend is weaker than the price trend. When the maximum of the divergence signal is negative, this is usually a very bad sign for the stock.

Becton Dickinson (BDX): Forced to Sell by Divergence, Only to Repurchase Later at a Higher Price Becton, Dickinson was studied in Chapter 2 when we discussed Active Boundaries (refer if necessary to Figures 2.3 through 2.6). As can be seen in Figure 3.37, the divergence analysis produced two large buy zones (zone A and zone B). Buying in zone A at around $60 was a good choice, since later on the price climbed to $70. The Active Boundaries analysis was also showing that zone A corresponded to the Lower Boundary, from which the price usually rebounds. You will also notice that the buy signal at the right of the graph in zone B is much

160

FIGURE 3.35 Ariba, Inc.: buy divergence analysis.

VALUE IN TIME

When Volume Diverges from Price

FIGURE 3.36 Ariba, Inc.: sell divergence analysis.

161

162

VALUE IN TIME

FIGURE 3.37 Becton, Dickinson: buy divergence analysis.

higher than historical signals, indicating a very unusual accumulation by large funds. What is interesting to note, however, is the large sell signal that occurred on October 10, 2006, just before the stock climbed up again in buy zone B. This sell signal was due to an important distribution of shares by

When Volume Diverges from Price

163

FIGURE 3.38 Becton, Dickinson: a rare case where the price trend was leading the Effective Volume trend.

large players. We can indeed see in Figure 3.38 that leading up to October 2 large players were heavy sellers, and that it took a strong price trend to move the Large Effective Volume trend back up. This clearly shows that large players are not always correct. As a matter of fact, nobody could have predicted that the price would continue to appreciate, and the large sellers probably just wanted to take their profits off the table. My policy, when facing such a heavy selling by large players just after a price run, is to also sell and get my profits off the table, even if I will later need to buy the same stock back at a higher price. It is too risky to bet against the large players’ move, even if it is later found to be wrong. When I was starting to trade, I had difficulties coming back into a stock at a price higher than the one at which I had last sold it. It was some sort of an admission of a failed decision to sell when it would have been wiser to keep the stock. But in fact, the market does not care about the past. It does

164

VALUE IN TIME

not care about your or my position. Therefore, there is no reason for me to care about such past positions. Once you sell a position, it is over. If the conditions change, it is perfectly correct to buy the same stock at a price higher than the one at which you previously sold it. You will come back into the stock only if the probability of making a profit is high enough. In other words, the technical analysis could very well tell you that it is less risky to buy a stock later on at a higher price than to hold the stock during a period of uncertainty. IMAX: How to Avoid Catastrophic Situations Using Divergence Analysis We studied the IMAX case in Chapter 2 (Figures 2.9 and 2.10). We can see in August 2006 a catastrophic gap downward that took out about 40 percent of the value of the stock. The problem is that the divergence analysis was showing a strong buy signal just before the price plunge. As shown in Figure 3.39, if you had bought during buy zone D, you would certainly have experienced a catastrophic loss. In Figure 3.39, I show the standard buy zone limit, as well as the strong buy zone limit, which is 1.5 times stronger. Let’s be clear: Divergence analysis will not protect you from catastrophic losses all the time. The divergence analysis helps you determine if something abnormal is happening, something that may require you to buy or sell the stock. However, if bad news is closely held, you may not be able to detect it in advance through the divergence or Large Effective Volume analysis. It is therefore a safer choice to make decisions based on a combination of indicators. Let’s see how this worked for IMAX during buy zone D.

r On July 28 (see Figure 2.9b), the Active Boundaries signal hit the Upper Boundary, a traditional place for reversal. Buying at the Upper Boundary must be done with extreme care. r The Effective Volume analysis (see Figure 3.40) shows that before reaching the price gap on the morning of July 27, large players were accumulating the shares, which is typical of information leaks before positive news (trend A in Large Effective Volume flow). However, two days after the gap-up, large players started selling, and the price headed down. As we saw in Chapter 2, it is forbidden to buy a stock against the large players’ trend. r We can see in Figure 3.40 that price trend B was down from July 28. One of my rules is that I do not buy when the stock price is decreasing, and I do not short when it is increasing, because the price trend has its own momentum that is often slow to change. As I wrote earlier, I even prefer buying when the price crosses above its 9-day average. Buying at the buy divergence signal would therefore go against this basic trading rule.

When Volume Diverges from Price

FIGURE 3.39 IMAX: buy divergence analysis.

165

166

VALUE IN TIME

FIGURE 3.40 IMAX: Effective Volume analysis.

The whole problem of the false buy signal generated by the divergence analysis lies in the fact that we have made gap corrections. The consequence of these gap corrections is that we are viewing the price trend as if there had been no gap. However, after a large gap (such as the one on July 27), the market players’ reactions often change. This is why, after a large gap, all the signals must be reevaluated for their correctness: The Active Boundaries could form new boundaries, the Large Effective Volume analysis may not be valid anymore, and the divergence analysis could give opposite results depending on whether you include the price gaps. Indeed, if you look at Figure 3.41, you can see that the middle panel (which calculated the divergence after gaps are corrected) gives a buy signal, while the lower panel (which calculated the divergence without gaps correction) produces a sell signal. Similarly, we can study the buy zones E and F in Figure 3.39. We can see that both zones E and F are false

When Volume Diverges from Price

167

FIGURE 3.41 IMAX: buy divergence analysis with or without gaps correction. Divergence analysis just after a large gap can give very different signals depending on whether the gaps are corrected.

168

VALUE IN TIME

FIGURE 3.42 IMAX: first false buy signal. This first false buy signal was indicated by the divergence analysis on buy zone E in Figure 3.39, but contradicted by the Effective Volume analysis.

buy zones, because none of these two signals is sustained by the Effective Volume trends (see Figures 3.42 and 3.43).

Combining Divergence and Active Boundaries In Chapter 2, we extensively studied how to monitor the trend of Tellabs using the Active Boundaries (for reference, see Figures 2.7a and b, and 2.11a and b). Let’s go again through the full up and down cycle of Tellabs (see Figure 3.44) and study what the divergence analysis is telling us in relation to the Active Boundaries. The numbers in Figure 3.44 are different points that will be discussed with only one focus: Should we buy or sell?

When Volume Diverges from Price

169

FIGURE 3.43 IMAX: second false buy signal. This second false buy signal indicated by the divergence analysis on buy zone F in Figure 3.39, but contradicted by the Effective Volume analysis.

FIGURE 3.44 Tellabs: price cycle. Source: Chart courtesy of StockCharts.com.

170

VALUE IN TIME

Let’s first summarize the four trading rules that we must follow: 1. Buy close to the Lower Boundary if:

r The divergence analysis is in a buy zone and r The price is not decreasing (it is above the 9-day moving average) and

r The Large Effective Volume is increasing. 2. Buy close to the Upper Boundary if:

r The divergence analysis is in a buy zone and r The price is not decreasing (it is above the 9-day moving average) and

r The Large Effective Volume is increasing and r The Active Boundaries signal did not cross below 0 percent between the last time it hit the Upper Boundary and now. 3. Sell close to the Upper Boundary if: r The divergence analysis is in a sell zone or r The Large Effective Volume is not increasing. 4. When the price passes through the Lower Boundary, do not buy. Instead, wait for new boundaries to be formed. Please note among these rules that in order to buy, we need a combination of all the conditions to be met, but in order to sell, we need only one of the conditions to be met. This means that we must use great caution when buying and sell at the first sign of trouble. Trade Analysis: A, B, and C Trends of Figure 3.44 We are now going to review the important decision points that will lead us to successful trading decisions.

r Points 1, 2, and 3 in Figure 2.7 were used to define the boundaries. We therefore start our analysis at point 4 in Figure 3.45, close to the Lower Boundary. The Large Effective Volume is not increasing (see Figure 3.46), so even though we are in a buy zone (see Figure 3.47) and the price is flat, we cannot buy here. r At point 5, we just went through the Lower Boundary. It is a very bad idea to buy here; some large player wanted badly to be out (see Figure 3.46). Furthermore, we are in a sell zone (see Figure 3.48). r At point 6, large players start moving up, but we are not in a buy zone yet, and we are at the Upper Boundary, a traditional reversal point. It is impossible to buy here. r Point B (buy) in Figures 3.45, 3.46, and 3.47 is interesting. It lies close to the Upper Boundary, in the buy zone, with Large Effective Volume

When Volume Diverges from Price

171

FIGURE 3.45 Tellabs: Active Boundaries for the A, B, and C trends in Figure 3.44.

FIGURE 3.46 Tellabs: Effective Volume analysis in the uptrend.

172

VALUE IN TIME

FIGURE 3.47 Tellabs: buy divergence analysis in the uptrend.

increasing and the price trend not decreasing. This is a clear buy at $10.7. r Since we bought, we now have to look for a selling point. Point 7 is definitively not a selling point, even if we are at the Upper Boundary. Indeed, large players are still heavily buying, and we are far from a sell zone on the divergence analysis chart (see Figure 3.48). r Point 8 makes us very worried: We are back at $10.7, our buying price, even though we reached $11.88 at point 7. What should we do? Should we sell in order not to turn our paper gain into a loss? However, at point 8, we are located at the Lower Boundary, with large players heavily buying, and the divergence analysis signal is at its highest. We keep the

FIGURE 3.48 Tellabs: sell divergence analysis in the uptrend.

When Volume Diverges from Price

r

r

r

r

r r

173

stock. There is no reason to sell when large funds are buying on such a scale. At point 9 we have a nice paper profit. What should we do? Should we take our profit and “sell the news”? Large players are not selling the news. We keep the stock. At point 10, we are somewhat higher than the Upper Boundary, an exuberance situation that cannot last. However, large players are still accumulating, the price trend is healthy, and the divergence analysis does not show a sell signal. We keep the stock. At point 11, we are back to the Upper Boundary. However, the Large Effective Volume flow is flat, and the divergence analysis signal is strongly in the sell territory. We actually sell before reaching point 11, as soon as we reach the sell zone. We sell at point S (sell), for $14.71. At point 12, on the drop in price, even if the price returns to its positive uptrend, the Large Effective Volume stays flat, and the divergence analysis signal is not back into the buy zone. We wait. Points 13 and 14 show the same pattern: Both lie deep in the sell zone. Point 15 is interesting: We are back in the buy zone. Should we thus buy? At that point, the Effective Volume flow is flat or slightly decreasing. The divergence is producing a “buy” signal simply because the price dropped much more quickly than the drop in Effective Volume. Furthermore, we passed through the Lower Boundary, indicating that something is happening. As a matter of fact, the large price drop has been attracting sellers wanting to protect their past gains or shortsellers wanting to profit from the shaky states of the shareholders. It is safer to wait.

Trade Analysis: D, E, and F Trends in Figure 3.44 As we finished point 15 of the analysis without rebuying, we are still looking for a good entry point, if ever. Just as a reminder, a good entry point is a value point (in terms of Active Boundaries) where large players are accumulating with abnormal strength.

r Since we crossed through the first Lower Boundary at point 15, we may think that point 16 is forming a new set of boundaries (see the second set of boundaries in Figure 3.49). Since we are not sure, it is better to wait until the second set of boundaries is formed. r At point 18, we may think that the second set of boundaries has been fixed. However, the Large Effective Volume is trending down (Figure 3.50), so even though point 18 is in a buy zone (Figure 3.51), we wait. r Point 19 is still worse in terms of the Large Effective Volume trend. We are in a sell zone (Figure 3.52). The stock is now in a confirmed downtrend. We continue to wait.

174

VALUE IN TIME

FIGURE 3.49 Tellabs: Active Boundaries for the D, E, and F trends in Figure 3.44.

FIGURE 3.50 Tellabs: Effective Volume in the downtrend.

When Volume Diverges from Price

175

FIGURE 3.51 Tellabs: buy divergence analysis in the downtrend.

r Points 20 and 21 show false buying points (Figure 3.51). Indeed, at these points, the divergence between Effective Volume and price is positive simply because the price fell too sharply compared to the Effective Volume. However, the Large Effective Volume trend is also negative (Figure 3.50), forbidding us to buy. r The next buying point is point 24 ($11), which combines a strong uptrend in Large Effective Volume with a flat price. But then, unfortunately, point 25 ($10.6) would be hit quite quickly thereafter, and we would sell at a small loss (−3.6 percent). r Buying at point 26 ($11.5) would also result in selling at point 27 ($11.3) for a small loss (−1.7 percent).

FIGURE 3.52 Tellabs: sell divergence analysis in the downtrend.

176

VALUE IN TIME

Between points 23 and 27, the Large Effective Volume flow analysis shows a big battle that is taking place between funds. Some funds that bought in at the beginning of the year and could not sell earlier in the downtrend are caught with losing positions, while their reporting period of the end of the year is approaching. Some of these funds are tempted to sell their Tellabs holdings on any new price strength. In the meantime, other funds that sold at the top are encouraged to return to the stock, whose valuation is still low compared to the previous high. As you can see in this Tellabs example, a combination of the different volume-based tools produces very few good trading signals. This method is indeed much more restrictive than many other methods. As a consequence, by following this method, you will be invested in a specific stock for only a small amount of time. In our example, we made a one-shot 37 percent gain by being invested only between the end of December 2005 and the end of February 2006 (a two-month period out of 13 months), and two small losses for a total loss of 5.3 percent. What the high return on the short time period tells us is that if we need to be invested in a diversified portfolio of 20 different stocks, and if for each stock our trading method allows us to stay invested for only 15 percent of our trading time, then we will need to follow at least 133 stocks (20 ÷ 0.15 = 133). In fact, you will likely need to follow many more, because trading signals coming from different stocks usually overlap. Nobody can follow so many stocks on a daily basis, except with the help of an automatic scanning system. I myself follow more than 800 stocks daily, but actually analyze fewer than 20 every day—the ones that are flagged by the system.

HOW TO SET THE OPTIMAL ANALYSIS WINDOW So far in this chapter, we have followed a basic principle: We need to buy when there is a strong positive divergence between the Large Effective Ratio and the price rate of change, and we need to sell when there is a strong negative divergence between the total Effective Ratio and the price rate of change. Unfortunately, life is not that simple, because volume and prices have some very different characteristics, which makes divergence analysis rather tricky. The area in which they differ is called volatility. Price volatility refers to the number of price reversals within a given period of time, together with the amplitude of such reversals. Thinly traded stocks (for example, those listed in the Pink Sheets) have higher price volatility than the well-traded blue chips.

When Volume Diverges from Price

177

Volume is much more volatile than price. The amplitudes of the volume variations can be very significant to the point of rendering divergence analysis useless. The cause of such volatility comes from large funds that have to trade a significant number of shares in order to make a meaningful profit. Since trading hours are limited, funds need to move large amounts of shares in and out during a short period of time. Even if the extreme volumes are filtered out, during a short period of time volume volatility will be much higher than price volatility. Let’s come back to Figure 3.17 (Darden Restaurants Inc.). In Figure 3.17, I took an analysis window of 5.3 days to calculate the Effective Ratio and the price rate of change as they are represented in the lower panel of Figure 3.17. In that case, the divergence analysis of the middle panel allowed us to make some trading decisions. Let’s now take a much shorter analysis window of one day, and process again the divergence analysis, as in Figure 3.53. You will at once notice that the divergence signal closely follows the Large Effective Ratio signal, simply because with such a small analysis window, the gray signal that represents the price rate of change is very small in amplitude. This only shows that for very short analysis windows, volume is much more volatile than price. If you follow the buy signals generated by this analysis, you can see that you have a very large number of buy signals (you buy when divergence is high). What do these signals represent? First, let’s remember that we are looking at the large players only. That is, we are looking at the large spikes of volume that were responsible for a price change from one minute to the next. What the model says is that these volume spikes come by waves, and since the analysis window is short, a few consecutive volume spikes will be responsible for these violent moves. It is also important to note that when an analysis window moves minute by minute from older trading minutes to newer trading minutes, we have to add new Effective Volume and subtract old Effective Volume. This has an important consequence: If a new spike of Effective Volume must be added when we enter the corresponding trading minute, this same spike will have a negative effect on the signal when leaving the analysis window. This largely explains the periodic ups and downs that we see in the short-term divergence signal, such as the one in Figure 3.54. This problem occurs only for small analysis windows, for which the volume volatility is higher than price volatility. You can see in the lower panel of Figure 3.53 that Large Effective Ratio spikes can lead to day-by-day variations going from −5 percent to +15 percent, for a total span of 20 percent. If you look at the price rate of change during the same one-day analysis period, you will notice that this rate of change is moving between −3 percent and +5 percent on average, for a total span of 8 percent. On a one-day analysis window, the Large Effec-

178

VALUE IN TIME

FIGURE 3.53 Darden Restaurants: divergence analysis for a very short analysis period.

tive Ratio signal offers an amplitude 2.5 times stronger than the price rate of change. We now understand that on very short analysis windows, trading on the divergence signal is identical to trading on the Large Effective Ratio signal only. Is this what we need as traders? Would it be meaningful to buy when some large player is buying and sell when some other large player is selling? This makes little sense. First, the buyers and the sellers are probably not the same, and second, “large” does not especially mean “correct.” It is not because someone is buying one ton of oranges that we may conclude that the price of oranges is going to increase. It is better to wait and see if a sustained number of large-quantity orange buyers appears, without attracting more orange supplies.

When Volume Diverges from Price

179

In other words, when you use a short analysis window, you are saying: “Oh! Yesterday, large players were buying heavily. I’ll buy, too!” Saying that, you do not even know if the buyers are latecomers who want to catch an uptrend, trendsetters who are accumulating before a price surge, or bargain hunters who are jumping in on a sudden price drop. Trading on a volume signal only is the worst method you could follow. Indeed, when directly investing in stocks, the only two ways to turn a profit are: 1. Buy at a low price and sell at a higher price. 2. Short at a high price and cover at a lower price.

If you read the previous statement again, you will notice that volume has nothing to do with profit. Only price matters: To turn a profit, you need to have the price right more than you need to have the volume right. Let’s not forget the goal of this model: We want to buy when the stock is cheap and when there is buying momentum by large players that is not yet built into the price. We want to sell when the stock price is higher (much higher, if possible) and when there is a selling momentum starting in the market. In terms of the model, this simply means that price is the main indicator. Therefore, the price signal that we use in the model should be stronger or at least equal to the volume signal. You should not, under any circumstances, use a volume signal that is much stronger than the price signal (as is shown in Figure 3.53). Let’s examine Figure 3.54, for which I took an analysis window of 10 days. You can easily notice in the lower panel of Figure 3.54 that the price rate of change is varying much more wildly than the Large Effective Ratio. The price rate of change is moving from −8 percent to +10 percent, while the Large Effective Ratio is moving only from +2 percent to +6 percent. For a 10-day analysis window, the price rate of change offers an amplitude 2.25 times higher than that of the Large Effective Ratio. As a consequence, the divergence between the two (plotted on the upper panel) gives buy and sell signals that are mainly a consequence of the price rate of change variations. The problem with such a long analysis window (10 days) is not only that you get few signals, but that:

r These signals come too late. You will miss the moves by insiders, who typically accumulate during only a few days before the news strikes.

r The limit between false and true signals is not reliable since we do not have enough tops and bottoms to calculate an average that is statistically valid. To get more tops or bottoms to calculate an average, we

180

VALUE IN TIME

FIGURE 3.54 Darden Restaurants: divergence analysis for a very long analysis period.

would need much older data, which in fact does not properly reflect the behavior of the actual pool of traders. The conclusion is that primarily following the price rate of change is much more effective than following only the volume signal. However, this will not give you an edge compared to other indicators that also use measures of price variations. Selecting the right size for the analysis window is critical in order to get the right buy/sell signals. The right size is obviously between 1 day and 10 days, such as the 5.3-day window in Figure 3.17. Since I like to trade on the price trend but also get enough Effective Volume information to tell me if something is happening behind the scenes, I usually decide on the analysis window size in such a way that the maximum amplitude of the Large Effective Ratio is between 75 percent and 100

181

When Volume Diverges from Price

percent of the maximum amplitude of the price rate of change pattern. In Figure 3.17, we can see that the Large Effective Ratio is evolving between −0.5 percent and +8.5 percent (amplitude of 9 percent), while the price rate of change is evolving between −4 percent and +8 percent (amplitude of 12 percent). Thus, we have 9 divided by 12, or 75 percent.

EMPTY TRADING MINUTES If you look closely at Table 3.11, which represents the minute-by-minute data that feeds all the tools presented in this book, you will notice that there are occasionally jumps between trading minutes. For example, you can see that there are three lines lacking between 15:03 and 15:07. We should have seen lines for 15:04, 15:05, and 15:06. The fact is that no trade was executed during these three minutes. Indeed, when there is no activity during a specific trading minute, the data feed usually does not indicate “0” for that minute, but simply does not send data at all regarding that minute. We then end up with a data feed such as the one in Table 3.11. If we want to be mathematically correct, we need to insert “0” lines for the minutes that have not seen any trading, such as in Table 3.12 (see all the new lines with “0” volume). For the different tools that have been introduced so far, neither the Effective Volume Flow analysis nor the Active Boundaries analysis will be affected by “0” or blank lines. However, it is clear that the tools that use a moving analysis window of fixed length will be affected, since including “0” lines will reduce the volume included in the analysis window. This will affect the divergence signal, since the price and the Effective Volume have a very different volatility. In theory, adding “0” lines will slightly increase

TABLE 3.11 A Data Feed That Does Not Include Nontrading Minutes

4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006

15:10 15:09 15:08 15:07 15:03 15:01 15:00 14:59 14:55

Open

High

Low

Close

Volume

8.00 8.00 8.00 8.00 8.01 8.02 8.01 8.01 8.02

8.01 8.00 8.01 8.00 8.01 8.02 8.01 8.01 8.02

8.00 8.00 8.00 8.00 8.01 8.00 8.01 8.01 8.01

8.00 8.00 8.00 8.00 8.01 8.00 8.01 8.01 8.01

300 1,417 2,303 100 1,800 1,884 1,992 1,000 8,897

182

VALUE IN TIME

TABLE 3.12 A Data Feed That Includes Nontrading Minutes

4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006 4/10/2006

15:10 15:09 15:08 15:07 15:06 15:05 15:04 15:03 15:02 15:01 15:00 14:59 14:58 14:57 14:56 14:55

Open

High

Low

Close

Volume

8.00 8.00 8.00 8.00 8.01 8.01 8.01 8.01 8.00 8.02 8.01 8.01 8.01 8.01 8.01 8.02

8.01 8.00 8.01 8.00 8.01 8.01 8.01 8.01 8.00 8.02 8.01 8.01 8.01 8.01 8.01 8.02

8.00 8.00 8.00 8.00 8.01 8.01 8.01 8.01 8.00 8.00 8.01 8.01 8.01 8.01 8.01 8.01

8.00 8.00 8.00 8.00 8.01 8.01 8.01 8.01 8.00 8.00 8.01 8.01 8.01 8.01 8.01 8.01

300 1,417 2,303 100 0 0 0 1,800 0 1,884 1,992 1,000 0 0 0 8,897

the amplitude of the price rate of change signal and slightly decrease the amplitude of the Effective Ratio signal. However, as shown in Figure 3.55, these changes will have a very small effect on the divergence signal itself. This effect is indeed well within the measurement error, and could not possibly affect trading decisions, at least

FIGURE 3.55 Comparison of divergence analysis signal with or without the insertion of blank or “0” lines for trading minutes for which there was no trading activity.

When Volume Diverges from Price

183

if these “0” lines are not too numerous. In the case of Ariba, Inc., the stock is traded only 77 percent of the time. This means that since every trading day includes 390 trading minutes, the stock is traded on average during only 300 minutes. The 90 nontrading minutes are added as “0” lines, but that addition hardly affects the divergence signal, as shown in Figure 3.55. Although one example does not take the force of proof, it is quite obvious that inserting or not inserting blank or “0” lines when a trading minute has seen no trading activity will not affect the divergence signal of stocks that are relatively well traded (well-traded stocks see trading activity for most of the trading minutes). Because the systematic insertion of blank or “0” lines uses a lot of computation power, I have made the decision in my day-to-day trading not to insert them.

WHAT WE LEARNED REGARDING DIVERGENCE ANALYSIS In this chapter, we learned that a high positive divergence between volume and price trends indicates buying points. Such positive divergences occur for one of two reasons: 1. The price is dropping more quickly than the Effective Ratio, meaning

that traders are selling at a slower pace than what the price movement may indicate. In other words, the price drop is preventing an increasing number of traders from selling—more and more traders are thinking that the price is getting too cheap to sell. 2. The price is increasing more slowly than the Effective Ratio, meaning that traders are buying at a quicker pace than what the price movement may indicate. This typically occurs when traders believe that the price’s upward momentum will continue, attracting still more buyers, until the price becomes too high. Trading decisions are made by comparing the divergence signal to the price trend:

r If the price is decreasing but the divergence signal (between the Effective Ratio and the price rate of change) is increasing, it means that the price is dropping only on a few sellers (meaning that nobody is buying), or that bargain hunters have started to come in, slowing the downtrend of the Effective Ratio. Simply put, we do not know what is going on. The best course of action is to wait for the price to stabilize or reverse up.

184

VALUE IN TIME

r If the price is stable but the divergence signal is increasing, this is a sign of underlying change. However, accumulation may continue for days before the price increases. In such a case, we should wait for the divergence to become stronger than its historical average maximum before buying in; and, more important, I would advise waiting for the price to cross above its 9-day average before buying. We learned that trading rules must use a combination of complementary indicators. My set of four trading rules is: 1. Buy close to the Lower Boundary if:

r The divergence analysis is in a buy zone and r The price is not decreasing (it is above the 9-day moving average) and

r The Large Effective Volume is increasing. 2. Buy close to the Upper Boundary if:

r The divergence analysis is in a buy zone and r The price is not decreasing (it is above the 9-day moving average) and

r The Large Effective Volume is increasing and r The Active Boundaries signal did not cross below 0 percent between the last time it hit the Upper Boundary and now. 3. Sell close to the Upper Boundary if: r The divergence analysis is in a sell zone or r The Large Effective Volume is not increasing. 4. When the price passes through the Lower Boundary, do not buy. In-

stead, wait for new boundaries to be formed. We also learned that the divergence method when used in combination with the Active Boundaries signal, because it is very selective, does not provide abundant trading signals. Therefore, this combination of methods forces us to follow a large number of stocks. Some sort of scanning system is thus necessary to automatically analyze a large number of stocks and produce alerts on the stocks that offer good trading opportunities.

CHAPTER 4

Supply and Demand The Key to Trading

t this stage in the book, you know what my basic hypothesis is: In order to make a profit trading, it is necessary to grasp how the supply/demand equilibrium functions. We saw in the previous chapters that this equilibrium is affected by a motive called profit (measured in terms of price) and is moved by a force called strength (measured in terms of volume). There is, however, a third concept that influences the supply/demand equilibrium. It is the resistance to change. It is a well-known fact in physics that when you apply a force to an object, the object responds with an opposite force. In the market, this means that when you apply a force in one direction, you instantaneously generate a counterforce in the opposite direction. When the price of a stock increases, propelled by buying power, that instantaneously attracts active sell market orders as well as passive sell limit orders. We have thoroughly studied the active players’ behavior using the Effective Volume method, but we would do well to also study the behavior of the passive players, those who place limit orders. This is what this whole chapter is all about. The first part of the chapter deals with a measure of the overall supply of shares. We will see that when the supply of shares dries up, the probability of a rebound in price is very high, especially when large funds are net buyers. The second part is somewhat more technical and deals with a minuteby-minute measure of the passive supply of shares. We will see that passive players have only a minor influence on a stock’s price movements. We will see that the market battle is mainly fought between large players—usually

A

185

186

VALUE IN TIME

through very fast computer-based order-generation and routing systems. We will see how this fight prevents market manipulation. This will lead us to the third part of the chapter: a comparison of the strengths and weaknesses of funds compared to retail investors.

SUPPLY/DEMAND EQUILIBRIUM In Chapter 2, we learned that during trends, traders’ expectation is cyclical, moving within two limits called the Upper and the Lower Boundaries. This cyclical movement is mainly influenced by the turnover of the pool of active traders: New buyers lower the average profit of active traders and increase the average expectation that the price will move up. The more numerous the new buyers, the stronger the change in average expectation. When the average profit hits the Upper Boundary, the average expectation of active traders for the price to further increase is at its lowest; at this point, new buyers stop coming in and the price reverses down. However, if you remember, in Chapter 1 (“Effective Volume”), we learned that the buyers/sellers equilibrium is influenced by the supply/ demand equilibrium (Are shares available? Is money coming in?) and volatility (How quickly are shares becoming available for sale? How quickly is money becoming available to purchase shares?). We also learned that 25 percent of the volume involved in stock trading is responsible for 75 percent of the price movements. This 25 percent is called the Large Effective Volume. The buyers/sellers equilibrium depends primarily on the supply/ demand equilibrium. Indeed, the stock market is above all a market (see Figure 4.1). You therefore need to go to that market in order to find a buyer for the shares that you want to sell and a seller for the shares you want to buy. The demand is a measure of the number of shares that people want to buy or the incoming buying orders (see point 2 in Figure 4.1) that come as a reaction to a price change. Only a part of the demand turns into real transactions (see point 4 in Figure 4.1). The supply is a measure of the number of shares that are for sale (see point 6 in Figure 4.1) or that come into the selling book as a reaction to a price change. Only a part of the supply turns into real transactions (see point 4 in Figure 4.1). Many orders that are placed too far outside of the bid/ask range are indeed never executed and will eventually be canceled. Furthermore, the supply of shares is represented not only by what lies in the order book, but also by all the hidden limit orders that are waiting to be placed once a certain price level is reached. To a larger extent, it is also

Supply and Demand

187

FIGURE 4.1 How the stock market works. 1. 2. 3. 4. 5. 6.

Potential buyers follow a specific stock. A buy decision is made; an order is sent. The order reaches the book (market order or limit order). Market orders are executed at market price. The pool of stockholders follows the stock price. Available shares are presented in the market by rank of expectation; the shares with the lowest expectations are usually sold earlier than the shares with higher expectations.

7. For large transactions, some orders can go directly between large holders.

represented by all the shares belonging to the float (the shares available for trading) and by those that could be extended for sales if an attractive price is offered. We learned in the Introduction that decimalization has killed market visibility. Because, post-decimalization, traders no longer had an incentive

188

VALUE IN TIME

to “show their hand,” it became quite difficult to evaluate the size of the supply and the demand for shares. Normally, this equilibrium is indicated in the order book, which lists the limit orders. Limit orders are the waiting buy orders at and below the bid and the waiting sell orders at and above the ask (see point 3 in Figure 4.1). As a trader, if you see a very large amount of shares at the ask, you know that the market is offering shares for sale, and that waiting a little could allow you to buy shares under better conditions. Before decimalization, having access to the book of orders was very helpful and even critical for clearly understanding the buyers/sellers equilibrium. In my opinion, today, after decimalization, the order book is of not much help in figuring out the direction of the market (except in day-trading activities, where one cent can make a difference). This is because today institutional players display only a part of their orders, and usually at the very last moment. The conclusion that we can draw from these observations is that looking at the order book is not very helpful in determining the supply/demand equilibrium. It only gives an instantaneous picture of the market at a given time. In order to trade successfully, we need something more comprehensive. Furthermore, the difference between the buy/sell equilibrium concept and the supply/demand equilibrium concept is not only the number of buy/sell orders that get executed compared to the size of the supply/ demand; it is also the will that is behind each of the two concepts. Indeed, a share that was bought reflects a will, a taken decision, and carries an expectation from the buyer (the buyer expects to make a profit). A share that is coveted (to be bought) is part of the demand and reflects either an intention or a potential expectation. It is not as strong as the real buying or selling act. The same difference exists between placing a limit order and placing a market order. A limit order will be executed if a given price is reached. A limit order enters into the supply/demand basket, but is not immediately executed. Indeed, a limit order needs to meet a corresponding opposite order at the requested price to be executed. A market order is an order to sell or buy the stock at the current price, even if it causes some slippage between the displayed price and the executed price. A market order carries more will than a limit order. There is slippage when a market order is executed at a price that is slightly different from the price at which the order was placed. This small difference is the cost of having instantaneous execution at the best available price. Volatility is the second element that influences the supply/demand equilibrium. There are two types of volatility: the price volatility (How quickly is the price changing?) and the volume volatility (How quickly are shares being supplied and exchanged in the market?).

Supply and Demand

189

Volatility Volatility is defined as the relative rate at which a variable (price or volume) changes. Daily price volatility is mathematically calculated as the standard deviation on an annual basis of daily price changes. We say that the price of a stock has a high volatility if it moves up and down quickly. The one-minute volatility is often very different from the daily volatility, because prices move much less during one minute than during one trading day. When we talk about volatility, it is therefore important to state the period of time we are referring to. Usually, we talk about short-term volatility as opposed to long-term volatility. It is important to remember that, as we saw in Chapter 3, the price pattern has a very small short-term volatility but a significant long-term volatility. However, the volume pattern has an opposite volatility: a stock’s short-term volatility is often significant (100 shares can be exchanged during one minute, and then 100,000 shares could be exchanged during the next minute), while its long-term volatility is small (the 50-day average volume variations are not so large).

The price and volume volatilities are related:

r The more quickly the buying volume appears, the more quickly the price moves up.

r The stronger the price moves, the stronger the volume activity (strong price movements attract stronger supply and demand).

How to Measure the Supply/Demand Equilibrium There are two practical ways to evaluate the supply/demand of shares. The first one is static: a calculation of a price histogram corresponding to the number of shares available for trading (the float). Each share carries its load of emotions, expectations, and so on. Therefore, knowing how many shares have been bought and sold at each price level allows you to know if many shares will be offered for sale when the price increases (indicating resistance), or if many buyers will appear when the price decreases to a certain level (indicating support). Figure 4.2 shows a support line that joins consecutive troughs for the company Reliant Energy Inc. (RRI). This line shows a support that is close to $10.80. The interpretation of such a support line is that in the past, buyers were finding this price level to be cheap enough to buy, and sellers found it to be too cheap to sell; hence, the price would reverse up. Because

190

VALUE IN TIME

FIGURE 4.2 Support line. Source: Chart courtesy of StockCharts.com.

the pool of traders has not changed much since the last time the support line was hit by the price, we may believe that this time buyers and sellers will make a similar analysis as to the value of the stock, and that the price has a fair probability of moving up again from the support line if it falls that far. In Figure 4.3, I represented the volume histogram of all the last-traded shares that belong to the float of Reliant Energy, as of September 25, 2006. Each vertical bar represents the number of shares that were exchanged at the corresponding price. In such a volume histogram figure, the support line is a vertical line that can be drawn at a price location were the volume

FIGURE 4.3 Price histogram.

Supply and Demand

191

histogram shows the lowest number of shares. In the case of Figure 4.3, the support line is set around $10.80. Note that the volume histogram figure gives an instantaneous picture of the repartition of shares on a given date and at a specific time. As we will see later, the volume histogram figure evolves with time. I call this measure of the supply/demand level static because it measures only supply: how many shares will be available. It does not measure the volatility (how quickly shares and money are coming in or out). The second way to measure the supply/demand equilibrium is dynamic, but is used only by players with access to large enough funds. It is how you experimentally measure volume volatility. This dynamic method is equivalent to market testing, and costs money to execute. Some market makers will indeed quickly push the price down and measure how fast a new supply of shares comes into the market or see if the new low price attracts many new buyers. This measure is often executed by trying to move through a support line. If the support line holds, this indicates that the market is ready for a new up leg. This is a well-known testing method. A significant price decline that attracts fewer sellers than an equivalent recent price decline indicates that the supply of shares is drying up. As you may notice, these two methods (either static or dynamic) only measure one part of the supply/demand equilibrium: the supply side. The supply side is indeed easier to measure, since we have access to the past transactions. As we know how many shares have been bought at what price, we can draw some conclusions as to how many could eventually be sold at the current price. This is what I call the supply analysis. The demand side is much more difficult to evaluate. We cannot look into the wallets of potential buyers and read their minds as to how they will use their cash. However, as we saw in Chapter 3, a change in demand can still be analyzed by looking at the divergence between the Large Effective Ratio and the price. A strong positive divergence indeed indicates that an unusual accumulation by large players is taking place. The accumulation itself originates from an increase on the demand side. If we catch that demand increase early enough, and if the supply analysis shows that only a few sellers are present, a price increase will probably occur. As we can see, the supply/demand equilibrium can be measured through the use of two different tools: the supply analysis tool and the divergence analysis tool.

Supply Analysis Tool Before explaining how the supply analysis tool works, it is important to come back to the impact of the evolution of the volume/price repartition (histogram) to understand the dynamics of stock trading. Let’s look at

192

VALUE IN TIME

FIGURE 4.4 IMAX: stock price evolution at the end of 2006. Source: Chart courtesy of StockCharts.com.

an example. In Chapter 2, I presented the case of the IMAX Corporation, which on August 10, 2006, saw its stock price cut in half overnight. In Figure 4.4, I’ve pointed out three specific days that correspond to different sets of volume/price histograms. These three sets of histograms are shown in Figure 4.5a, b, and c: 1. August 9 was the day prior to the big sell-off. The histogram is repre-

sented in Figure 4.5a. We can see that as of August 9, all of the 40.21 million shares that constitute the total issued shares of IMAX had been exchanged at prices ranging between $8.17 and $10.92. During the last 10 days preceding August 9, 545,000 shares on average had been exchanged daily. 2. August 21 was the day that ended the big sell-off. During the eight trading days between August 10 and August 21, 22.6 million shares exchanged hands, or about 2.8 million shares per day (four times the daily average of the 10 days preceding August 9). Figure 4.5b shows the volume histogram as of August 21. As we can see, bargain investors appeared and created a new group of shareholders. Bargain investors are typically investors who compare the present price to a past price and invest “because the stock is now cheaper.” 3. September 5 was the day that ended the second sell-off. During the 10 trading days between August 21 and September 5, 11.6 million shares exchanged hands, or about 1.16 million shares per day. Figure 4.5c shows the volume histogram as of September 5. As we can see, value investors appeared and created a third group of shareholders. After a price drop, value investors typically wait longer than bargain investors

Supply and Demand

193

FIGURE 4.5a IMAX: volume histogram, August 9, 2006.

do to start purchasing shares. Typically, value investors do not react to a price drop, but are attracted by the valuation of the company in terms of its price-earnings (P/E) ratio. In November 2006, the average shares exchanged per day fell to about 350,000 shares, indicating that the double waves of selling indeed killed the demand by antagonizing all the long-term shareholders. The first lesson that we may learn about the analysis in Figure 4.5a, b, and c is this: When a company goes through a catastrophic situation that instantly affects all the shareholders, the probability that the company will lose its active shareholders is very high. Usually, when active shareholders

FIGURE 4.5b IMAX: volume histogram, August 21, 2006.

194

VALUE IN TIME

FIGURE 4.5c IMAX: volume histogram, September 5, 2006. The evolution of the volume histogram over the course of time indicates the positions of the different groups of shareholders and is the starting point in analyzing availability of shares. In Figure 4.5a, before the first sell-off, the last 40.21 million shareswere exchanged at prices ranging between $8.17 and $10.92. In Figure 4.5b, a large part of these shares were sold and purchased by a group of bargain hunters at prices ranging between $5.45 and $6. Note at the right of Figure 4.5b that the initial group of shareholders whose shares are shown in Figure 4.5a shrank and was replaced by the group of bargain hunters. After the second sell-off, the leftover of the initial group of shareholders shrank further and was replaced by a group of value investors, as shown in Figure 4.5c.

are caught in such a bad position, they take their losses and leave the company forever. In the case of IMAX, bargain investors who were looking for a reversal replaced the departing shareholders. However, that reversal did not materialize, and continuous bad news forced both traditional shareholders and some of the new bargain investors to sell at a loss. A new set of shareholders appeared: the value investors. Of course, after August 21, it is hard to know if the price really represented the value of the company. Nobody could tell if more bad news was on the way. As we saw in Chapter 2, to determine whether a share at a given time represents value, you have to determine the probability that you will be able to sell it later on to someone else at a higher price. To increase your chances of finding value, you must find a buying price at which there will be very few sellers (the price will be so low that few are willing to sell at that price). At the same time, you also need to find buyers other than you who will push the price higher. As shown in Figure 4.6, large players have been keenly selling shares during the different sell-off phases. In addition, no accumulation by large players can be seen at any time, indicating that demand for the IMAX stock completely dried up. Dried-up demand is not a place to find value, even if the price itself is cheap compared to past prices.

Supply and Demand

195

FIGURE 4.6 IMAX: Effective Volume analysis. After a large price drop, the Large Effective Volume will indicate if large players find value in the company. If they don’t, there is no specific reason to buy at this point.

The second (and more important) lesson that we may learn from the example is this: Among the group of shareholders who follow a company, the set of shareholders effectively invested in the company stock is, at any given time, constantly changing. In the case of IMAX, the change occurred at a very quick pace, but this evolution in shareholders’ positions is occurring for all companies’ stocks, admittedly at a somewhat lower speed. It is because of this evolution that the supply/demand equilibrium is continually evolving.

Where Does the Supply Come From? In this section I show how the level of supply constantly evolves and how monitoring it can lead to successful trading decisions. But first, we need to answer a very simple question: Where does the supply come from?

196

VALUE IN TIME

FIGURE 4.7 Tellabs: stock price leading to September 21, 2006.

This is quite simple: The supply comes from shareholders who are selling their shares. Why would people want to sell their shares? Let’s have a look at Figure 4.7, which represents the share price evolution of the company Tellabs for the last 448.5 million shares. This means that between June 26, 2006 (the beginning of the graph) and September 21, 2006 (the end of the graph), 448.5 million shares were exchanged. This is exactly the number of issued shares of Tellabs. The dotted line separates winners from losers, measured as of the price of the last trading day. Those who purchased their shares at a price lower than the closing price of September 21 are earning money. The others are losing money, at least on paper. To try to analyze the possible behavior of shareholders who purchased the Tellabs shares in the past, let’s turn to Figure 4.8. Figure 4.8 represents

FIGURE 4.8 Tellabs: shareholders’ profits/losses compared to stock price on September 21, 2006.

Supply and Demand

197

the profits or losses on September 21, 2006, that shareholders who bought in the past are experiencing, compared to the closing price at the right of the graph. As you may have noticed, Figure 4.8 is the inverted image in Figure 4.7. I have pointed out four zones of interest on which I will comment. For each zone, we have to ask ourselves: Are the shareholders who bought during this price zone ready to sell their shares? What is the probability that they will consider selling their shares on the next trading day? If nobody is ready to sell, then on the next day any buying move will push the price higher. 1. Zone A represents the profit/loss of the most recent buyers. These

recent buyers are experiencing a profit ranging from 0 percent to 15 percent. I do not believe that people who recently bought and are experiencing a profit of between 0 percent and 5 percent would sell. They will probably wait. However, those who are experiencing a profit of between 5 percent and 15 percent are more likely to be ready to sell. I believe that their readiness to sell is proportional to their profit. Let’s assume that those who are experiencing a 5 percent profit are not ready to sell, that those who experience a 15 percent profit are entirely ready to sell, and among those who experience, for example, a 10 percent profit, only half are ready to sell. This means that at a profit of 15 percent or above, all the recent buyers would be potential sellers. This does not mean that they would be certain to sell. It means that the probability of finding sellers among shareholders experiencing a profit of 15 percent or more is higher than among those experiencing only a 5 percent profit. 2. Zone B represents the profit/loss of shareholders who bought earlier than the most recent buyers. Zone B shareholders are experiencing between 15 percent and 20 percent profit. There are two interesting characteristics about zone B shareholders: r The first characteristic is that many of the zone B shareholders who bought at that time have probably already sold their shares on the way to the current share price. Some of them indeed took their profits between 5 percent and 15 percent. We may therefore say that the further back we look into past buyers, the fewer shares will be found for sale, since many of these past buyers would have sold earlier. r The second characteristic of past buyers compared to more recent ones is that those who did not sell at an early stage are probably looking for higher returns. 3. Zone C represents the profit/loss of still previous shareholders who bought earlier than the more recent buyers of zones A and B. Among this group, shareholders who regret not selling before the price drop of

198

VALUE IN TIME

July 27 are now relieved that the price is coming back into their buying range, and may therefore be happy to sell their shares at around their buying price. 4. Finally, Zone D investors, if they did not cut their losses at the beginning of the price descent, are still locked in and will probably not sell before the price rises again. In Figure 4.9, I show zones A, B, C, and D, but in terms of number of shares exchanged during the corresponding periods. You can now see that for each zone, it is possible to count the number of shares that were exchanged and calculate a probability that these shares will be sold. The total for the four zones forms the number of shares that could eventually become available for sale on the day following September 21, 2006. In fact, a true mathematical model is more complex, because we have to consider that different price zones often overlap. Also, active traders would usually sell more quickly than long-term investors. Therefore the real mathematical model that I use separates both types of shareholders. Although the model is quite complex, it simply counts the number of shares that are available for trading and then weights this number with a probability that the shares will be offered for sale. Such a probability is calculated using mathematical expressions that factor in the selling pattern distribution compared to the actual profit/loss and the delay since the purchase. Let’s see how the model works on a few real-life examples.

Practical Examples I would first like to come back to a comment that I usually read in traders’ journals: “You must buy when everyone else is selling.” For me, this is a

FIGURE 4.9 Tellabs: volume histogram as of September 21, 2006.

Supply and Demand

199

sure recipe for financial disaster. There are only two clear times when you should buy: 1. You buy when everybody else is buying, but you do it early in the trend. 2. You buy when everybody else has stopped selling. In other words, you

buy when the supply of shares has dried up, when only a few shares are available for sale. From what we learned in the previous section, it is obvious that at a specific point of time, the only way to correctly evaluate the availability of shares is to count the number of shares that have a good probability of being made available for sale at that point in time. Supply analysis is important not because it allows us to discover value, but because it allows us to detect the right timing for when the selected stock will be prone to rise. This price rise will occur if large players decide that the stock price is compelling enough for them to take a significant position. In other words, we need to wait to see that value is recognized by large players, being sure that our floor is protected by a lack of potential sellers. The Case of Reliant Energy (RRI) The lower panel in Figure 4.10 shows a measure of the supply of the Reliant Energy stock (refer also to Chapter 2, Figure 2.19a and b). You can see that the supply signal is very sensitive to small price variations: A price increase of a few percentage points could potentially attract a large number of new sellers, depending on their relative profits. Because that sensitivity could give misleading trend signals, I smoothed the signal in the lower panel by using a four-day simple moving average. As can be seen in the lower panel in Figure 4.10, I arbitrarily drew two horizontal lines that separate the chart into three interesting supply zones: 1. The low supply level is set for a supply of less than 10 percent of the

total number of issued shares. At such a low supply level, it is almost assured that any significant buying activity from a large fund will result in a price increase. Do not forget that we can detect the large funds’ buying patterns through the Effective Volume tool. However, a low supply level does not ensure that the price will not continue to fall. Indeed, if the prospects for the company are very bleak, no new investor will invest and any small selling pressure will push the price further down. The low supply level determines a buying zone. (After I wrote this chapter, I did a sensitivity analysis on the supply level in

200

VALUE IN TIME

FIGURE 4.10 Reliant Energy, supply analysis. The supply analysis will help to point out unusually low supply patterns, where any significant purchase by a large fund will put pressure on the price to increase. Above the medium supply level of 20 percent, shares are available for sale in high enough quantities to satisfy a large fund. Hence, the buying by large funds above the medium supply level may not necessarily result in a price increase.

terms of trading performance; I found out that the optimum low supply level is not 10 percent, but rather between 5 percent and 7 percent. This sensitivity analysis is explained in Chapter 6.) 2. The medium supply level is set for a supply that is between 10 percent and 20 percent of the total issued shares. Twenty percent is

Supply and Demand

201

relatively low compared to the total number of issued shares, but it is already twice the low supply limit of 10 percent. By experience, this 20 percent level is the maximum limit at which one can say that there is a potential share availability problem. Do not forget that if there is a lack of shares available, any fund that wants to take a significant position will need to push the price higher to attract the new supply. 3. Above the 20 percent limit, we can say that a fund will be able to take a significant position without having a real impact on the price. In Figure 4.10, I have also pointed out two zones of high price, zone A and zone B. Although these zones correspond to high supply zones, it is difficult in my experience to establish a correlation between the possible selling pressure on a stock and the supply level. In other words, a 70 percent supply level does not offer a much stronger selling pressure than a 50 percent supply level. The Case of Tellabs (TLAB) Tellabs also shows two interesting low supply zones where buying would have led to a significant profit. (See Figure 4.11; refer also to Chapter 2, Figure 2.7a and b.) In this example, it is also important to note that the peaks of supply do not necessarily correspond to selling points. A peak of supply simply indicates that shares could potentially be made available for sale. However, what will trigger a selling move is a price pullback. Indeed, when supply is high, it means that most shareholders are turning a paper profit of 15 percent or more. A price pullback will attract these shareholders into selling, in order for them to protect their profits. This selling pressure will be either a temporary pullback or a real new downtrend. Once again, as we saw in Chapter 3, the Effective Volume tool and divergence analysis will give you a much clearer picture of where you stand. In the lower panel of Figure 4.12, I indicate the buy zone 2, as calculated by the supply analysis tool represented in Figure 4.11. The upper panel in Figure 4.12 indicates that the downtrend 1 in Large Effective Volume was preventing us from buying the stock, even at a low supply level. It is only when large players moved in (as indicated by the small uptrend 3) that the trading rules allowed us to start buying. The Case of Openwave Systems (OPWV) Figure 4.13 shows how dangerous it can be to trade only on the basis of the supply signal. (Refer also to Chapter 2, Figure 2.16a and b.) You can indeed see in the lower panel that the two correct buy signals (buy zone 1 and buy zone 2) are followed by an incorrect buy signal (buy zone 3). A catastrophic drop in the stock price of Openwave Systems triggered this incorrect buy signal.

202

VALUE IN TIME

FIGURE 4.11 Tellabs: supply analysis. This example clearly shows that the peaks of supply do not automatically correspond to the peaks in prices. Also, notice that the buy zone 2 came too early in the downtrend. This shows that other tools such as the Effective Volume tool are necessary to select the best entry point.

Unfortunately in this case, the Effective Volume could not prevent us from entering during the Large Effective Volume uptrend B in the upper panel of Figure 4.14, just before another price collapse (downtrend C in the lower panel). This is the reason I also use the rule to buy when the price is above its 9-day simple average. This avoids being trapped into most of the catastrophic situations.

Supply and Demand

203

FIGURE 4.12 Tellabs: Effective Volume analysis. The Effective Volume tool is an excellent complementary tool to the supply analysis tool.

One of the problems with the Large Effective Volume analysis is that we cannot discern if a buyer is buying the new-found value or if he is simply covering a short position. We all know that large down-trends attract shortsellers. The owner of a share is always more eager to take action (sell) than someone who does not own it (yet who would be eager to buy it). Therefore it is probable that after a long down-trend, the first buying signs will be due to shorts covering their position. As we will see in Chapter 6, this is the reason why we need to look at the Large Effective Ratio signal that we studied in Chapter 3. Indeed, the Large Effective Ratio allows us to compare the present share accumulation to past accumulations, and this comparison allows us to judge the real strength of the buying movement: short covering must be accompanied by genuine buying in order to produce a significant signal on the Large Effective Ratio tool.

204

VALUE IN TIME

FIGURE 4.13 Openwave Systems: supply analysis. The supply analysis cannot distinguish between a low supply signal that came from a normal pullback in price and a low supply signal that happened because of a large catastrophic reversal. Other tools such as Effective Volume or the Effective Ratio are necessary to avoid untimely entries.

The supply analysis signal is a very useful tool that gives unexpected (but good) results when used in combination with other tools. We will see in Chapter 6 how this signal can be used in a successful trading strategy. Before going into the real world of how to make money, it is important for us to briefly study how funds get in and out of investment positions, since they are the ones that provide liquidity to the markets.

Supply and Demand

205

FIGURE 4.14 Openwave Systems: Effective Volume analysis. Figure 4.14 shows that the Effective Volume analysis does not always prevent wrong calls. The Large Effective Volume issued a buy signal inside of buy zone 3 indicated by the supply signal. Both signals proved wrong later on.

FUNDS’ STRATEGIES We saw in Chapter 1 that large players have a critical influence on the price direction, because funds have a propensity either to trade a large volume in the same price trend or to force a trend change. However, if trading a large volume can give a large player real force to move the market, it is also a giant weakness. Trading volume implies that a counterparty exists: If you want to buy one million shares, you need someone ready to sell them to you. If the market is illiquid, large sales or purchases could have an impact on the price. This is the main

206

VALUE IN TIME

reason there are parallel markets that allow funds to exchange shares among themselves. For a retail investor, markets are fundamentally liquid. Retail investors can indeed buy and sell shares at will without worrying about the influence their trading will have on the share price. For funds, however, markets are in a constant illiquid situation: The size of the positions that funds must take is so large that they often need many days to load or unload positions. The supply/demand analysis is therefore critical for funds. We already know that managing a large position will have a significant impact on the selection of tactics, depending on whether the fund wants to buy or sell shares. Suppose that a fund wants to buy from the market five million shares of a company trading at $10. Suppose also that the average number of shares traded is one million per day. If the fund does not want to buy more than 10 percent of the daily volume, it can only purchase a maximum of 100,000 shares per day, and its accumulation tactics will be carried out over 50 days. The Effective Volume tool will surely detect such tactics. However, such an accumulation could at some point start to have an impact on price. Indeed, since the continuous buying could change the balance between supply and demand, supply could dry up, forcing the stock price to increase. Assume, for example, that after the purchase of three million shares, the price starts to increase, and is increased by 10 percent within a few days. Even if the fund did not get its targeted five million shares, it is already showing a paper gain on the three million shares that have been previously accumulated. The fund manager could then elect either to stop the share accumulation or to continue to accumulate at a higher price at the ask. This would push the price even higher and signal to the market that “a buyer is in town.” A new uptrend could then be triggered to the funds’ advantage. In summary, buying at the ask will work in the fund’s favor as long as the fund has already accumulated a position in the stock. Let’s now suppose that the fund needs to sell five million shares on the market. It would be foolish to dispose of these shares in large blocks sold at the bid. This would certainly push down the price and the fund would incur an instant paper loss on its remaining shares. Therefore, selling a large position into the market takes a lot of time and is more effectively achieved by increasing the shares offered at the ask without pushing the price down, or by selling small quantities of stocks at the bid without totally erasing the supply of shares. In summary, selling blocks of stocks at the bid will work in a fund’s disfavor if the fund still holds a large number of shares. The fund will be better off making sure that the bid/ask balance is not disturbed.

Supply and Demand

207

The question of supply is therefore critical for allowing large funds to execute their moves. Funds need to start selling while the price is still moving up. If they wait too long, they could be forced to sell their large positions into a downtrend, when the market is crowded with traders wishing to sell. Position size has another important negative effect: Traditional funds need to take large positions, but they cannot use stop losses on these large positions and therefore need to diversify to a great extent. Their extreme diversification does not allow them to profit from good investment decisions in any significant way—hence the poor return of pension funds. Let’s look at a few very tough situations that are in fact very common occurrences. First, let’s have a look at Table 4.1. Every quarter, institutional players have to report their buying/selling activity for all the securities in which they are active. These figures are readily accessible on the NASDAQ web site (www.nasdaq.com). In Table 4.1, I’ve compiled the figures related to the eight companies for which we studied the Active Boundaries indicator in Chapter 2. The first and the second columns represent the number of outstanding shares and the percentage of institutional holdings, respectively, as reported by the NASDAQ web site at the end of June 2006. The third and fourth columns show, respectively, the number of shares that were traded during the prior quarter and the institutions’ position changes as reported by the NASDAQ. Finally, the last column gives a measure of the institutional activity during the last quarter. It is not entirely correct, because the institutions’ position changes are normally lower than their real activity. Indeed, if an institution buys one million shares during the quarter and sells it before the quarter ends, these two million shares will not be reported. However, the last column gives you a good sense of the importance of institutional activities for each stock. Table 4.2 shows the detail of institutional activities. It is interesting to see that even though the global institutional activity is significant, the net activity itself is small in percentage points compared to the number of shares exchanged. For example, we can see that for Chico’s FAS, the net activity was only –7.6 percent of all the shares exchanged during the quarter ending on June 30, 2006. As a comparison, the last column shows the price change during the same period of time. You can see by comparing the last two columns of Table 4.2 that there is no clear correlation between a price drop and net selling activity by institutions, or between a price increase and net buying activity. Let’s take a closer look at Openwave Systems. During the quarter ending on June 30, 2006, institutions had been sellers of 30 million shares and buyers of 24 million shares. During that period, Deutsche Bank AG was the largest buyer. This German bank increased its position in Openwave

208 Number of Shares Outstanding (Million)

19 448 94 87 176 308 245 40

Company

Envoy Communications Group Tellabs, Inc. Openwave Systems Inc. Meridian Resource Corporation Chico’s FAS, Inc. Reliant Energy, Inc. Becton, Dickinson and Company IMAX Corporation

4.3% 72.8% 87.7% 43.9% 73.3% 89.2% 84.1% 67.5%

Percentage Institutional Holdings

5.5 383.0 183.8 37.0 210.0 150.5 56.0 19.5

Number of Shares Exchanged (Million)

TABLE 4.1 Institutional Activity for the Quarter Ending June 30, 2006

0.4 83.3 54.0 12.5 82.0 59.4 25.2 11.9

Position Changes by Institutions (Million)

6.6% 21.7% 29.4% 33.8% 39.0% 39.5% 45.0% 61.0%

Institutional Activity as Percentage of Shares Exchanged

209

Position Changes by Institutions (Million)

0.4 83.3 54.0 12.5 82.0 59.4 25.2 11.9

Company

Envoy Communications Group Tellabs, Inc. Openwave Systems Inc. Meridian Resource Corporation Chico’s FAS, Inc. Reliant Energy, Inc. Becton, Dickinson and Company IMAX Corporation

0.05 42.30 24.00 7.60 33.00 34.60 10.20 5.40

Position Increased by Institutions (Million)

0.3 41.0 30.0 4.9 49.0 24.8 15.0 6.5

Position Decreased by Institutions (Million)

−0.3 1.3 −6.0 2.7 −16.0 9.8 −4.8 −1.1

Net Activity of Institutions (Million)

TABLE 4.2 Institutional Position Change for the Quarter Ending June 30, 2006

−4.9% 0.3% −3.3% 7.3% −7.6% 6.5% −8.6% −5.6%

Institutional Net Activity as Percentage of Shares Exchanged

4.6% −17.0% −46.7% −13.1% −32.7% 13.2% −0.6% −11.0%

Price Change during Last Quarter

210

VALUE IN TIME

Systems by 180 percent, adding 2.6 million shares. During that quarter, the stock fell from $22 to $11. Suppose that Deutsche Bank AG bought at $16.50 on average. Since later on the stock price dropped to $6.10, we can easily calculate that Deutsche Bank AG lost close to $10 per share, or $26 million of paper loss. It may sound very large, but if we compare it to Deutsche Bank AG’s total portfolio value of $168.4 billion, it is a loss of only 0.0154 percent. Let’s also have a look at Chico’s FAS. During the quarter ending on June 30, 2006, institutions had been sellers of 49 million shares and buyers of 33 million shares. TIAA-CREF Investment Management LLC was the largest buyer during that period. This fund increased its position in Chico’s FAS by 360 percent, adding 4.9 million shares. During that quarter, the stock fell from $40 to $27. Suppose that TIAA-CREF Investment Management bought at $33.50 on average. Since later on the stock price dropped to $18, we can conclude that TIAA-CREF lost close to $15.50 per share, or a $76 million paper loss. Compared to the total portfolio value of $127.7 billion, it is a loss of only 0.059 percent. Both institutions that are singled out run more than 4,000 positions simultaneously, meaning that they are very well diversified. As a matter of fact, they both manage a wide range of funds in different market sectors. A loss of $10 million on a position is probably compensated by a gain on another position. Diversification is important if you cannot use stop-loss strategies to get out of a losing position quickly. This is the case when you build a large position in a single stock, which is what funds do. However, because of their very large diversification, none of the large funds can ever expect to beat the markets on a consistent basis. They will gain when the market increases and they will lose when the market decreases. The main goal of diversification is risk control; the drawback is a lack of significant return. This is usually the case with pension funds, which follow strategies that are even more defensive. The press is very fond of going after “the big, ugly market manipulators and rogue traders.” Everybody has read about such-and-such hedge fund going under because of the very large positions that were taken against a nonconsenting market. I am in no position to praise or criticize; however, the reality is that managing a large fund is a very difficult task. That task is often more about managing a sizable position than it is about the right entry and exit timing. I do not really sympathize with the two fund managers who had to go through $26 million and $76 million of paper loss because they bought into a falling stock price. However, I think that fund managers who can consistently beat the market managing a large fund deserve our respect. Now, wait a minute! Didn’t I write in Chapter 1 that large players are the ones responsible for price movements? In such a situation where even

Supply and Demand

211

large funds can routinely lose large amounts of money, is it worthwhile to follow the moves of large players? Yes, it is! At least if you know what you are looking at. I believe that the terms active and passive are indeed accurate for describing different types of funds. Passive funds invest in baskets of shares that follow longterm sector trends. Active funds follow specific investment strategies that involve both fundamental and technical analysis. Active funds have a tendency to carefully select the stock in which they are invested. The best active funds have key analysts who spend quite a large amount of time studying the fundamentals of companies or checking the sales channels. Active funds invest both in blue-chip companies as well as in midsize ones. However, if you are a retail investor who wants to follow large players, you are better off investing mainly in midsize stocks. Indeed, blue chips are traded by both active and passive funds. This makes it more difficult to distinguish Effective Volume movements triggered by active funds, since many passive funds will also trade large volumes. Passive funds seldom invest in midsize companies that are not part of the main stock indexes, because they would carry considerable volatility. Following the moves of large funds in these midsize companies provides many opportunities for profit.

FUNDS AND MARKET MANIPULATION Playing the market is very difficult for funds. By contrast, look at the many key advantages that a retail investor can enjoy. Retail investors:

r Are always eager to study and gain knowledge, because it is their money on the line.

r Can wait for a trend to start before entering a position. r Can quickly enter and exit positions. r Can concentrate on a few good plays. In such a difficult environment for funds, the next natural question is: How do funds make money? Maybe it’s just an innocent question, but after what we learned in the previous section, it is quite natural to wonder how these traditional funds still give an acceptable return to their investors. We all know that hedge funds use specific strategies that involve complex financial instruments in related markets. The question is more applicable to traditional funds: How do they make money? Funds that trade successfully find pools of temporary illiquid situations from which they can

212

VALUE IN TIME

profit. The most common illiquid situations are those explained earlier in this chapter, where the price of a stock has decreased more than what the underlying fundamental value dictates, but where the strong decline has locked most investors into the stock. The supply analysis method described earlier in this chapter gives a good general view of the supply situation. However, one question has repeatedly come to my mind: When a large fund accumulates a stock that is in an illiquid situation, can this accumulation be conducted without a manipulation that would keep the price low or cheap enough? Is it possible for a fund to accumulate a significant number of shares on the market without inducing a price increase? In other words, can funds manipulate the market? Do they need to manipulate markets in order to make a profit? These are not innocent questions, since we know from Chapter 1 that major news is already shown in the Effective Volume signal before it becomes public. This means that insider trading is a standard practice, not an exception. My next innocent (but very scary) question, then, is: Is stock price manipulation also a standard practice? In the Introduction, I discussed the decimalization revolution that rendered the Effective Volume method possible. If you remember, I wrote: “Decimalization killed market visibility and, as some believe, may have encouraged price manipulation.” What does this mean?

r Market visibility. Decimalization lowered the price spread by a factor of 6.25. This greatly reduced the disadvantage of placing market orders (buying at the bid or selling at the ask) compared to placing limit orders (placing a buy order at or below the bid or a sell order at or above the ask). Large players therefore simply stopped placing large limit orders and instead placed repetitive and fragmented market orders. r Price manipulation. Since the price spread reduction caused by the decimalization eliminated the incentive to place limit orders, it became theoretically possible for a large fund to place a large buy market order followed by a small sell market order that would push the price back down. Before decimalization, price manipulations were quite difficult since there were large buy and sell limit orders at each bid and ask position, respectively. Because we’ve covered the different concepts of the book in so much detail, it is now time to come back to these bold statements regarding market visibility and price manipulation. To address these issues, we need to study the resistance to change that I introduced at the beginning of this chapter. The resistance to change is defined as the market’s natural reaction to a price change in a specific direction. To measure a resistance, we must naturally first get a change.

Supply and Demand

213

Let’s look at the resistance to a buying pressure. We saw that the Effective Volume is, by definition, the volume responsible for small price changes. When large players buy a stock, they must fight the resistance of sellers who will place market sell orders. They must also fight against the resistance of sellers who placed or will place limit sell orders. I call those who place market sell orders active sellers and those who place limit sell orders passive sellers. The Effective Volume can be separated into two components: the positive Effective Volume and the negative Effective Volume. The negative Effective Volume will appropriately measure the active selling force. However, how can we measure the passive selling force? This force is clearly to be found in the “Non-Effective Volume.” This Non-Effective Volume is the volume that has no impact on price changes. I briefly inserted it into the divergence analysis (presented in Chapter 3) by using the Effective Ratio. If you remember, the Effective Ratio measures a change in the ratio of the Effective Volume to the total volume. This total volume includes both the Effective and the Non-Effective Volume. However, if you remember the details of Chapter 3, the level of the Effective Ratio usually stays within a few percentage points of the total exchanged volume. This means that the equilibrium between buyers and sellers is rather thin. I have been wondering, then, if the divergence we often see between the Effective Ratio and the price rate of change cannot also be influenced by the passive players. For example, suppose that a large fund places a very large number of shares for sale at the ask. This means that these shares will simply wait for buyers to come in, and it is clear that the price will not move up until the whole ask has been bought out by these buyers. Because of the large volume at the ask, the only way for small sellers to sell their shares is to sell them at the bid, therefore lowering the price by one cent. However, if there is active buying on the stock, midsize buyers will come buy the stock available at the ask, increasing the price by one cent on midsize volume. In this example, this means that the Effective Volume pattern will be an increasing Effective Volume; it represents the active buyers together with a flat price trend that is formed because a price limit was set by the large volume at the ask. The same pattern would also appear if a large seller were to send a regular number of shares for sale at the ask. This is an extreme case, of course, because I have not experienced such behavior yet, but it may be worthwhile trying to measure it. Recall that traditional technical analysis tools linking volume to price usually weight volume by some price spread. The Effective Volume theory says, however, that since we are at the microscopic level, the price spread itself is not important. The mere fact that the price changed from one minute to the next is what is important. Therefore, the theory says that

214

VALUE IN TIME

if 10,000 shares move the price by one cent, it has the same importance as if the price had moved by five cents. Indeed, in terms of Effective Volume, if the 10,000 shares move the price up by five cents, it does not mean that these shares are five times stronger than if the price had moved by only one cent. It just indicates that someone has been buying. It also indicates something different, though: The price moved up five cents not because the 10,000 shares were exceptionally strong, but simply because the supply of shares was not enough to offset this buying push of 10,000 shares. This gives us a clue about how to measure the static resistance to change. Indeed, we have learned that a positive Large Effective Volume is what really moves the price up. For example, we saw in Table 3.6 in Chapter 3 that for Darden Restaurants, the Large Effective Volume was pushing the price up or down by almost four cents on average, while the Small Effective Volume was moving the price by only a little more than two cents on average. Therefore, to measure the resistance to the buying power of large players, we only need to measure for a fixed period of time the size of positive Large Effective volume necessary to move the price up by one cent. Indeed, if large players need fewer shares to push the price up by one cent on average, it means that the resistance from passive players is decreasing. The calculation method must use the following steps: 1. Take an analysis window of your choice (usually the same window as

the one used for the divergence analysis. In the case in Figure 4.15, I took 3.3 days). 2. Add all the positive Large Effective Volume during the analysis window. 3. During the analysis window, add all the price increases that occurred during the price inflections linked to the positive Large Effective Volume. 4. Divide the positive Large Effective Volume obtained in step 2 by the result obtained in step 3. 5. Move the analysis window by one minute, redo the calculation, and

plot the results as in Figure 4.15, which represents the resistance to change. Figure 4.15 must be compared to Figure 4.16, which represents both the Large Effective Volume flow and the price during the same analysis period of 100 trading days. When looking closely at Figure 4.15, we can see a strange pattern that does not make much sense: The start of the buying activity by large players,

Supply and Demand

215

FIGURE 4.15 Tellabs: number of positive Large Effective Volume shares necessary to move the price up by one cent on average.

FIGURE 4.16 Tellabs: Large Effective Volume versus price analysis.

216

VALUE IN TIME

FIGURE 4.17 Tellabs: resistance to buying versus resistance to selling.

around December 2 (point A of Figure 4.16), corresponds to a strong resistance (R1 in Figure 4.15). Then, from December 2, it seems that the resistance progressively slows down (down arrow from R1). This is very counterintuitive: If the price of a stock increases (up arrow from point A in Figure 4.16), more passive sellers should normally appear and the resistance should increase. Why do we see the opposite? Since I did not have a good explanation, I decided to do the same exercise by measuring the resistance to the selling activity by large players. I then depicted those results with the results of the resistance to the buying activity by large players (see Figure 4.17). Figure 4.17 is rather interesting: It shows that the resistance to selling pressure is following the same pattern as the resistance to buying pressure. That goes against all common sense, so I conclude that passive resistance is negligible compared to active resistance. Indeed, when you look at the 3.3-day pattern of positive Large Effective Volume and you compare it to the negative Large Effective Volume pattern, you get the results shown in Figure 4.18a. You can see that the positive Effective Volume and the negative Effective Volume are pushing the market in opposite directions at the same time:

r When large buyers are getting stronger (arrow is up on the positive Large Effective Volume), large sellers are also getting stronger (arrow is down on the negative Large Effective Volume). r When large buyers are getting weaker (arrow is down on the positive Large Effective Volume), large sellers are also getting weaker (arrow is up on the negative Large Effective Volume).

Supply and Demand

217

FIGURE 4.18a Tellabs: positive and negative Large Effective Volume.

In practical terms, this means that the markets are very efficient: Whenever a large player buys enough shares to slightly move the price up by a few cents, another large player will sell almost an equivalent number of shares that will bring the price back to its previous position. In light of my statement that decimalization killed market visibility while favoring price manipulations, I now need to comment on market visibility and price manipulation.

r Market visibility. The fact that it is very difficult to measure the passive buying or selling resistance levels shows that limit orders are not an important component of the general resistance to change compared to active resistance. This means that the order book does not influence the market and does not give any visibility to the direction of the balance between buyers and sellers. Decimalization indeed killed visibility. r Price manipulation. What Figure 4.18a and 4.18b tells us is that price manipulation is virtually impossible due to the very fine balance between positive and negative Large Effective Volume. Indeed, with such a balance, buying or selling of small amounts of shares is not likely to push the price up or down and therefore is not a good means for controlling the share price during accumulation or distribution. We can therefore conclude that due to market forces, price manipulation is not a possibility, at least for those stocks that trade significant volume. In order to confirm my findings quantitatively, I ran a detection algorithm on the minute data on many stocks that were in a price trading range, but could not find any mathematically significant discrepancy that could be explained by price manipulation.

218

VALUE IN TIME

FIGURE 4.18b Tellabs: Large Effective Volume balance. In a trading range, the fine balance between the positive and negative Large Effective Volume shows that markets are efficient for very liquid stocks, and that therefore no significant price manipulation can take place.

WHAT WE LEARNED REGARDING THE SUPPLY ANALYSIS This chapter concludes the set of new concepts that I wanted to introduce:

r r r r

Chapter 1: Effective Volume. Chapter 2: Active Boundaries. Chapter 3: Effective Ratio and divergence analysis. Chapter 4: Supply analysis.

We started this chapter by learning how important it is to measure the level of the supply of shares, and how to effectively perform such a measure. We then moved on to study how funds must play in an illiquid environment, and learned that funds have great difficulty making money primarily because the size of the position a fund must take is so large. We also discovered that markets are very efficient and that therefore price manipulation is not a likely scenario for funds. This leads us to the next section of the book: how to make money, not only for retail players, but also for traditional funds that need to manage large positions. Indeed, we now understand how the market really works, and we know which tools can show us how it works. What we do not have yet is a system that allows us to make money. It is a trading system that we need, one that will constantly generate a profit. In Part Two we will first review in Chapter 5 how to measure the risk/return balance of a trading strategy. We will then move on to the study in Chapter 6 of a variety of successful trading strategies that are based on a combination of the tools presented in Part One.

PART TWO

Trading Strategies

CHAPTER 5

Performance The Risk/Return Balance

t this point in the book, what we have is a set of trading signals. Before going on to the next chapter, which compares different trading strategies, it is worthwhile to take a fresh look at the risk/return balance. Indeed, it makes no sense to compare trading strategies in terms of the risk/return balance if we do not have a clear idea of what we are comparing. “But wait a minute,” you might say. “It is simple: The return is the money you expect to make, and the risk is the money you may eventually lose.” This is entirely true. There are many books that deal with portfolio performance analysis and comparison using the risk/return balance. You can indeed find many tools that allow the ranking of funds within that framework. (In this field, I highly recommend the book Hedge Funds: Quantitative Insights, by Franc¸ois-Serge Lhabitant.) However, the same tools cannot be used at the level of trading strategy, because a trading strategy produces a set of investment opportunities from which the portfolio manager must choose. Therefore, the risk/return rating of a portfolio mixes both the trading strategy’s efficiency and the portfolio manager’s stockpicking skills. In Chapter 6 we will be comparing different trading strategies, so it is important to first understand what we are measuring and how to measure it. In this chapter, I review some of the formulas used to evaluate the performance of a trading strategy in terms of both risks and returns. Some of these formulas are complex and could be misleading, since they use hypotheses that are not always correct. Just because a formula is complex does not mean that it is useful. On the contrary, because the more

A

221

222

VALUE IN TIME

complex formulas do not necessarily reflect a trader’s everyday reality, I prefer to use much simpler performance measures: the yearly expected return (YER) and the monthly loss transferred (MLT) by the trading strategy to the portfolio. The YER is used as a measure for the return of the trading strategy, while the MLT is used to measure risk. A successful trading strategy must not only generate good performance figures, but also be easily integrated into an automated trading system. I liken trading stocks to milking cows. When I was a boy I used to work on my uncle’s farm over the school holidays. My uncle had just bought an automatic milking machine, and he explained that with it he could milk his cows much more quickly. In other words, he got more milk per unit of time with the machine than he could get milking by hand. Have you ever tried to milk a cow by hand? It is very difficult to do: You sit with your head on the cow’s flank, firmly grabbing one teat in each hand. Then you pull on the teats while gently squeezing them to get the milk out, all while the cow continuously hits your head with her tail. Trading stocks without a system is a similarly tedious process. Using a trading system to generate trades is similar to introducing a new milking machine on a farm: We need to evaluate its performance and then try to optimize it. In the farm example, we need to rate the automatic milking machine against what the herd can produce by hand-milking. In the same fashion, when we evaluate our trading system, we also need to rate it against a buy-and-hold strategy executed on the same sample of stocks over the same time period as the one used by the trading system. Furthermore, in order to test the efficacy of a milking machine, we could separate the good milk-producing cows from the rest, to see if, compared to handmilking, the machine gives better performance for each group of cows. We can do something similar with stocks. We can measure what a group of stocks can produce with a buy-and-hold strategy versus some other control group. This will allow us to rate the performance of our trading system. We will see how to do that later in this chapter. The term trading system usually implies that the trading strategy is used in a systematic (almost mechanical) way. However, in this chapter and the next I use the terms trading strategy and trading system interchangeably since I believe that a true trading strategy must be turned into a mechanical system to be objectively effective. Let’s now examine what sorts of risks are inherent in a trading system. In the same way that we should not confuse the risk of a portfolio with the risk of a trading strategy, we should not confuse the risks of the stock market in general with the risk of a trading strategy. The stock market risks are very well known, and measures of market volatility or periodicity of crash occurrences can indicate how risky a market is. In order to lower

Performance

223

our risk, we could select stocks from different sectors, with different beta values. (William Sharpe defined the term beta in his capital asset pricing model theory. Without going into the detail of the formula, the beta value of a stock is a measure of its volatility relative to the asset class. Stocks that enjoy a beta value higher than 1 are more risky than the general stock market, but offer potentially higher returns. Stocks with a beta value lower than 1 are less risky than the general stock market, but offer lower returns.) However, this stock selection strategy has nothing to do with measuring the risk of the trading system.

THE TRADING STRATEGY Let’s first try to answer two basic questions: What is a trading strategy, and what are the objectives of a trading strategy?

What Is a Trading Strategy? A trading strategy is a method that allows the production of a regular flow of trades in a repetitive process. It can be compared to a production line that includes four components: 1. The raw materials. These are the trading signals that allow us to enter

a trade. 2. The tuning parameters. These are the different ways to control exiting of a trade. Stop loss levels, profit targets, and time limit levels are three tuning parameters that I use (the time limit level consists of selling the stock after a certain number of days). 3. The products. These are the trade returns, linked to the number of days that it took to produce each return. 4. The by-products. There are two types of by-products: the risks

linked to each trade and the pain generated by the temporary price drawdowns.

What Are the Objectives of a Trading Strategy? The first objective of a trading strategy is to generate trades that will allow a better return, on average, than the returns produced by a buy-and-hold strategy.

224

VALUE IN TIME

The second objective of a trading strategy is to shield us against a variety of risks:

r The first risk is simply the risk of not being able to reach our expected return.

r The second risk is market risk. The trading strategy must provide good protection against the expected negative impact that a bad market would have on the group of stocks we are following. r The third risk is the trader’s erratic reactions when facing conditions of exuberant joy or excruciating pain. r The fourth risk is simply the trader’s wrong analysis.

Applying the Strategy My objective in this chapter is to help you study the products and byproducts generated by the trading strategy and see how each of the tuning parameters may influence them. We need to understand how the machinery functions so that in the following chapter we will be able to select the best trading strategies and optimize them. It would be very difficult for me to explain the different ideas of this chapter without using practical examples. I have therefore generated as many trades as possible using a sample trading strategy explained next. We will then be able to see by what mechanism and to what extent the different tuning parameters help meet the sample trading strategy’s objectives. For this work, I used the minute data of 159 stocks. As shown in Figure 5.1, the majority of the data covers the years 2005 and 2006, a period that corresponds to two years of positive market trend (see Figure 5.2). Any trading strategy used during that period would naturally produce positive returns and low risks. Finally, I sorted the 159 stocks by performance, collected the 69 worst-performing stocks in a laggards group, the 97 best performers in a highfliers group, and took some stocks from each group to form a standard group (see Table 5.1). Please note that I also included 7 of the worst “highfliers” among the “laggards” group, where they became the best “laggards.” Please note three points: 1. Even though 159 stocks may look like a large sample, it is still relatively

small compared to the thousands of stocks and the different markets on which the method can be used. 2. Data availability dates back only to April 2001, a period that is not long enough to qualify the method during changing market conditions. 3. Past results do not indicate future results.

Performance

225

FIGURE 5.1 Distribution of the sample of stocks used to test the trading strategies.

Although the different trading methods presented in Chapter 6 should be tested on more samples and over a longer period of time, I decided to publish these very promising preliminary results. In this chapter, I use data from only the standard group of stocks. However, in Chapter 6, the optimization process makes use of the three different groups.

FIGURE 5.2 S&P 500 trend during the last six years.

226

VALUE IN TIME

TABLE 5.1 Types of Stock Groups Used for the Analysis

Group of Stocks

Number of Stocks

Number of Trading Days Analyzed

Average Number of Trading Days per Stock

Laggards group Standard group Highfliers group

69 101 97

35,023 41,843 47,584

508 414 491

The sample strategy that I use in this chapter consists of buying a stock whenever the price is higher than its nine-day average, and keeping it unless obliged to sell due to one of the tuning parameters. Since before investing, the price was lower than the nine-day average, the objective of this strategy is to run price uptrends, catch them early enough, and hope that they develop into full long-term trends. The idea of using this very simple trading strategy is to be able to study how the different tuning parameters work, which are separate from the trading signals generated by the strategy itself. The goal of this chapter is to work out the relationship between the different tuning parameters (the stop loss level, the time limit level, and the profit target) and to study how they influence the performance of a trading strategy, in terms of both returns and risks. I believe that the behavior of the three tuning parameters is independent of the trading strategy and is more related to the market cycle.

OPTIMIZING PROFITS Makeup is to beauty as profit optimization is to trading strategies. Makeup on its own cannot turn a homely girl into a beauty queen; it only enhances already attractive features. Similarly, profit optimization only works with a trading strategy that is already robust. Let’s first define what we call profit and then see how to enhance it.

How to Measure the Return of a Trading Strategy Running the sample strategy will produce a set of independent trades such as those shown in Table 5.2. The buy order would be triggered by the trading signal produced by the sample trading strategy, while the rightmost column indicates the four reasons for selling the stock: reaching a stop loss level, a profit target, a time limit level, or a sell signal generated by the indicators.

227

Performance

TABLE 5.2 Trades Produced by a Trading Strategy Symbol

Buying Date

Buying Price

Selling Date

Selling Price

Reason for Selling

KG RHT BDX CTSH UPL HAL CVO CTSH CME CDNS

10/25/2006 10/24/2006 10/20/2006 10/18/2006 10/10/2006 10/6/2006 10/4/2006 9/29/2006 9/27/2006 9/26/2006

$16.94 $19.42 $72.14 $77.41 $49.16 $27.68 $19.02 $74.01 $479.35 $16.82

12/15/2006 10/26/2006 12/26/2006 11/8/2006 12/18/2006 11/15/2006 12/26/2006 10/5/2006 11/17/2006 10/26/2006

$16.31 $14.83 $70.57 $77.12 $48.07 $33.57 $19.85 $75.44 $534.80 $17.79

Time limit Stop loss Time limit Sell signal Time limit Profit target Time limit Sell signal Sell signal Sell signal

You do not need more information than what is stated in Table 5.2 in order to measure the return of a trading strategy: You only need to know how much you are winning or losing and the number of days during which you held your shares. There is a fascinating book written by Ralph Vince, called Portfolio Management Formulas. In it he uses the mathematical methods employed in casino games and sports bets to rate stock-trading systems. The idea is that a stock-trading system produces a number of bets (or trades) that generate either a positive or a negative return. By analyzing the past flow of returns, Vince says that it is possible to anticipate what the future returns of the trading system will be. What we are trying to do is to build a trading system that will give us a positive mathematical expectation of winning. What does this mean? It means that if we invest $1,000 using the trading system, then we can reasonably expect the return on that $1,000 sum to be, on average, similar to the returns we had in the past using that system. It is true that sometimes retail investors find themselves in the same situation as the casino gambler, except that the odds are supposed to be in their favor, while the odds of the casino gambler are in favor of the house. The similarities between casino gambling and trading are that there are an almost unlimited number of bets and you can play any amount of money; you are only limited by the amount of your wealth. (This is very reasonable for traders; otherwise they would not be traders, for individuals who have a billion dollars would probably not spend their time trading.) Vince therefore defines the performance ratio (PR) and the pessimistic return ratio (PRR) as the two measures that will rate a trading system,

228

VALUE IN TIME

using only the number of trades and the return generated for each trade. He also introduces the optimum “f,” which is simply the optimum amount of money that we must risk on each trade in order to obtain the maximum PR or PRR. I advise day traders and short-term traders to have a good look at what Ralph Vince says. Using Vince’s formula, if our trading system generates 50 positive trades with an average return of 12 percent and 40 negative trades with an average return of −8 percent, it is obvious that the mathematical expectation of winning is positive. It is indeed calculated as: Performance ratio =

Ratio of positive trades × Average positive return Ratio of negative trades × (−Average negative return)

Performance ratio =

(50/90) × 12% = 1.875 (40/90) × 8%

However, if on average each positive trade lasted 20 days and each negative trade lasted 50 days, we can see that the trading method generated 1,000 positive trading days with an average daily positive return of 0.6 percent and 2,000 negative trading days with an average daily negative return of −0.16 percent. 50 × 20 = 1,000 positive trading days with an average daily positive return of 12%/20 = 0.6% per day 40 × 50 = 2,000 negative trading days with an average daily negative return of −8%/50 = −0.16% per day We could therefore say that the method produces a return of 0.0933 percent on average for each trading day, calculated using the following formula: (1,000 × 0.6%) − (2000 × 0.16%) = 0.0933% 3,000 If you consider that in one year there are about 250 trading days, the method could statistically generate 0.0933% × 250 = 23.33% per year (without using the compounding effect). If you use calendar days, you need to multiply the daily percentage by 365.

229

Performance

However, consider the reverse of the previous situation: If each positive trade lasted 50 days and each negative trade lasted 20 days, the method would generate 2,500 positive trading days (with an average daily positive return of 12 percent) and 800 negative trading days (with an average daily negative return of −8 percent). 50 × 50 = 2,500 positive trading days with an average daily positive return of 12%/50 = 0.24% per day 40 × 20 = 800 negative trading days with an average daily negative return of −8%/20 = −0.4% per day We could then expect a return of 0.08465 percent for each trading day: (2,500 × 0.24%) − (800 × 0.4%) = 0.08465% 3,300 This would statistically generate a yearly return of 21.21 percent, lower than the yearly return of 23.33 percent obtained when the negative trades took longer than the positive trades. We can easily see that two sets of trade productions that generate the same performance ratio could perform differently in a portfolio, simply because of the average time difference it takes between the sets to complete each positive and negative trade. As shown by Ralph Vince, the typical measure of performance in a casino is the performance ratio, which only involves the number of bets and the return on each bet. However, since the return on trading strategies is not instantaneous, it is better to rate the trading strategy by measuring the average return per holding day multiplied by the number of days during a reference period. Since the reference period is usually one year, I call this measure the yearly expected return (YER). The YER is the ideal return that you will have using the trading strategy if you are invested 100 percent of the time. Since it is very hard to be invested 100 percent of the time, you could consider yourself fortunate to generate 80 percent of the optimal YER.

230

VALUE IN TIME

The Pessimistic Return Ratio Ralph Vince defines the pessimistic return ratio (PRR) as the ideal measure for ranking different trials during the optimization process of a trading strategy (Portfolio Management Formula, p. 46). The PRR is defined as the performance ratio (PR) with one fewer square root winner and one more square root loser. This adjustment will give more weight on sets of trades that include more trades.
 √ (W − W)/T × AW  PRR =
√ (L − L)/T × AL where

W = the L = the T = W + L = the AW = the AL = the

number of winning or positive trades number of losing or negative trades total number of trades average winning trade amount average losing trade amount

As indicated by Ralph Vince, a PRR value greater than 2.0 indicates a very good system. A ratio greater than 2.5 is excellent.

My small comment on this PRR measure is that although it is excellent at measuring a set of trades, it does not include the time that it takes for these trades to unwind. Hence the need for the YER measure. YER =

Average return per trade × AD Average number of holding days per trade

AD is 250 (the number of trading days in one year) if you use the real number of days open for trading to calculate the average number of holding days per trade. AD is 365 if you use the number of calendar days to calculate the average number of holding days per trade. Note that the YER figure does not include the compounding effect of the daily gain, since it considers each trade as independent from the next. It is not the same as when you are using a set of trades inside a portfolio. The compounding effect can be dealt with by using the average logarithmic daily return for every day in the trade. Let’s see how the YER is calculated in Table 5.3. In Table 5.3, I calculated the return per trade, as well as the number of days it took to realize each of these returns. I then calculated the average number of days (42)

231

KG RHT BDX CTSH UPL HAL CVO CTSH CME CDNS

Symbol

10/25/2006 10/24/2006 10/20/2006 10/18/2006 10/10/2006 10/6/2006 10/4/2006 9/29/2006 9/27/2006 9/26/2006

Buying Date

$16.94 $19.42 $72.14 $77.41 $49.16 $27.68 $19.02 $74.01 $479.35 $16.82

Buying Price

12/15/2006 10/26/2006 12/26/2006 11/8/2006 12/18/2006 11/15/2006 12/26/2006 10/5/2006 11/17/2006 10/26/2006

Selling Date

Selling Price

$16.31 $14.83 $70.57 $77.12 $48.07 $33.57 $19.85 $75.44 $534.80 $17.79

TABLE 5.3 Correct Calculation of the Yearly Expected Return (YER)

Time limit Stop loss Time limit Sell signal Time limit Profit target Time limit Sell signal Sell signal Sell signal Average

Reason for Selling

51 2 67 21 69 40 83 6 51 30 42 YER

Invested Days

−3.72% −23.64% −2.18% −0.37% −2.22% 21.28% 4.36% 1.93% 11.57% 5.77% 1.28% 11.11%

Profit/ Loss

232

VALUE IN TIME

per trade as well as the average profit/loss generated per trade (1.28 percent). The YER is simply the 1.28 percent return obtained in 42 days, extrapolated to 365 days, which gives 11.11 percent. Of course, this is just an example calculated for trades initiated between September 26, 2006, and October 25, 2006. In reality, the YER calculation is realized using hundreds of trades. What Table 5.3 means for practical purposes is that this trading strategy generates trades that, on average, last 42 days and yield 1.28 percent per trade. This is not a very good strategy, but it is interesting as an example, because of the −23.64 percent loss that occurred on RHT on October 24, 2006. Another way to calculate the YER consists in first calculating the YER per trade, and then taking the average for all the trades. Table 5.4 shows such a calculation, which yields a YER of −388.63 percent. This second method is not valid, because it assigns more weight to the returns generated during short-term trades than to the returns generated during longerterm trades. In fact, we see in the example of Table 5.4 that the RHT loss that occurred during a very short period (two days) is measured as a trade carrying a YER of −4313.47 percent. Since the Table 5.4 method allocates the same weight to the YER of each trade to calculate the average YER, we end up with an aberrant −388.63 percent YER for the method. Furthermore, to make the simulation more realistic, I will also introduce a cost per trade of 0.25 percent of the amount invested, to cover the commission and the slippage costs. This is to say that an investment of $10,000 would cost us $25 when we buy and $25 when we sell. Out of this $25 sum, a commission of $10 is normal for online trading, while $15 for slippage costs seems realistic. For example, if the stock we want to buy is trading at $15 and we want to buy at the ask, we will pay $0.01 more than if we buy at the bid (if there are enough shares offered at the ask). For an amount of $10,000, we will have to purchase 667 shares, incurring a slippage cost of $0.01 × 667 shares = $6.67. By taking instead a higher figure of $15 as slippage cost, I foresee that we might have to bid the price still higher to buy the shares. Subtracting a 0.25 percent cost per trade means subtracting 0.5 percent from each line of Table 5.3, which produces the results shown in Table 5.5. In order to analyze the influence of the three parameters (the profit target, the stop loss, and the time limit) on the return of the trading strategy, independent of buy/sell signals that could be generated by the trading tools, the sample trading strategy was run over the three groups of stocks of Table 5.1: the laggards group, the standard group, and the highfliers group. As a reference, Table 5.6 summarizes the return produced by using a buy-and-hold strategy on the three groups. A buy-and-hold strategy simply states that we buy the stock on the first day and sell it at the last day for which we have valid data. I use the same method to calculate the return of

233

KG RHT BDX CTSH UPL HAL CVO CTSH CME CDNS

Symbol

10/25/06 10/24/06 10/20/06 10/18/06 10/10/06 10/6/06 10/4/06 9/29/06 9/27/06 9/26/06

Buying Date

$16.94 $19.42 $72.14 $77.41 $49.16 $27.68 $19.02 $74.01 $479.35 $16.82

Buying Price

12/15/06 10/26/06 12/26/06 11/8/06 12/18/06 11/15/06 12/26/06 10/5/06 11/17/06 10/26/06

Selling Date

$16.31 $14.83 $70.57 $77.12 $48.07 $33.57 $19.85 $75.44 $534.80 $17.79

Selling Price

Time limit Stop loss Time limit Sell signal Time limit Profit target Time limit Sell signal Sell signal Sell signal

Reason for Selling

51 2 67 21 69 40 83 6 51 30

Invested Days

−3.72% −23.64% −2.18% −0.37% −2.22% 21.28% 4.36% 1.93% 11.57% 5.77%

Profit/ Loss

TABLE 5.4 Incorrect Calculation of the Yearly Expected Return as the Average of the YER per Trade

−0.07% −11.82% −0.03% −0.02% −0.03% 0.53% 0.05% 0.32% 0.23% 0.19% Average

Profit/Loss per Day

−26.62% −4,313.47% −11.86% −6.51% −11.73% 194.17% 19.19% 117.54% 82.79% 70.16% −388.63%

YER per Trade

234

KG RHT BDX CTSH UPL HAL CVO CTSH CME CDNS

Symbol

10/25/2006 10/24/2006 10/20/2006 10/18/2006 10/10/2006 10/6/2006 10/4/2006 9/29/2006 9/27/2006 9/26/2006

Buying Date

$16.94 $19.42 $72.14 $77.41 $49.16 $27.68 $19.02 $74.01 $479.35 $16.82

Buying Price

12/15/2006 10/26/2006 12/26/2006 11/8/2006 12/18/2006 11/15/2006 12/26/2006 10/5/2006 11/17/2006 10/26/2006

Selling Date

$16.31 $14.83 $70.57 $77.12 $48.07 $33.57 $19.85 $75.44 $534.80 $17.79

Selling Price

Time limit Stop loss Time limit Sell signal Time limit Profit target Time limit Sell signal Sell signal Sell signal Average

Reason for Selling

51 2 67 21 69 40 83 6 51 30 42

Invested Days

TABLE 5.5 Calculation of the Yearly Expected Return (YER) Including Trading Costs

−3.72% −23.64% −2.18% −0.37% −2.22% 21.28% 4.36% 1.93% 11.57% 5.77% 1.28%

Profit/ Loss

YER

−0.50% −0.50% −0.50% −0.50% −0.50% −0.50% −0.50% −0.50% −0.50% −0.50%

Trading Costs

−4.22% −24.14% −2.68% −0.87% −2.72% 20.78% 3.86% 1.43% 11.07% 5.27% 0.78% 6.77%

Total profit/ Loss

235

Performance

TABLE 5.6 Return of the Buy-and-Hold Strategy Group of Stocks

Yearly Buy/Hold Return

Laggards group Standard group Highfliers group

−2.1% 13.6% 38.9%

the buy-and-hold strategy as the one I used to calculate the YER in Table 5.3. The returns of Table 5.6 will be compared later to the return of the sample trading strategy.

How to Get the Best Return from a Trading Strategy It is already well known to traders that by quickly taking profit, the risk of a downturn is limited. However, we also limit our future profit potential, since the stock we sold at a profit could continue climbing up and would have perhaps given us a better return. To analyze the impact of the profit target level on the return of the trading strategy, we need to disable the two other parameters (the stop loss level and the time limit level). We can then see in Figure 5.3 that profit targets have practically no influence on the yearly expected return (YER). Without any use of stop loss or time limit, the profit target of our sample trading strategy will yield approximately what a buy-and-hold strategy

FIGURE 5.3 YER of the standard group of stocks, calculated for different profit targets, without using stop loss, time limit, or transaction cost.

236

VALUE IN TIME

FIGURE 5.4 YER of Figure 5.3, but including a 0.5 percent transaction cost.

could have produced. This is not only true for the standard group of stocks as shown in Figure 5.3, but also for the two other groups. This is due to the fact that as soon as we take our profit, the trading strategy will quickly indicate that the stock price is above the nine-day average and force us to buy. In other words, the sample trading strategy is not much different from a buy-and-hold strategy. Including the 0.5 percent trading cost per completed trade, the yearly expected return is getting weaker for lower profit targets, as shown in Figure 5.4; this is logical, since lower profit targets imply a higher churn rate (we will buy and sell more frequently than when using higher profit targets). From this point forward in Chapter 5 and 6, all the YER calculations will include the 0.5 percent amount for transaction costs. If we now use different stop loss levels on the same trading strategy while also applying a 10 percent profit target and then a 20 percent profit target, we obtain the interesting data shown in Figure 5.5. Figure 5.5 mainly indicates that tighter stop loss levels are hurting profits, especially when using smaller profit targets. This is primarily due to the more numerous trades generating a high number of transaction costs. The fact that lower profit targets generate lower returns is not something that can be generalized to every trading strategy. With this small example, we can already see why we need to use realistic transaction costs when analyzing the return of a trading strategy. Therefore, whenever you use the services of a stock-picking web site, be sure that the historical performance includes realistic transaction costs. I even know of a “timely” stock-picking service that will send its pick only after the trade has moved at least 0.5 percent in its direction. That stockpicking service uses a −2 percent stop loss and 5 percent to 10 percent

Performance

237

FIGURE 5.5 YER of the standard group of stocks, calculated for different stop loss levels, using no time limit.

profit targets, but beats the market by a jaw-dropping margin year after year. Needless to say, its performance does not include transaction costs. Figure 5.5 also indicates that we cannot analyze profit target and stop loss levels independently of each other. It is their combination that will influence the return of the trading strategy. Indeed, let’s have a look at Figures 5.6 and 5.7, which show the number of positive and negative trades depending on different stop loss levels and profit targets. We can clearly see that lower profit targets increase the

FIGURE 5.6 Number of winning versus losing trades as a function of the profit target.

238

VALUE IN TIME

FIGURE 5.7 Number of winning versus losing trades as a function of stop loss levels, for a 20 percent profit level.

number of winning trades (Figure 5.6), while tighter stop loss limits increase the number of losing trades (Figure 5.7). By comparing Figure 5.7 to 5.8, we can also see that the level of stop loss that produces more winning than losing trades varies according to the profit target we use: If we use the 20 percent profit target of Figure 5.7, we will need to place relatively wide stop loss levels (higher than −13 percent) in order to generate more winning than losing trades. By comparison, Figure 5.8 shows that using a

FIGURE 5.8 Number of winning versus losing trades as a function of stop loss levels, for a 10 percent profit level.

Performance

239

10 percent profit target allows us to place tighter stop loss levels (higher than −8 percent) to obtain more winning than losing trades. However, it is clear that even if the tighter stop loss levels generate more losing trades, the average loss per losing trade will be lower. Symmetrically, if lower profit targets produce more winning trades, the average profit per winning trade will be lower. This is the reason why people like Ralph Vince use the pessimistic return ratio (PRR) to compare different trading strategies, since this measure includes both the number of trades and the profit/loss per trade. Figure 5.9, by contrast, shows that none of the varied stop loss levels leads to satisfactory PRR levels (a PRR of 2.0 or above is rated as very good by Ralph Vince), while Figure 5.10 shows that any profit target will give very good PRR figures when no stop loss or time limit is involved. In reality, however, we also have to consider the duration of each trade. Indeed, if we let our negative trades run longer than the positive trades, we will soon have a portfolio that will be riddled with negative trades. It is called the “loss of opportunity” factor. A higher loss of opportunity will be generated by the trading strategy in which losing trades take more time than winning trades. Indeed, in a portfolio we have a finite amount of cash to invest, and by investing a portion of our cash in a losing trade we will lose the opportunity to invest it in a winning trade. When we quickly cut our losing trades, we consequently increase the average proportion of winning against losing trades in our portfolio. As an example, Figure 5.11 shows that beyond point A (a −12.5 percent stop loss level), losing trades will last longer than winning trades, directly hurting the portfolio. The wider the

FIGURE 5.9 Pessimistic return ratio as a function of stop loss level.

240

VALUE IN TIME

FIGURE 5.10 Pessimistic return ratio as a function of profit target.

distance (S) between the two curves, the uglier our portfolio will become over time. This is why, as shown in Figure 5.12, any level of profit target would quickly render a portfolio very ugly if the distance S is positive (losing trades take longer to unwind than winning trades). This will be somewhat mitigated, however, if the number of winning trades generated is much higher than the number of losing trades. As a rule of thumb, an acceptable trading strategy must generate trades whose average winning-versus-losing duration must be higher than 1.5. As

FIGURE 5.11 Average trade duration as a function of stop loss level.

Performance

241

FIGURE 5.12 Average trade duration as a function of profit target.

shown in Figure 5.13, a fixed winning/losing duration ratio necessitates the use of stricter stop loss levels (point B: −5.5 percent) when applying a 10 percent profit target than when applying a 20 percent profit target (point A: −12.5 percent). If you remember, however, Figure 5.5 was showing miserable returns for the sample trading strategy when applying a −5.5 percent stop loss to the trading strategy while using a 10 percent profit target. This set of parameters can therefore be eliminated, which leaves us with higher profit

FIGURE 5.13 Winning-to-losing trade duration ratio as a function of stop loss level.

242

VALUE IN TIME

targets combined with wider stop loss levels (a 20 percent profit target and a −12.5 percent stop loss level). It should be noted that other trading strategies could work well for low levels of profit target and stop loss, as we will see in Chapter 6. Figure 5.14 shows that not using any means to limit our losses (either a time limit or a stop loss level) will lead to very poor returns on the portfolio, as indicated by the low winning/losing duration ratio.

What about the Time Limit Parameter? The time limit parameter works in a way that is similar to placing a market order on a fixed number of days after the purchase of the stock. As shown in Figure 5.15, this parameter does not improve the general profitability of the trading method. Nor, as shown in Figure 5.16, does it effectively cut the number of losing trades the way the stop loss parameter did (see Figure 5.7). Indeed, it is only after the time limit is reached that losing trades are cut. Before that limit, losses are allowed to mount. The consequence is the poor PRR obtained by the use of this time limit parameter (see Figure 5.17). In other words, compared to the use of stop loss, the time limit parameter produces a smaller ratio of winning to losing trades. However, the winners will evolve in the portfolio for a long time while the losing trades will be small and cut short. The time limit parameter thus has the propensity to produce a good winning-to-losing trade duration ratio (see Figure 5.18).

FIGURE 5.14 Winning-to-losing trade duration ratio as a function of profit targets.

Performance

243

FIGURE 5.15 YER of the standard group of stocks as a function of time limit levels.

This small exercise in trading logic using one simple example allowed us to understand the three different trade-offs that a trader is faced with: 1. The stop loss and time limit parameters work like an insurance policy

against bad trades, but both hurt profitability. This is especially true for tighter stop loss or shorter time limit levels (see Figures 5.5 and 5.15). 2. However, looser stop loss levels hurt the ratio of the winning/losing trade duration, leaving the portfolio open to too many losing trades (see Figure 5.13).

FIGURE 5.16 Number of winning versus losing trades as a function of time limit levels, for a 20 percent profit target.

244

VALUE IN TIME

FIGURE 5.17 Pessimistic return ratio as a function of time limit levels, for a 20 percent profit target.

3. The time limit parameter is a good solution for the winning/losing trade

duration ratio (Figure 5.18), but produces many more losing than winning trades (Figure 5.16). To conclude this section, it is obvious that:

r The adjustment of the trading parameters will not turn a losing trading strategy into a winning one.

FIGURE 5.18 Winning-to-losing trade duration ratio as a function of time limit levels, for a 20 percent profit level.

Performance

245

r If you want to use an insurance policy against losing trades, the stop loss parameter will work better than the time limit parameter.

r If you use the stop loss parameter, you have to adjust its level depending on your profit target (higher profit targets require wider stop loss levels).

MINIMIZING RISKS Even the best plan sometimes fails. Traders are optimistic people, so planning for failure is often a very difficult task: It forces us to look at the deficiencies of the trading strategy. We will first examine what we call risk, and then study how to best tune the risk/return balance of a trading strategy.

How to Measure the Risks of a Trading Strategy Of the four risks mentioned earlier (the risk of not being able to reach the expected return, the market risk, the trader’s behavior risk, and the trader’s wrong analysis risk), only the risk of being unable to reach the expected return can be measured directly. All the other risks can only be measured indirectly. Indeed, when we make a bad trade, it is difficult to know (for an outsider) whether the markets were responsible or the trader was responsible for the negative outcome. For an outsider, the outcome is identical: Time and money have been wasted. It is also interesting to note that for all the risks mentioned (except the first risk, the risk that we will not be able to reach the expected return), there are trading parameters available to provide protection:

r The stop loss parameter provides some protection against market downtrends or against a sudden price drop of the stock we are trading. Stop loss levels limit a trader’s pain and hence a trader’s potential erratic decisions when experiencing large losses on a few positions. r The profit target parameter forces the taking of profits on a regular basis, avoiding the predominance of the trader’s optimism or exuberant emotions. r The time limit parameter allows the trader to avoid continuation of an analysis mistake, since an analysis is valid for only a limited period of time. The measure of the risk of not reaching the return objective is simply a measure of the robustness of the trading strategy in terms of return consistency. This could be important, especially for a fund manager whose

246

VALUE IN TIME

compensation is often proportional to the return above a certain benchmark. The risk will then be measured by how much the return deviates from that benchmark. Measuring a deviation from an average is called standard deviation and is usually referred to as a measure of volatility. The standard deviation measures how consistent the returns have been in the past. If we successfully passed the optimization test, we may then suppose that such consistency will continue in the future. However, mathematically speaking, the standard deviation formula is valid only for normal or symmetrical distributions of returns. Such distributions seldom occur in financial analysis. This common simplification of assuming normal distributions makes calculations easier while research has not shown that this simplification generates important errors—hence its wide acceptance. Figure 5.19 shows the distribution of daily returns for the sample trading strategy. Note how the daily return is calculated: For each trade, we divide the return of the trade by the number of days it took to produce that return. In financial applications, many distributions are skewed either positively or negatively. As an example, Figure 5.20 shows a positively skewed

FIGURE 5.19 Probability that any given day will produce a return that is within the stated bracket, for the sample trading strategy using a 20 percent profit target and a −10 percent stop loss level.

Performance

247

FIGURE 5.20 Probability that any given day will produce a return that is within the stated bracket, for a good trading method using a 20 percent profit target and a −10 percent stop loss level.

distribution of daily returns produced by a good trading strategy (the supply/Large Effective Ratio trading strategy that we will study in Chapter 6). Not only is the distribution’s center of gravity clearly on the right side of the 0 percent value, but the right tail is fatter than the left tail, indicative of a good trading strategy. In order to see how to measure the robustness of a trading strategy, let’s revisit the example involving milking equipment. We know by experience how much milk our herd of cows would produce by hand-milking. We also know that cows produce less milk during a dry summer than during a normal summer. If the milking machine yields a better production than hand-milking during either a dry or a normal summer, we would be happy about it, but this is to be expected. Otherwise, why would have we invested in the equipment? What is not normal is to have cows that produce less with the milking machine than with hand-milking. This is called the downside risk. It is the variability of the returns below a return deemed normal under the actual market conditions: the return of the buy-and-hold strategy. In our case, what we have to measure is the downside risk per trading day. Indeed, our trading strategies are producing returns per trading day (i.e., daily returns).

248

VALUE IN TIME

Downside Risk Considering that BH is the average daily return of the buy-and-hold strategy, the downside risk is calculated by the following formula:   n  (ri − B H)2 /N Daily downside risk =  i=1

where

ri = represents only the daily returns that were below the BH value N = represents the total number of trading days n = represents the total number of days with daily returns lower than BH

The yearly downside risk of the trading strategy would then be calculated by multiplying the daily downside risk by the square root of the number of days in a year (250 if we count trading days, or 365 if we count the total number of days). To come back to our herd of cows, there are two other useful pieces of information that we would like to have: First, in our herd, what is the average number of cows that are allergic to the milking machine, and thus will produce less milk than by handmilking? This is called the downside frequency. For our trading strategy, it is simply the ratio of the number of trading days that produce a daily return lower than the daily buy-and-hold return to the total number of trading days. During the optimization process, the downside frequency is to be minimized. Downside frequency =

n N

Average Downside Deviation The average downside deviation is calculated as follows: For all the trading days that produce a daily return lower than the daily buy/hold return, we add the sum of their difference from the BH return, and divide that sum by the number of days that produce a return lower than BH. Average downside deviation =

n  i=1

(ri − B H)/n

Performance

249

Second, when a cow is allergic to the milking machine, how bad is it? Will she kick the machine and spill all the milk on the floor, or will she just be uneasy and produce a little less milk than by hand-milking? This is called the average downside deviation. The three risk measures that I have explained (the daily downside risk, the downside frequency, and the average downside deviation) are in my opinion useful only in two instances: 1. When we want to compare the risks of different trading strategies,

these three measures give us rational points of comparison. 2. If you are a fund manager, it is always cool to show nice PowerPoint

presentations demonstrating the risk/return evolution of your fund. For the rest of us, these measures of risk are of very little use. First, except for the downside frequency, these measures use quite abstract concepts, and second, how significant are they really? Does it make a big difference to know that 40 percent of your trades would be below the target? What is really important is to know that even in an adverse market condition, your trading strategy will meet your objectives and will shield you against bankruptcy risks. This is the reason why, in terms of risk management, I prefer to use the “average maximum drawdown” measure, which is close to the trader’s reality. The maximum drawdown related to a specific trade is the maximum loss that we may experience between the time we buy and the time we sell. For example, if we bought at $10 and sold at $15 but in the meantime the stock retreated to a minimum of $9, the maximum drawdown is ($10 − $9)/$10 = 10%. (Note that the definition is different from that of maximum drawdown in a portfolio, where we take a peak-to-valley measure.) The average maximum drawdown gives a physical picture of the pain that, on average, we will have to endure when using a specific trading strategy. The average maximum drawdown figures in Table 5.7 are the averages of the maximum temporary loss that we take for each stock during the whole period of time. It is the measure of the average maximum pain that we are going to endure for each group of stocks.

How to Tune the Risk/Return Balance This part is far trickier, since, as we saw earlier, there are quite a few different types of risks. Let’s look at the influence of the profit level on the average maximum drawdown, a realistic measure of risk. It looks as if

250

VALUE IN TIME

TABLE 5.7 Risk/Return Produced by a Buy/Hold Strategy Group of Stocks

Yearly Buy/Hold Return

Average Maximum Drawdown

Laggards group Standard group Highfliers group

−2.1% 13.6% 38.9%

−32.2% −18.2% −13.5%

Figure 5.21 indicates that lower profit targets limit the risk, since the maximum average drawdown is lower for smaller profit targets. However, this may not be the case. We know that lower profit targets reduce the average duration of each trade, and as a consequence also limit the average maximum drawdown. In other words, we are probably replacing a few large drawdowns with more numerous smaller drawdowns. This also means that the maximum average drawdown is probably not the perfect measure of risk. There are indeed two events that could bankrupt us as traders: a small string of large losses occurring together or a large string of small losses occurring together. If you ask me whether I would prefer to be eaten by a tiger or by an army of ants, I’d say that except to shorten the pain, I would not care much. However, I would certainly appreciate avoiding both scenarios. Figure 5.22 shows the probability of occurrence of drawdowns for a buy-and-hold strategy applied to the standard group of stocks. We can see that it would be wise to avoid the 5 percent of the drawdowns that are very steep (between −50 percent and −99 percent) and the 22 percent of the

FIGURE 5.21 Drawdown of the standard group of stocks, calculated as a function of profit targets, using neither stop loss nor time limit.

Performance

251

FIGURE 5.22 Probability of occurrence of drawdowns.

drawdowns that lie between −25 percent and −50 percent. The maximum average drawdown is not a good measure of risk, because often in the stock market, very large drawdowns have a tendency of occurring concurrently. This is evidenced in Figure 5.23. Hence, when markets enter into a price correction, the majority of the stocks are affected.

FIGURE 5.23 Time distribution of maximum drawdowns.

252

VALUE IN TIME

Clearly we want to avoid tigers (a few large drawdowns occurring together) and ants (many small losses occurring together). So what if we try to measure their probability of occurrence depending on the different parameters used? First, we need to define what together is. If I were a short-term investor, with an average holding time of five days, together would mean a period somewhat longer than five days—for example, two weeks. If I were a longterm investor, I would need to use a longer analysis period such as one quarter. In this chapter, I use a period of one month. This means that I will look at the drawdowns during consecutive one-month periods. A Measure of Large Drawdowns Considering our standard group that includes 101 stocks, we need to measure the percentage of stocks that will find themselves hit at the same time by a drawdown larger than −25 percent (tigers). Please note here that to measure the risk linked to the largest losses, we need to measure it in terms of large drawdowns and compare it to the number of stocks that we are analyzing. This is the risk of what I call “trading bankruptcy.” It is when the concurrent large drawdowns are so numerous that we risk completely crippling our portfolio. What exactly is trading bankruptcy, anyway? I define trading bankruptcy as the moment when a trader has lost 50 percent of his capital. Indeed, someone who loses 50 percent of his capital will need to have a 100 percent gain to come back to his original capital. Moreover, after emerging from trading bankruptcy, it will take that person 10 years of average annual returns superior to 13 percent to reach the same no-risk return offered by 5 percent Treasury bonds over the same 10-year period. How daunting a task is it to emerge from trading bankruptcy? As a point of reference, consider that over the 10 years ending in December 2006, the average hedge fund returned a little more than 10 percent annualized. In my opinion, it is highly unlikely that someone who has just lost 50 percent of his portfolio capital will constantly beat the average hedge fund during the following 10 years. Because the road to recovery is very long, trading bankruptcy should be avoided. To make this theory more understandable, let’s work on practical examples. Figure 5.24 shows a time distribution of maximum drawdowns linked to the trades generated by our sample trading strategy when applied to our standard group of stocks. In this example, I used only a 10 percent profit target as a parameter—no stop loss or time limit. I am limiting the analysis to the two years 2005 and 2006 since I have more data for that period. Figure 5.25 shows the proportion of occurrence of large drawdowns. This chart is better, in my opinion, than any mathematical formula at

Performance

253

FIGURE 5.24 Time distribution of maximum drawdowns using a 10 percent profit target.

FIGURE 5.25 Proportion of occurrence of drawdowns larger than −25 percent using a 10 percent profit target, without stop loss or time limit.

254

VALUE IN TIME

showing what the risk level of the trading strategy is, at least in terms of occurrence of possible disaster. For example, consider that in July 2006, during the market reversal illustrated in Figure 5.2, 11 percent of the generated trades were showing a drawdown larger than −25 percent. This is quite a large figure, and it is to avoid such consecutive occurrence of losses that traders use stop-loss orders. By comparison, consider Figures 5.26 and 5.27. They show what happens when the same trading strategy is used, but with a −20 percent and a −10 percent stop loss, respectively. These figures show that the use of stop-loss orders greatly reduces the risk of trading bankruptcy. Indeed, as shown in Figure 5.26, a −20 percent stop loss level would have reduced the worst occurrence of large drawdowns from 11 percent to 4.6 percent, while a −10 percent stop loss level would have lowered it to almost 1 percent, or one occurrence in 100 trades (see Figure 5.27). You may wonder why it is still possible to get drawdowns larger than −25 percent with stop loss levels of −10 percent or −20 percent. This is due to the price gaps that may occur on bad news. A price gap could easily pass through the stop loss level and effectively force the trader to exit the stock at a lower price than the one indicated in the stop-loss order. A Measure of Small Losses A measure of small temporary drawdowns is not very useful, because most of the stocks go through small

FIGURE 5.26 Proportion of occurrence of drawdowns larger than −25 percent using a 10 percent profit target and a −20 percent stop loss level.

Performance

255

FIGURE 5.27 Proportion of occurrence of drawdowns larger than −25 percent using a 10 percent profit target and a −10 percent stop loss level.

drawdowns anyway after we purchase them. What is more useful is measuring the ratio of the small monthly losses—for example, those that are larger than −5 percent—to total trades for the month. This is more useful, because these will be transferred to the portfolio. Our trading strategy will therefore have to produce good enough profits to cover these losses. This measure is calculated by adding all the losses larger than −5 percent for each month and then dividing that number by the number of trades that occurred during the same month. Using a stop loss level of −20 percent, we obtain the results shown in Figure 5.28. We can see that in April 2005, 25 percent of all the ongoing trades during the month produced a loss larger than −5 percent. Is this significant? It depends on how much larger than −5 percent the losses were. If the average loss is large, then there is a 25 percent chance of transferring a high level of loss to the portfolio. Figure 5.29 shows the level of such monthly losses. You can see that even if we call them small losses, their average is somewhat larger than −20 percent. We see here that in April 2005, 25 percent of the ongoing trades for the month generated a loss of −20 percent. If we multiply the two figures, we obtain −5 percent, which is the expected level of the loss that was transferred in April 2005 from the trading strategy to the portfolio. This is not the real loss level that will hit the portfolio, since the portfolio loss also depends on positive trades of the month. However, I find this figure to be a good

256

VALUE IN TIME

FIGURE 5.28 Proportion of occurrence of losses larger than −5 percent using a 10 percent profit target and a −20 percent stop loss level.

FIGURE 5.29 Average monthly loss, for all trades ending with a loss higher than −5 percent, while using a 10 percent profit target and a −20 percent stop loss level.

Performance

257

FIGURE 5.30 Expected MLT, for all trades ending with a loss higher than −5 percent, for a trading strategy using a 10 percent profit target and a −20 percent stop loss level.

reference for a loss that could hit, if profits do not show up as expected. I call this calculation the monthly loss transferred (MLT) calculation. We can see in Figure 5.30 that the average MLT is −1.22 percent. This means that all the small losses higher than −5 percent generated by our trading strategy will transfer an average loss of −1.22 percent every month to the portfolio, unless we generate positive trades to balance these losses. This figure shows the force we are fighting against. What happens if we do the same work but with a −10 percent stop loss level instead of a −20 percent stop loss level? We obtain a higher occurrence of small losses (Figure 5.31) than what we saw in Figure 5.28 for a −20 percent stop loss. However, even if the −10 percent stop loss produces lower losses, their occurrence is so frequent that the MLT level worsens to −1.81 percent, as shown in Figure 5.33. This is very disturbing, because I had always believed that a stop-loss order was good protection. The use of stop loss indeed offers good protection against larger hits, but the level of the stop loss itself will not help to better protect us against a string of small losses occurring during a market downturn. As we have seen, tighter stop loss levels protect us well against large losses, but either lose their efficiency against small repetitive losses or hurt our profit by forcing us to overtrade. This trade-off is shown in Figure 5.34.

258

VALUE IN TIME

FIGURE 5.31 Proportion of occurrence of losses larger than −5 percent using a 10 percent profit target and a −10 percent stop loss level.

FIGURE 5.32 Average monthly loss, for all trades ending with a loss higher than −5 percent, while using a 10 percent profit target and a −10 percent stop loss level.

Performance

259

FIGURE 5.33 Expected MLT, for all trades ending with a loss higher than −5 percent, for a trading strategy using a 10 percent profit target and a −10 percent stop loss level.

FIGURE 5.34 Expected MLT as a function of the stop loss level for a trading strategy using a 10 percent profit target.

260

VALUE IN TIME

FIGURE 5.35 Expected MLT without stop loss or time limit levels for a trading strategy using a range of profit targets.

By experience, an MLT level greater than −1.5 percent does not bode well for our portfolio. As a reference, Figure 5.35 shows the corresponding MLT for trade losses larger than −25 percent when using a trading strategy without stop loss or time limit levels. This chart speaks for itself. Does the Time Limit Parameter Help the MLT? As shown in Figure 5.36, the time limit level performs better than the stop loss level, in terms of expected MLT. For those who like stop loss, you can see in Figure 5.37 that a combination of a 30-day time limit level with a stop loss strategy also performs nicely in terms of expected MLT.

MEASURES OF RISK-ADJUSTED PERFORMANCE: THE SHARPE AND BURKE RATIOS The Sharpe ratio, originated by the economist William Sharpe, is a measure of the excess return achieved over a risk-free investment per unit of volatility. It is calculated as follows: Sharpe ratio =

Return of the trading strategy − Risk-free return Standard deviation of return of the trading strategy

Performance

261

FIGURE 5.36 Expected MLT as a function of the time limit level for a trading strategy using a 10 percent profit target.

The higher the Sharpe ratio, the better the portfolio’s performance. The investment firm Morningstar, Inc. states that a Sharpe ratio greater than 1.0 is good; outstanding funds have a ratio greater than 2.0. In the past, the Sharpe ratio of the S&P 500 has mostly been below 0.4. The Sharpe ratio is widely used to compare funds, which makes it very practical as a single

FIGURE 5.37 Expected MLT as a function of the stop loss level for a trading strategy using a 10 percent profit target and a 30-day time limit.

262

VALUE IN TIME

comparative measure. This is the reason why, in the following chapter, I use mostly the Sharpe and Burke ratio (explained below) as a measure of risk, instead of the MLT. Several other economists have proposed using the drawdown value on ratios similar to the Sharpe ratio. From among them, I selected the Burke ratio, devised by Gibbons Burke. Instead of using the standard deviation of returns to measure risk, the Burke ratio uses drawdowns, using the square root of the sum of the squares of each drawdown. This measure allows us to give more importance to large drawdowns than to numerous small ones. Burke ratio =

Return of the trading strategy − Risk-free return Daily Downside Risk

   n   Daily Downside Risk =  (Drawdown)2 i=1

In Chapter 6, where I use this Burke ratio, I take 5 percent as the risk-free return and consider only the drawdowns that are larger than −5 percent. The interested reader may refer to Franc¸ois-Serge Lhabitant’s book Hedge Funds: Quantitative Insights.

WHAT WE LEARNED IN THIS CHAPTER As you know now, this chapter is the necessary foundation for a better understanding of Chapter 6, which examines automated trading systems. In this chapter, I used a sample trading strategy to generate the buy signals. My objective was to show that after the trade is initiated, the management of the trade itself using the different parameters that I described here is independent of the generation of the signals (unless, of course, the Large Effective Volume trading signal flashes that we need to quickly cash out, for example). We saw that at the level of the trading strategy the risk/return balance is well measured by:

r The yearly expected return (YER) of the trading strategy. The YER should outperform the return of the buy-and-hold strategy.

r The monthly loss transferred (MLT) by the trading strategy to the portfolio, both for small −5 percent losses and for larger −25 percent temporary drawdowns.

Performance

263

Both the YER and the MLT, because they are commonsense performance measures that “speak to you,” should be preferred over mathematical formulas. However, to be honest, both the Sharpe and Burke ratio are easier to program than the MLT. Most important, we also saw how the profit target, the stop loss, and the time limit parameters can be used to manage an opened trade. We learned how they mechanically work on the trades, and are able to be altered independently of the trading signals themselves.

CHAPTER 6

Automated Trading Systems

ome people think that automated systems are no good because they take the decision making away from the investor. I for one do not want to use a system that leaves the investor (me) powerless or ignorant. Since the money I invest is the money that I’ve earned from working hard, it seems unnatural to me to let a bunch of wired transistors make decisions in my stead. Besides, when I go to a cocktail party, I always have good trades that I can talk about (let’s forget about the bad trades, which aren’t cool enough for cocktail parties). I cannot imagine myself standing among seasoned investors and telling them: “Yep! I am heavily invested in the stock market, but I have no idea what sectors or what stocks I am in. My computer does the trading for me, and it beats the market quite consistently. By the way, my hobby is fishing. . . .” Why, then, am I writing a chapter about automated systems? It is not because I’m concerned with how others perceive me at cocktail parties, but rather because automated systems are a very good way to generate profits on a consistent basis while controlling risk. Human beings cannot trade profitably with consistency over a long period of time—unless they are backed up by a trading system. One reason is that human beings are greatly influenced by their emotions and memories of past trades (both good and bad). These memories carry an emotional impact that may eventually influence their future decisions. It is well known to traders that computers remove emotion from trading. Computers, however, do much more than that. They:

S

r Can scan many more stocks than a human trader can. r Do not miss opportunities. 265

266

VALUE IN TIME

r Act whenever necessary (for example, by taking profits or cutting losses).

r Have neither a boss nor a spouse to whom they need to report. r Allow you to back-test and optimize your ideas. So despite my preference for human decision making over computerized decision making, there is a mathematical reality that I can’t ignore: Automated trading systems work. In Chapters 1 through 4 we studied several new tools that generate valuable and interesting trading signals. As we saw, one of these new tools—the Effective Volume signal—can greatly improve an existing trading system that is based on a traditional method. It can be used as a confirmation of other systems’ signals appearing simultaneously, but maybe also as a signal to turn a short-term trade into a long-term trade and switching from one system to the other. I believe that it is the easiest tool for a trader with a successful system to use in a practical way what I have discussed so far in this book. In this chapter we will see how the use of different combinations of the tools presented in the first part of the book allow us to develop trading strategies that perform. These can be used both by dedicated private investors and by fund managers. As a reminder, each tool is based on a specific concept that already carries with it the principle of a trading strategy:

r The Large Effective Volume signal tells us that it pays to buy a stock when that signal has been increasing for a few days while the stock price is still within a trading range, because it indicates shares accumulation that does not yet appear in the price. r The total Effective Volume signal tells us that it pays to buy when traders are accumulating (for example, you could decide to buy when the total Effective Volume signal crosses over its 20-day average); this indicates a change in the supply-balance equilibrium, which is a reason for a future price increase. r The Active Boundaries signal tells us that it pays to buy at the Lower Boundary and sell at the Upper Boundary. This is because the Upper Boundary indicates the point at which the average shareholder does not expect the price to increase further, while the Lower Boundary indicates the point at which the average shareholder expects the price to begin increasing. r The divergence analysis signal tells us that it pays to buy when the divergence between the Large Effective Ratio and the price rate of change is greater than the average of its historical maximum. This

Automated Trading Systems

267

indicates that some unusual distortion is happening between the volume accumulation trend and the price trend. r The supply signal tells us that it pays to buy when that signal is lower than five to seven percent, indicating that few shareholders are ready to sell their shares. r Finally, we should not forget the very important price trend signal, which states that “the trend is your friend” and that you should not bet against it. Turning these various instruments into a profitable and automated trading system is no small task. I use such a system every day myself, but I seriously doubt that individual investors would be able to do, practically speaking, what I do. That’s why this chapter is intended primarily for fund managers who enjoy better access to technical and human resources. Applying the different tools on a daily basis is a very daunting task: A one-day automated calculation of the different signals for 300 stocks takes about six hours of computing time on a 3.2GHz dual-processor computer. I think that a standard trading platform would be even slower; it would probably not be able to record intermediate calculations for hundreds of stocks and store them in a way for the user to be able to execute on a daily basis only incremental calculations on newly downloaded data. In other words, unless the different tools that I present here are fully integrated in a modern trading platform by an organization that offers the necessary professional training and technical support, individual traders will most likely not have the opportunity to use these tools to enhance their trading. The only exception would be the Effective Volume Excel add-on (after Microsoft corrects the execution speed problem on Excel 2007—an issue that has not been resolved as of this writing; earlier versions of Excel work fine). However, many funds do have the necessary computer power and inhouse technical support. All they need to do is train one technician to manage the set of computers that produce the trading signals. These signals can then be fed either to an automated trading platform or to a team of traders. This chapter is divided into two sections: 1. The production section deals with the questions of how to produce and

how to report the different trading signals. The production section will give you the potential for money creation. 2. The trading strategies section shows how to combine these trading signals to create successful trading systems. Each trading system will be determined by a fixed set of trading rules that can be unequivocally interpreted by a computer. This section shows how to make money in practice.

268

VALUE IN TIME

PRODUCTION OF TRADING SIGNALS Most modern trading platforms allow users to define either alert rules or trading rules. When certain conditions are met, the alert rules will trigger an instant computer alert message (for example, in case of volume surge). The set trading rules are used to scan a database and produce trading signals for the best matches. I will now review two screens that are based on the different tools introduced in Chapters 1 through 4: the alert screen and the production screen. The alert screen indicates something like: “Look out—large players are moving in or out of a stock and this is not reflected in the price.” The production screen sorts out the stocks that produce the best trading signals.

Alert Screen In terms of signal production, the first and most basic signal is an alert that calls for your attention: “Look here, maybe something is happening!” It is not a trading signal per se, but simply an alert so that you will not miss something that could be important. The question is: What are the things you can’t afford to miss? What you do not want to miss is something that is taking place under your nose and that will come to light only later on when a price change happens. In technical terms, you want to know if some large player has been accumulating or distributing shares for a period of more than three days while the price is trending in another direction or is still in a trading range—since funds need time to accumulate, taking a minimum of three days allows you to focus on only sustained accumulation patterns. This is useful for detecting:

r Accumulation while the stock is in a trading range or is finishing a downtrend.

r Distribution while the stock is in a trading range or is finishing an uptrend. If you are not yet invested in the stock, this alert is not a signal for you to jump in; instead, it simply alerts you that the price could soon move out. Therefore, if the positive Effective Volume flow trend is continuing, you should buy the stock as soon as the price breaks its base on the upside, as the breakout is likely to be real. However, if you are already invested in the stock and you see the Large Effective Volume turning flat for a few consecutive days while the price is still in an uptrend, this usually indicates that you should get out because

Automated Trading Systems

269

there is a lack of accumulation power. When you are already invested in a stock, you should consider the Large Effective Volume signal to be a real trading signal—it usually pays to err on the side of caution. (However, for shorting opportunities, the analysis of the total Effective Volume is more appropriate, as we will see in Chapter 7.) An example of an alert screen is shown in Table 6.1. As you can see, the rightmost three columns are the most important:

r The “Price 3 Days” column shows the price change within the past three days (between the closing price of the most recent day and the opening price three days before; positive price trends are shown in gray). r The “LEV 3 Days” column shows the strength of Large Effective Volume during the past three days, compared to the average three days’ strength calculated over the previous 15 days (positive figures are in gray). This strength is defined as the ratio between the rate of change of the Large Effective Volume flow during the past three days and the rate of change over the previous 15 days. r The last column is used as a ranking tool for the previous two columns. A positive number indicates that price and volume are trending in divergent directions. A negative number indicates that price and volume are trending in the same direction. The alert screen informs you of a developing situation: When LEV is trending positively while the price is still in a downtrend, it often indicates TABLE 6.1 Alert Screen for Price/Effective Volume Divergences

270

VALUE IN TIME

accumulation by large funds, which could be followed by a price trend change; when LEV is trending negatively while the price is still in an uptrend, this indicates profit taking, which could also trigger a price trend change later on. The lower part of Table 6.1, where the ranking figures are negative, indicates that the accumulation/distribution trend by large players is following the price trend. This is a trend confirmation signal. Every day, about 10 percent of all stocks show an early alert signal, while about 25 percent show a trend confirmation signal. The remaining 65 percent, which have been omitted from Table 6.1, do not indicate noteworthy information in LEV (this occurs late in a price trend or during price trading ranges for which no accumulation/distribution was signaled by the Large Effective Volume). As an example of an early alert signal, we can see in Figure 6.1 that large players are strong accumulators of Priceline shares even during the price downtrend B. This could indicate that the price pullback will be

FIGURE 6.1 Priceline: a valid early alert signal.

Automated Trading Systems

271

short-lived. (When I wrote these lines on July 10, 2007, I did not foresee that the price would jump and reach $80 one month later.) In contrast, Figure 6.2 shows a false early alert signal. KB Home has been experiencing a steep price downtrend, while large players were net buyers (arrow B). This, however, does not indicate that the price downtrend will change anytime soon. First, the sector does not attract positive attention from investors, and second, you can see in Figure 6.2 that arrow A has about the same strength as arrow B. This clearly shows that the large and small players neutralize each other. A stronger LEV signal associated with a changing price trend will be necessary to buy that stock.

Production Screen The production screen is the closest you can be to an automated trading system if you do not want to use or build one. Table 6.2 shows one section

FIGURE 6.2 KB Home: a false early alert signal.

272

VALUE IN TIME

TABLE 6.2 Production Screen

of my production screen. The original screen includes a pattern of colors that allows me to quickly see which indicators are positive. In Table 6.2, the cells that show a positive signal from the indicator are represented with a gray background. As indicated by its name, the first characteristic of a production screen is to show the level of the different indicators that are automatically produced. For example, you can see in Table 6.2 that the production screen shows percentage numbers for the four indicators introduced in this book: 1. Expectation. This number originates from the Active Boundaries cal-

culations (see Chapter 2). The number indicates the percentage price increase you may expect by investing at the current price level, until the price reaches the Upper Boundary. (In Table 6.2, the same number indicates the percentage price decrease you may expect by going short at the current price level, until the price reaches the Lower Boundary.) The color code indicates how far the signal is from the Upper or Lower Boundary. (The gray cell indicates a signal close to the Upper Boundary for short trades and close to the Lower Boundary for long trades. Short trade candidates are to be found among stocks that show a strongly negative global signal.)

Automated Trading Systems

273

2. Buy/sell divergence. This figure originates from the divergence anal-

ysis calculation (see Chapter 3). It shows in percentage terms the strength of the divergence between the Large Effective Ratio and the price rate of change. The color code indicates how far the signal is from the average of the historical maxima. (The gray cell indicates a signal greater than 1.5 times the historical maxima for the buy divergence signal, and less than 1.5 times the historical minima for the sell divergence signal.) 3. Large Effective Ratio. This figure measures the balance between large buyers and sellers, compared to the total volume exchanged during the analysis window (see Chapter 3). The color code indicates how far the signal is from the average of the historical maxima. (The gray cell indicates a signal greater than 1.5 times the historical maxima for the Large Effective Ratio signal—historical minima and maxima are defined in Chapter 3.) 4. Supply. This figure originates from the supply calculation model (see Chapter 4). It shows in percentage terms if the supply of shares is temporarily low enough to trigger a price increase as soon as the demand starts to increase. The color code indicates how strong the signal is. (The gray cell indicates a supply level that is lower than 10 percent.) A production screen should do two things. It should: 1. Produce individual indicators for each stock that allow the trader to

quickly ascertain the stock situation without analyzing a graph. 2. Produce a global signal that allows a ranking of the different stocks

and that automates trading decisions. A production screen also allows us to quickly confirm or invalidate the signals of the alert screen. For example, the KBH false signal of Figure 6.2 is also invalidated by the Large Effective Ratio value for KBH that stands at 1 percent, which is well below the historical high. As for PCLN, even if Figure 6.1 shows a valid alert signal, the expectation cell for PCLN (see Table 6.2) indicates a stock price appreciation of only 4 percent, which is still quite low. Obviously, for the first four indicators, what is important is not the percentage itself, but the color code that visually shows the relative strength of the indicator signal compared either to a reference level (for the supply signal) or to historical levels (for the other three indicators). This means that a trading method that combines several indicators first requires the calculation of these reference, or historical, levels.

274

VALUE IN TIME

If you remember, the various indicators introduced in this book tend to measure the behavior of different groups of shareholders:

r The Active Boundaries measure the limit levels of the expectation of the group of active shareholders.

r The Large Effective Volume measures the accumulation/distribution by large players.

r The divergence analysis measures how far the accumulation is built into the price.

r The supply indicator measures to what extent shareholders are ready to offer their shares for sale. The fundamental hypothesis of this book is that the group of shareholders who follow a given stock varies slowly over time. All things being equal, these shareholders will make similar buy or sell decisions. The calculation of the historical levels for a past period ranging between six and 18 months tries to capture a reference value linked to past decisions. This has two important consequences for the management of the signal production: 1. Some setup time must be allowed for each stock. This includes the

downloading and formatting of past data, as well as the detection of historical levels. If we anticipate 30 minutes of setup time for each stock, 1,000 stocks would require 500 hours of work, mainly computing time. 2. The historical levels must be automatically recalculated at least every month in order to incorporate the data of the past month—more frequent adjustments are too computing-intensive for those following a large number of stocks. In this way the trading system can adapt itself to the changing behavior of the pool of shareholders.

TRADING STRATEGIES As you can see in the rightmost column of Table 6.2, the production process also generates a global trading signal that serves to sort the best trading opportunities. It would be a mistake to think that this global trading signal is just a mathematical combination of all the previous indicators. What this global signal says is that, given (1) a trading strategy that has been formalized into a set of trading rules and (2) the weight that the trader puts on each of these rules, the global signal represents the degree to which the stock meets the predefined rules at this precise moment.

Automated Trading Systems

275

Does this mean that it is profitable to buy when the signal is high and sell when it is low? Assuming that the system works, it should be statistically profitable, right? Not really. At this point, it is impossible to know if the trading strategy will yield a profit and, if so, to what extent. Optimization is necessary: what signal strength to take, what time frame, and so on. Also, buying is only half the work. You still need to sell, and as we saw in Chapter 5, there are many reasons for selling: to buy a better play, to take your profit, to limit your losses, to shorten your investment time, or simply to follow a selling signal. Although short-term traders must carefully select the timing of their entries and exits, what matters most for long-term traders is the quality of entries. Short-term traders work with smaller profit margins. Therefore, a bad exit makes a significant dent on the profitability of the trade. Indeed, if a short-term trader has a 5 percent profit target, a 2 percent miss on the exit lowers the profit by 40 percent (2 percent out of 5 percent), while for a long-term trader who is looking at 20 percent or more profit per trade, a 2 percent miss on the exit of the trade lowers the profit by only 10 percent (2 percent out of 20 percent). In other words, although the trading strategies explained in this chapter manage both entries and exits, their main difference is in the quality of the entries. We will see that even if they use identical exit strategies, those trading strategies that have high-quality entries fare much better. But we will also see that for some strategies, early profit taking can greatly improve the results. Before studying the different successful trading strategies, let’s come back to Table 5.6 of Chapter 5. If you remember, I briefly introduced three different groups of stocks, including their average yearly expected return (YER) using a buy-and-hold strategy. I will be using the buy-and-hold return on these three groups of stocks as a benchmark against which we will compare the different trading strategies. What are the characteristics of a good entry strategy? My method is not different from other methods in its requirements: We need to find long-term value, detect a change trigger, and use common sense. Translated into Effective Volume vocabulary, the requirements are to:

r Find long-term value. The stock must be cheap, measured either in terms of Active Boundaries (see Chapter 2) or in terms of supply level (see Chapter 4). r Detect a change trigger. Large funds must be heavily buying, measured in terms of Large Effective Volume flow (see Chapter 1), Large Effective Ratio, or divergence analysis (see Chapter 3). r Use common sense. If you decide to be long, the long-term price trend should be positive, and the short-term price trend should not be negative.

276

VALUE IN TIME

Of course, these are obvious requirements, and you might think that anybody who can use the trading signals corresponding to these requirements would make a killing in the stock market. Unfortunately, this is not the case. There are four reasons why: 1. We saw earlier that it is not because a stock is cheap that it cannot

become cheaper. For example, a cheap uptrend reversal could easily become an expensive starting downtrend. 2. Large funds sometimes make large mistakes, especially during price downtrends. Price downtrends indeed attract large bargain hunters who have a positive long-term view of a stock. 3. It is sometimes very difficult to predict if a new slope (a small trend) will develop into a new long trend. 4. Finally, the market as a whole could turn against us, which often has

nothing to do with the specific signals generated by our stocks. Let’s now look at different trading strategies that meet these requirements. For the different trading simulations, I evaluate the signals at the close, but use the next day’s opening price as the buying price. I also take a 0.5 percent cost, including slippage, per round trade.

Trading Strategies Based on the Active Boundaries We saw in Chapter 2 that the Active Boundaries tool is excellent for capturing price trends. The Upper and Lower Boundaries capture the change in expectation of active traders. The Lower Boundary indicates that traders’ expectation for a price increase is high; hence the probability is high that the price will reverse up, especially if large players are net buyers at the level of the Lower Boundary. Let’s review how the trading strategy works, taking the example of the Todco company, a deepwater oil drilling company that was bought out in mid-March 2007. Todco’s price movements, represented in Figure 6.3, are captured within the Active Boundaries by Figure 6.4, in fact almost copying exactly the price swings. The question is: Is it allowed to buy at points A and B, which lie at the Lower Boundary? You can indeed see in Figure 6.5 and 6.6—which represent the Effective Volume flow separated by size leading to points A and B, respectively—that in both cases, the Large Effective Volume flow during the last three days was positive, while the price trend was not negative. The three requirements are thus met.

Automated Trading Systems

277

FIGURE 6.3 Stock price evolution for Todco. Source: Chart courtesy of StockCharts.com.

Let’s first study the influence of the Active Boundaries indicator without the use of the Effective Volume trigger. This gives the following trading rules: Buy r If the Active Boundaries signal is close to the Lower Boundary, r And the price trend is not negative. Sell r If the Active Boundaries signal is close to the Upper Boundary, r Or if one of the following selling parameters is met: profit-taking limit of 20 percent, stop loss limit of −10 percent, or time factor limit of 50 days. (Figure 5.13 showed that, for the test trading strategy used in Chapter 5, a 20 percent profit-taking limit combined with a −10 percent stop loss produced a winning/losing duration ratio higher than 1.5. The 50-day time factor limit was selected to give time for the Active Boundaries indicator to complete a swing from the Lower Boundary to the Upper Boundary.) Note that the trading strategy discovers a price trend change before the price change actually occurs. However, it is sometimes wiser to wait for the price trend to effectively begin before entering the trade. Table 6.3 shows that the results of this specific trading strategy are very good, especially in comparison to the return of the buy-and-hold

278

FIGURE 6.4 Active Boundaries for Todco.

FIGURE 6.5 Effective Volume for Todco, leading to point A.

VALUE IN TIME

279

Automated Trading Systems

FIGURE 6.6 Effective Volume for Todco, leading to point B.

strategy. The rightmost column of Table 6.3 indicates the proportion of invested days compared to the total number of days analyzed. The lower the average invested time, the more stocks we will need to research in order to find enough investment opportunities. TABLE 6.3 Return of the Active Boundaries Trading Strategy

Group of Stocks

Sharpe Ratio

Yearly Expected Return

Yearly Buy/Hold Return

Improvement over Buy/Hold Return

Average Invested Time

Laggards group Standard group Highfliers group

2.00 1.30 3.14

36.3% 21.9% 50.5%

−2.1% 13.6% 38.9%

38.4% 8.3% 11.6%

10.7% 16.7% 13.1%

280

VALUE IN TIME

First Improvement: Addition of the Large Effective Volume Condition Let’s now add the Large Effective Volume trigger condition on the trading rules: We buy only if the Large Effective Volume was positive during each of the previous three trading days. The trading rules are therefore adapted as follows: Buy r If the Active Boundaries signal is close to the Lower Boundary, r And the Large Effective Volume flow was positive during each of the previous three trading days, r And the price trend is not negative. Sell r If the Active Boundaries signal is close to the Upper Boundary, r Or if one of the following selling parameters is met: profit-taking limit of 20 percent, stop loss limit of −10 percent, or time factor limit of 50 days. As you can see in Table 6.4, this improved strategy gives better results for the standard and the highfliers groups, but at the expense of the average invested time. Although this trading strategy works very well, it is important to note two limitations that are inherent to the Active Boundaries indicator: 1. The Active Boundaries indicator catches trends between two parallel

lines, either for uptrends or for downtrends. Some people could be confused and be tempted to be long in a downtrend, simply because the visual pattern is similar: two parallel flat lines are less visually striking than two increasing or decreasing trend lines.

TABLE 6.4

Return of the Active Boundaries and Effective Volume Trading Strategy

Group of Stocks

Sharpe Ratio

Yearly Expected Return

Yearly Buy/Hold Return

Improvement over Buy/Hold Return

Average Invested Time

Laggards group Standard group Highfliers group

1.29 2.12 3.16

24.4% 31.3% 54.9%

−2.1% 13.6% 38.9%

26.5% 17.7% 16.0%

5.8% 8.6% 5.6%

Automated Trading Systems

281

2. The Active Boundaries indicator is very bad at catching long-term

trend changes. At the Lower Boundary, it is sometimes difficult to detect whether the Active Boundaries signal will slide below the Lower Boundary or reverse up, even with the use of the Large Effective Volume signal. Let’s review these two limitations using the example of the company Lexmark, a computer printer manufacturer. We can see in Figure 6.7 that Lexmark experienced a strong price uptrend (uptrend A) in 2006 followed by a similarly strong price downtrend in 2007 (downtrend B). As you can see in Figure 6.8, the Active Boundaries nicely captured these two trends. The trading strategy that says to buy at the Lower Boundary and sell at the Upper Boundary is valid only for uptrends, however. In downtrends, the trading strategy should be to short at the Upper Boundary and to cover at the Lower Boundary. It is evident that the uptrend A and the downtrend B are both eye-catching in Figure 6.7: They are very easy to recognize. But the same trends expressed in terms of Active Boundaries, as shown in Figure 6.8, request our close attention in order to be recognized. If this difference is true for the human eye, a computer will recognize the fixed levels of the Active Boundaries much more easily than a trend pattern. As a reminder, we rate the trend strength by taking the midpoint between the Upper Boundary and the Lower Boundary. For example, uptrend A was a 10 percent uptrend [15% − (15% − 5%)/2], while downtrend B was a −9.75 percent downtrend [−6.5% − (−6.5% + 13%)/2]. The analysis of what happened at point 1 in Figure 6.8 will easily illustrate the second limitation of the Active Boundaries. We can see that at

FIGURE 6.7 Stock price evolution for Lexmark. Source: Chart courtesy of StockCharts.com.

282

VALUE IN TIME

FIGURE 6.8 Active Boundaries for Lexmark.

point 1, we are at the Lower Boundary A, a location where the price should increase again. However, if the Large Effective Volume flow is either neutral or negative at point 1, we may conclude that there is no buying pressure. Therefore, we can expect that the price will not increase, but rather that it will continue to move further down, pushed by its own downward momentum. This continuous fall would break the trend (we notice in Figure 6.7 that at point 1, the uptrend seems already somewhat jeopardized). However, Figure 6.9 was showing a strong accumulation by large players (see arrow C), which in theory presages a new price surge. This was clearly a false signal, because the large players’ stock accumulation was not strong enough to trigger a new uptrend. If you remember, in Chapter 3, during the divergence analysis explanation, I introduced the Large Effective Ratio concept. The Large Effective Ratio is simply the ratio of the number of shares accumulated by large players during a certain period of time to the total number of shares that were exchanged during the same period. The Large Effective Ratio allows the monitoring of the buying and selling waves generated by the activities of large players. The comparison between the amplitude of the actual Large Effective Ratio and the past amplitudes allows us to judge whether the actual buying is strong by historical standards. Figure 6.10 shows that for Lexmark, on January 17, 2007, the Large Effective Ratio was well below its average of past maxima, indicating that the large players’ accumulation was probably too weak to influence a price trend change. Second Improvement: Addition of the Large Effective Ratio Condition Now let’s see how to improve the trading strategy by replacing the Large Effective Volume trading rule with a Large Effective Ratio

Automated Trading Systems

FIGURE 6.9 Lexmark: Effective Volume leading to January 17, 2007.

FIGURE 6.10 Lexmark: Large Effective Ratio, leading to January 17, 2007.

283

284

VALUE IN TIME

trading rule. The idea is not just to catch Large Effective Volume signals, but to consider only those that are exceptionally good. The trading rules are thus modified as follows: Buy r If the Active Boundaries signal is close to the Lower Boundary, r And the Large Effective Ratio is higher than the past maxima, r And the price trend is not negative. Sell r If the Active Boundaries signal is close to the Upper Boundary, r Or if one of the following selling parameters is met: profit-taking limit of 20 percent, stop loss limit of −10 percent, or time factor limit of 50 days. As can be seen in Table 6.5, the use of the Large Effective Ratio instead of the Large Effective Volume has significantly improved the results of this trading strategy, but at a cost in the average invested time. Third Improvement: Elimination of the Price Trend Condition The drawback of the previous improvement is that the investment opportunities that respond to this stricter condition are more difficult to find, although these opportunities are of better quality. To increase the number of opportunities, we therefore need to reduce the strictness of the trading rules. What about the elimination of the trading condition that states that the short-term price trend must not be negative? It is, of course, safer to wait for a price downtrend to stop or revert before buying a stock. However, because the Active Boundaries show us that the stock is cheap and the Large Effective Ratio shows us that there is a significant accumulation under way, the probability is high that the price trend will soon change, even if it is still negative in the short term. The elimination of this price trend condition will also procure more time for shares accumulation,

TABLE 6.5

Return of the Active Boundaries and Large Effective Ratio Trading Strategy

Group of Stocks

Sharpe Ratio

Yearly Expected Return

Yearly Buy/Hold Return

Improvement over Buy/Hold Return

Average Invested Time

Laggards group Standard group Highfliers group

3.89 3.90 5.10

66.1% 52.9% 67.0%

−2.1% 13.6% 38.9%

68.2% 39.3% 28.1%

2.8% 4.3% 2.8%

285

Automated Trading Systems

TABLE 6.6

Return of the Active Boundaries and Effective Ratio Trading Strategy without Price Trend Condition

Group of Stocks

Sharpe Ratio

Yearly Expected Return

Yearly Buy/Hold Return

Improvement over Buy/Hold Return

Average Invested Time

Laggards group Standard group Highfliers group

2.44 3.92 4.77

44.6% 57.8% 69.7%

−2.1% 13.6% 38.9%

46.8% 44.2% 30.8%

3.0% 5.5% 4.0%

which is an interesting characteristic for funds, since funds need many days to accumulate a position. Table 6.6 shows that for the standard and the highfliers groups, this strategy is showing slightly better results than the previous one, while improving the average invested time. Figures 6.11 through 6.15 compare the different improvements using several measures. Figure 6.11 shows the progression of the yearly expected return (YER) for the different improvements that have been brought to the original trading strategy using the Active Boundaries. Notice that the original Active Boundaries strategy offered a YER that was already higher than the 13.6 percent YER of the buy-and-hold trading strategy (shown by the dotted line in Figure 6.11). Except for the last one, each improvement brought a substantial YER increase to the results of the previous stage.

FIGURE 6.11 YER for the different improvements on the Active Boundaries trading strategies for the standard group.

286

VALUE IN TIME

FIGURE 6.12 Drawdown measures for the different improvements on the Active Boundaries trading strategies for the standard group.

As we could have expected, the improvements that offer higher returns are also those that procure lower risks. You will note that this contradicts the general belief that higher returns usually also generate more risks. We can see in Figure 6.12 that the average drawdown per trade decreases together with each improvement of the Active Boundaries trading strategy. As you will notice in Figure 6.13, the monthly loss transferred (MLT) to

FIGURE 6.13 MLT for the different improvements on the Active Boundaries trading strategies for the standard group.

287

Automated Trading Systems

the portfolio gradually diminishes with each strategy improvement, except for the last one. Consequently, both the Sharpe ratio (Figure 6.14) and the Burke ratio (Figure 6.15), which combine risk and return, show a strong increase almost with each improvement, indicating an increased efficiency for the trading strategies. Fourth Improvement: Early Profit Taking Now let’s turn to the sell side of the trade for which I have been using the trading rule to sell:

r If the Active Boundaries signal is close to the Upper Boundary, r Or if one of the following selling parameters is met: profit-taking limit of 20 percent, stop loss limit of −10 percent, or time factor limit of 50 days. We already saw in Chapter 5 that a fine-tuning of the three selling parameters will probably produce better returns. But let’s concentrate on the trading signal itself: Do we really need to wait until we reach the Upper Boundary in order to activate a sell order? When we buy at the Lower Boundary and sell at the Upper Boundary, our goal is to take advantage of the full swing. However, the price movement from the Lower to the Upper Boundary is seldom linear. Traders quickly move into an emerging new uptrend, strongly pushing the price up. But before the price reaches the Upper Boundary, its rise starts to falter with the arrival of early profit

Active Boundaries

Active Boundaries Effective Volume

Active Boundaries Effective Ratio

Active Boundaries Effective Ratio No Price Condition

FIGURE 6.14 Sharpe ratio for the different improvements on the Active Boundaries trading strategies for the standard group.

288

VALUE IN TIME

Active Boundaries

Active Boundaries Effective Volume

Active Boundaries Effective Ratio

Active Boundaries Effective Ratio No Price Condition

FIGURE 6.15 Burke ratio for the different improvements on the Active Boundaries trading strategies for the standard group, for drawdowns greater than −5 percent.

takers. Because most of the gain happens early in the trend, it is better to modify the trading rules and sell at one-third of the distance to the Upper Boundary. The modified trading rules then become: Buy r If the Active Boundaries signal is close to the Lower Boundary, r And the Large Effective Ratio is higher than the past maxima, r And the price trend is not negative. Sell r If the Active Boundaries signal is rising to one-third of the distance that separates the Lower from the Upper Boundary, r Or if one of the following selling parameters is met: profit-taking limit of 20 percent, stop loss limit of −10 percent, or time factor limit of 50 days. As you can see in Table 6.7, this improved trading strategy gives still better results, especially for the standard and the highfliers groups. Furthermore, since we have improved the sell side of the trading rules, we can now apply this improvement to all of the previous versions of the Active Boundaries trading strategies:

r The original Active Boundaries strategy. r The Active Boundaries and Effective Volume strategy.

289

Automated Trading Systems

TABLE 6.7

Return of the Active Boundaries and Effective Ratio Trading Strategy, without Price Trend Condition but Using an Early Profit-Taking Tactic

Group of Stocks

Sharpe Ratio

Yearly Expected Return

Yearly Buy/Hold Return

Improvement over Buy/Hold Return

Average Invested Time

Laggards group Standard group Highfliers group

3.43 10.01 12.71

46.0% 102.0% 140.1%

−2.1% 13.6% 38.9%

48.1% 88.4% 101.2%

1.5% 2.5% 1.8%

r The Active Boundaries and Effective Ratio strategy. r The Active Boundaries and Effective Ratio strategy with no trading condition on the price trend. The comparison of these four trading strategies using either a late or an early profit-taking tactic is shown in Figures 6.16 through 6.21. The data set in gray (which is at the left side) displays the results for selling the stock close to the Upper Boundary—in other words, to take profit later in the uptrend. The data set in black displays the results for selling the stock earlier in the uptrend, when the Active Boundaries signal is only one-third of the way between the Lower and the Upper Boundaries.

Yearly Expected Return (YER)

Active Boundaries

Active Boundaries Effective Volume

Active Boundaries Effective Ratio

Active Boundaries Effective Ratio No Price Condition

FIGURE 6.16 YER comparison for the different improvements on the Active Boundaries trading strategies using either a late or an early profit-taking tactic.

290

VALUE IN TIME

You can see in Figure 6.16 that rapidly selling the stock while it is still early in the uptrend produces much better results than if we want to ride the complete trend. This is because in an uptrend the price increase is more important at the beginning than at the end of the trend. Therefore, quickly selling has a positive impact on the YER, and will be positive for the portfolio, if we can find enough trading opportunities to use our cash. Although the YER improvement between the late and the early selling tactics is never double, Figure 6.17 and 6.18 show that both the Sharpe ratio and the Burke ratio improve by a ratio that is often larger than the YER improvement. This means that the early selling strategies carry fewer risks than the late selling strategies, and therefore are also more efficient. This string of improvements, however, comes at a cost: a sharp decrease in the average invested time. Indeed, except for the elimination of the price condition, since each improvement applies stricter trading conditions, it becomes more and more difficult to find good investment opportunities at any time. Figure 6.19 shows the average time you can be invested in each stock using the different versions of the Active Boundaries–based trading strategies. For the original Active Boundaries trading strategy, Figure 6.19 shows that the strategy produces an average invested time of 16.7 percent. This means that, on average, if we select a stock from our standard group of stocks, during one year we will be invested in that stock for 60.955 days

Active Boundaries

Active Boundaries Effective Volume

Active Boundaries Effective Ratio

Active Boundaries Effective Ratio No Price Condition

FIGURE 6.17 Sharpe ratio comparison for the different improvements on the Active Boundaries trading strategies using either a late or an early profit-taking tactic.

291

Automated Trading Systems

Active Boundaries

Active Boundaries Effective Volume

Active Boundaries Effective Ratio

Active Boundaries Effective Ratio No Price Condition

FIGURE 6.18 Burke ratio comparison for the different improvements on the Active Boundaries trading strategies using either a late or an early profit-taking tactic.

(365 days × 16.7%). If we select a second stock, we can also expect to be invested during 60.955 days; for these two stocks, if we choose to be invested only in one stock at a time, we will be invested during 121.91 days (2 × 60.955 days). This is true only if these two investment opportunities do not overlap in time. If we can find six investment opportunities that do

FIGURE 6.19 Ratio of time invested for the different improvements on the Active Boundaries trading strategies using either a late or an early profit-taking tactic.

292

VALUE IN TIME

not overlap, we then can be invested 365.73 days (6 × 60.955 days). Considering a 20-position portfolio, this would require us to find 120 investment opportunities. Statistically, following 120 stocks using the original Active Boundaries trading strategy will provide such opportunities. If, however, we consider that 40 percent of these opportunities overlap, we will need to scan 200 stocks every day (120 divided by 60 percent). If we perform the same type of calculation for the best improvement of the method (the combination of the Active Boundaries, the Large Effective Ratio, the “no price trend” condition, and the early profit-taking tactic), we see in Figure 6.20 that we will need to follow more than 1,333 stocks on a daily basis. We can therefore conclude that even if early profit-taking tactics offer an enhanced performance, they require much more work to find investment opportunities. The early profit-taking tactics also naturally shorten the trades and therefore require us to perform more trades. This will not only increase the commission and slippage costs, but also the stress of performing more buy/sell operations (all the results include a 0.5 percent trading cost per round trade). For example, Figure 6.21 shows that for the late selling trading tactics, the average trade duration is about 40 calendar days, which means that each position in the portfolio requires about 365/40 = 9.12 buy-and-sell trades per year. A 20-position portfolio requires 182 trades per year. Figure 6.21 shows that moving to an early

FIGURE 6.20 Number of stocks to scan to support being invested in a 20position portfolio on the Active Boundaries trading strategies using either a late or an early profit-taking tactic.

293

Automated Trading Systems

Active Boundaries

Active Boundaries Effective Volume

Active Boundaries Effective Ratio

Active Boundaries Effective Ratio No Price Condition

FIGURE 6.21 Average trade duration in days for the different improvements on the Active Boundaries trading strategies using either a late or an early profit-taking window tactic.

selling tactic shortens the average trade by more than half and forces us to perform more than twice the number of operations. An early reader of this book told me that I should give advice on what method is best to use. Well, the advice is straightforward: I go for the trading strategy that has the potential to produce a 102 percent profit (see Figure 6.16), even if this means a lot of work, especially if it is the computer that is working. In reality, the answer is straightforward only for a swing trader, a trader who closely follows the stocks’ swings. Large funds sometimes need many days to take a new position. Therefore, for them, the shorter the trade, the more difficult it will be simply to invest and get out of their positions later. The case of short-term traders is still different: How would our strategies fare for short-term traders who like to switch within a few days from one position to the next? To answer that question, it is important to see how an average trade evolves for each trading strategy. For example, if we follow 10 trades, how much profit will we earn on average after one day, after two days, and so on? For a given trading strategy, do we have a constant profit increase or a steeper one during the first days? In other words, do we earn 0.5 percent after one day, then 1 percent after two days, 1.5 percent after three days, and so on, to reach 5 percent after 10 days and 20 percent after 40 days? We can already sense that this is impossible, because on average, such a trading strategy would produce a YER of 182

294

VALUE IN TIME

percent (365/40 × 20% = 182%). Note that I use calender days here not days opened for trading. Figure 6.22 would have displayed a sharper rise of the average profit. Since the maximum profit reached as of now is 102 percent, it is clear that 182 percent is quite high. But is it really unattainable? Before answering that question, we need to look at the evolution of the average profit produced by each trading strategy as each trade progresses in time. Such an evolution is shown in Figure 6.22. I have represented three strategies that require about 40 days on average for trade completion. Figure 6.22 shows the evolution of the first 20 days for the following three trading strategies: 1. The AB strategy. This is the original Active Boundaries strategy, whose

results are shown in Table 6.3. 2. The AB ER strategy. This is the Active Boundary strategy that uses the Effective Ratio as a trigger to decide on the entry timing. Results for this strategy are shown in Table 6.5. 3. The AB ER NP strategy. This is the Active Boundary strategy that uses the Effective Ratio, but without the condition that the short-term price trend should not be negative at the time of purchase. Results for this strategy are shown in Table 6.6. The striking feature of Figure 6.22 is that all three trading strategies produce an average profit that is linearly increasing from the day of

FIGURE 6.22 Evolution of the average profit produced by each trading strategy depending on the progress of the trade from the date of initial purchase.

295

Automated Trading Systems

purchase but that suddenly stops progressing. For example, the AB strategy produces an average profit that increases regularly until day 11, and then progress ceases. The other two strategies produce an average profit that increases much more rapidly at the start of the trades, until day 5, and then shows either no sign of progress or even some weakness after more days pass. The explanation for why the AB strategy takes more days to produce its average maximum profit is simple: For the AB strategy, we did not use any buying trigger. This means that the AB strategy offers a good measure for the value of a stock (it shows when the stock is cheap), but is does not offer any indication as to the timing of the purchase. By comparison, the other two trading strategies shown in Figure 6.22 indicate more precisely when to enter the trade, which allows them to quickly produce value for each trade. The real question that comes out of Figure 6.22 is: Why should we continue to be invested in a stock after the trade has reached its full potential? For the AB ER NP strategy, keeping the stock longer than five days will not produce additional profit. After day 5, the risk of incurring a loss on that trade is now higher than the potential additional return. Fifth Improvement: Imposing a Five-Day Time Limit Table 6.8 shows the results after applying a five-day time limit rule to the trading rules used to produce Table 6.7. This last improvement produces a truly exceptional return, as you can see in Figure 6.23, which slows the YER evolution from one trading strategy to the next. In Figure 6.24 the Burke ratio, which is a measure of return by unit of risk (the risk being based on the drawdown calculation), shows a fourfold amelioration over the previous trading rule improvement. But let’s come back down to earth. Whenever I see a trading strategy that publishes exceptional returns with very low risk, even if I do not doubt

TABLE 6.8

Return of the Active Boundaries and Effective Ratio Trading Strategy, without Price Condition but Using an Early Profit-Taking Tactic and a Five-Day Time Limit

Group of Stocks

Sharpe Ratio

Yearly Expected Return

Yearly Buy/Hold Return

Improvement over Buy/Hold Return

Average Invested Time

Laggards group Standard group Highfliers group

15.21 22.70 17.50

137.2% 174.9% 139.8%

−2.1% 13.6% 38.9%

139.3% 161.3% 100.9%

0.9% 1.1% 0.8%

296

VALUE IN TIME

Yearly Expected Return (YER)

Active Boundaries

Active Boundaries Effective Volume

Active Boundaries Effective Ratio

Active Boundaries Active Boundaries Effective Ratio Effective Ratio No Price Condition No Price Condition 5 Days Limit

FIGURE 6.23 YER evolution for the different improvements on the Active Boundaries trading strategies using either a late or an early profit-taking tactic.

the published performance, I must ask myself: What are the costs linked to such a performance? The main cost is shown in Figure 6.25: It is the very high number of stocks that we need to scan in order to be fully invested in a 20-position portfolio. Do not forget that the methods I am presenting in this book are

Active Boundaries

Active Boundaries Effective Volume

Active Boundaries Effective Ratio

Active Boundaries Active Boundaries Effective Ratio Effective Ratio No Price Condition No Price Condition 5 Days Limit

FIGURE 6.24 Burke ratio for the different improvements on the Active Boundaries trading strategies using either a late or an early profit-taking tactic.

Automated Trading Systems

297

FIGURE 6.25 Number of stocks to scan to support being invested in a 20position portfolio with the Active Boundaries trading strategies using either a late or an early profit-taking tactic.

very computing-intensive, since they go down to the level of the minute data. For example, if it takes one minute of computing time to produce the trading signals for one stock, the 2,924 stocks that we need to scan (as shown in the last column of Figure 6.25) will require about 48.7 hours of computing time. This would require a few computers working in parallel, which is not a common practice among individual traders. Another way to attain such returns would be to reduce the portfolio from 20 to five positions. The 48.7 hours of computing time would then become a more manageable 12.18 hours, which could quickly diminish with faster computers coming out on the market. When you reduce the number of positions, though, you face two problems: 1. A large portfolio means more diversification, while a reduced portfo-

lio means increased risk. For example, assume that we invest 20 percent of our capital in a single stock. If, because of very unfortunate circumstances that even insiders could not have predicted, the stock price drops by 50 percent overnight, we would lose 10 percent of our capital. 2. Less diversification also means that more capital must be allocated to each single trade. However, we saw in Figure 6.22 that the majority

298

VALUE IN TIME

of the gains are realized in the first few days of the trade. Therefore, it is clear that these larger amounts of capital must be invested very soon after the appearance of the first buy signal; otherwise we will start losing profit opportunities. For the AB ER NP strategy, if we are one day late, we lose a 1.2 percent gain opportunity. If we are two days late, we lose a 2.1 percent gain opportunity, or more than half of the average maximum 4 percent gain that the strategy can offer after five days. In other words, to reach the calculated returns, we must invest the whole allocation within a single day. For example, if we manage $1 million in capital that must be allocated among five positions, we will need to invest $200,000 within one day in a single stock. This seems entirely possible. But what if we have $100 million in capital to invest? That would lead us to invest $20 million in one day in a single stock, something that is impossible for the majority of small- and mid-cap stocks. It is therefore clear that for large funds, the main limitation regarding the use of the best-performing short-term trading strategy is the amount of capital that they need to invest. As a matter of fact, a fund that would like to use one of the trading strategies presented here would have to adjust it considerably to allow for the possibility of investing in a large number of positions. The fund would also adjust the strategy so that it offers some time to invest in these positions before the price climbs again significantly. Let’s come back to Figure 6.23 and analyze the “late profit-taking tactics” results that are represented by gray bars. We had four improvements over the original Active Boundaries strategy. The Active Boundaries indicator is a value indicator. This indicator alone produces an acceptable return of 21.9 percent, which is already higher than the 13.6 percent return of the buy-and-hold method (a relative increase of 61.2 percent). The first improvement was to find a good trigger for entering the trade. The analysis of the Large Effective Volume flow during the preceding three days was found to be a good enough trigger, improving the return from 21.9 percent to 31.3 percent (a proportional increase of 42.9 percent). The second improvement was to instead use the Large Effective Ratio as a trigger. This increased the return from 31.3 percent to 52.9 percent (a proportional increase of 68.8 percent). The third improvement was a technicality regarding a trading condition about the price trend. This increased the return from 52.9 percent to 57.8 percent (a relatively small proportional increase of 9.3 percent). The most striking improvement on return of the trading strategy is by limiting the duration of the trade. This increased the return from 57.8 percent to 156.5 percent (a relative increase of 170.8 percent).

Automated Trading Systems

299

What we have just discovered here are the three pillars of successful trading strategies: 1. The first pillar is to find value. If we misjudge the value of a stock com-

pared to its current price, then it is very likely that we will miss the trade. (The value is measured in terms of trading opportunity, which, with the Active Boundaries, is defined as traders’ collective expectation of a future price increase.) 2. The second pillar is to detect a good time to enter the trade. I call it the trigger. The trigger will produce an indication that the stock is ready to make a move. 3. The third pillar is the time management of the trade; it is the knowledge of the time it takes (1) between the signal from the trigger and the beginning of the stock price movement, and (2) between the signal from the trigger and the reaching of the maximum average return of the trading strategy. Figure 6.26 shows the YER improvement over the buy-and-hold method using the three pillars of successful trading. The value pillar improves the buy/hold return by 1.6 times, the trigger pillar improves the buy/hold return by 3.9 times, and the time pillar improves it by 11.5 times. Of the three pillars, the most important is the one that brings the least

FIGURE 6.26 YER improvement over the buy-and-hold method using the three pillars of successful trading strategies.

300

VALUE IN TIME

improvement: the value. Indeed, if we misjudge the value (we buy when the price is too high), the other two pillars will be of no use, because we will most likely miss the trade. In order to illustrate this, I ran three trading strategies using different trigger signals, first without the value signal from the Active Boundaries, and then in combination with this value signal. The trading rules for the strategies are: Buy r As soon as the buy trigger signal is set, r And the price trend was not negative during the past five days. Sell r As soon as the sell trigger signal is set, r Or if one of the following selling parameters is met: profit-taking limit of 20 percent, stop loss limit of −10 percent, or time factor limit of 50 days. The three trigger signals are the Large Effective Volume, the Large Effective Ratio, and the buy divergence signal (the divergence between the Large Effective Ratio and the price as explained in Chapter 3). The results of these three trading strategies are shown in Figures 6.27 and 6.28.

FIGURE 6.27 YER produced by trading strategies based on different triggers, with or without the Active Boundaries value signal.

Automated Trading Systems

301

FIGURE 6.28 Ratio of invested days to total days available for the trading strategies based on different triggers, with or without the Active Boundaries value signal.

Figure 6.27 clearly demonstrates that without a value indicator, the buying triggers produce very poor results. The main reason for these poor results is that a buying trigger based on a measure of the accumulation of shares by large players produces too many signals. Funds do indeed accumulate or sell shares on a constant basis, and since they often need many days to take a position, this generates a large number of short-term Large Effective Volume signals. But this accumulation does not mean that funds are right to accumulate (i.e., that they correctly evaluated the value of the stock). If their evaluation was not correct, the high price will attract selling by other players; hence the accumulation by large players will not be sufficient to move the price upward. However, if such a strong accumulation occurs at a time of good value (low price), it is probable that this good value will also be recognized by other players. At some point, the collective buying pressure will change the supply/demand equilibrium, and the price will have to increase. It can be seen in Figure 6.28 that the triggers, without use of the Active Boundaries signal, generate so many trades as to be undiscernibly invested in the stock: Triggers are not good for value assessment and therefore cannot be relied upon without a very good value detection method. More stress must be put on the value than on the buy trigger; furthermore, I don’t think this depends on the type of indicators used. After I came to this “miracle” conclusion, I revisited Come into My Trading

302

VALUE IN TIME

Room by Alexander Elder (p. 129). Dr. Elder explains his triple screen method: The method of Triple Screen is to analyze markets in several timeframes and use both the trend-following indicators and oscillators. We make a strategic decision to trade long or short using trendfollowing indicators on long-term charts. We make tactical decisions to enter or exit using oscillators on shorter-term charts. Dr. Elder continues: The original method has not changed, but the system—the exact choice of indicators—has evolved over years, as have the techniques. Indeed, in this chapter, you did not find new ways to trade the market, but you saw how to best combine a new set of indicators to apply the trading principles that have created wealth since human beings began trading for a profit: “Buy value at the right time.”

Trading Strategies Based on Supply Analysis I promised in Chapter 4 that I would describe a trading strategy that uses the supply analysis tool. If you remember, I wrote: To determine whether a share at a given time represents value, you have to determine the probability that you will be able to sell it later on to someone else at a higher price. To increase your chances of finding value, you must find a buying price at which there will be very few sellers (the price will be so low that few are willing to sell at that price). At the same time, you also need to find buyers other than you who will push the price higher. The general idea of the supply trading strategy is that:

r The price has fallen so much that most shareholders are locked in and are no longer willing to sell their shares at such a low price.

r The price has not risen far enough from its base to attract profit taking by more recent buyers. Therefore, if the company is financially sound and in no danger of bankruptcy, we may decide to step in as soon as the price downtrend is over.

303

Automated Trading Systems

TABLE 6.9 Return of the 10 Percent Supply Trading Strategy

Group of Stocks

Sharpe Ratio

Yearly Expected Return

Yearly Buy/Hold Return

Laggards group Standard group Highfliers group

0.38 1.22 2.45

11.0% 21.4% 37.7%

−2.1% 13.6% 38.9%

Improvement over Buy/Hold Return

13.2% 7.8% −1.2%

Average Invested Time

10.3% 24.4% 22.8%

This strategy is simply to buy the stock whenever the probabilistic mathematical model of the supply level shows a supply that is less than an arbitrarily fixed ratio of the total number of issued shares. (In Chapter 4, I showed examples using a 10 percent figure as a measure of a supply level.) We sell under the same conditions as those used in the previous trading strategies (profit-taking limit of 20 percent, stop loss limit of −10 percent, or time factor level of 50 days). Table 6.9 shows that the results of the supply trading strategy using a 10 percent level are not very good. This 10 percent supply strategy does slightly better than the buy-and-hold strategy, but this is a poor reward for our efforts. I initially selected the 10 percent level believing that it was a supply level low enough to trigger a buy signal. However, we can see in Figure 6.29 that the supply method is very sensitive to the selection of the supply level, and it is clear that supply levels as low as 2 percent or 3 percent give

FIGURE 6.29 YER for the supply trading strategy calculated for various supply levels.

304

VALUE IN TIME

much better YER than higher supply levels. The cost for the use of very low supply levels is shown in Figure 6.30: the number of stocks that we need to scan in order to find trading opportunities. We can see in Figure 6.29 that going from a 10 percent to a 5 percent supply level increases the YER from 21.4 percent to 25 percent, while the number of stocks to scan is multiplied by three (Figure 6.30). The supply trading strategy is similar to the Active Boundaries strategy, because both are based on the calculation of shareholders’ return since the time they purchased their shares. To produce a buy signal on the supply strategy, the price must have experienced a significant decrease, typically bringing it well below the 50-day or the 200-day moving average. To produce a buy signal on the Active Boundaries, the price simply has to pull back to the support line of the trend, which is the Lower Boundary. If the price drops below the Lower Boundary, the Active Boundary method will stop issuing buy signals, since this means that there is a good chance that the situation for the company or the market has changed, and that the trend could then also change its direction. It is usually when the Lower Boundary is broken that you may start looking to modify your trading stance from a long position to a short position. However, at the Lower Boundary, the supply level itself is often still high. As a reference, I measured that at the Lower Boundary 84 percent of the stocks had a supply level higher than 5 percent, and 64 percent had a supply level higher than 10 percent. This means that after the Lower Boundary is broken on the downside, the price still usually has a lot more to fall in order for the supply level to become low enough to trigger a buy signal. Many traders would not even buy in

FIGURE 6.30 Number of stocks to scan to support being invested in a 20position portfolio using the supply trading strategy at various levels of supply.

Automated Trading Systems

305

such situations. Nor would funds chase after distressed stocks, which is the commonly used term for stocks that have experienced a prolonged price decrease. But, in most cases, the sharp overselling situation of a stock may no longer accurately represent its true value or its potential. We might then conclude that the price drop is temporary. In these types of distressed plays, timing is everything. If we buy cheap at the wrong time, we will most certainly be stopped out in a matter of days, since the stock has a nonnegligible probability of continuing its price decrease. In other words, the supply calculation tool is not as good a measure of value as the Active Boundaries tool. Indeed, the biggest risk of this strategy is a value miscalculation: The fact that the price has fallen a lot doesn’t mean that it will not continue further down. Let’s examine the risk of this trading strategy with an example from the service and security company American Science & Engineering (ASEI). As we can see from Figure 6.31, around May 20, 2006, the supply model was indicating that the supply had fallen to less than 5 percent of the total number of issued shares, a level that can be seen as low enough for a price to increase as soon as buyers appear. However, this simple supply strategy does not wait for buyers to step in; it dictates a buy decision as soon as the supply is lower than 5 percent. From $56, the lowest buying point of the buy zone of Figure 6.31, the price dropped another 36 percent. This clearly shows that a large price drop does not guarantee that a positive reversal is due to occur soon. For a reversal to happen, new buyers must appear. This was not the case, as we can see in Figure 6.32. On the contrary, Figure 6.32 shows a continuous selling pattern from large players (arrow A) even while the price trend turned flat (arrow B). This indicates that the selling pressure was far from over. Based on this small example, it is evident that the supply model must be linked to an Effective Volume–based tool (Large Effective Volume, Large Effective Ratio, or divergence analysis) that offers a very good measure of the demand strength. If the demand is indeed very strong at a point where the supply of shares has dried up, the price, in theory, should increase. As shown in Table 6.10, the return produced by the 10 percent supply trading strategy is already better when it is combined with the Large Effective Ratio tool (which detects significant increases of large players’ buying activity). Furthermore, Figure 6.33 shows that the returns produced by the combined supply/Large Effective Ratio trading strategy are always superior to the returns produced by the original supply trading strategy, independently of the supply level. If we follow the same improvement gradation as we used for the Active Boundaries trading strategies, we could start using the time limit parameter

306

VALUE IN TIME

FIGURE 6.31 Supply model for ASEI.

in the trades and see if a five-day time limit performs better than a 50-day time limit. Unfortunately, the time limit factor does not work for the supplybased trading strategies, and the maximum YER that can be squeezed out of that strategy is 70 percent, as shown in Figure 6.33. If you remember, the very low supply levels are usually seen in distressed stocks, and we all know that distressed stocks need a lot of time to come back to their original strength—if ever. You can see in Figure 6.34 that independent of the level of supply used, the progression in terms of average profit generated by the trading strategy is very linear: We cannot say that the early portions of the trades perform better than later portions. This was not the case for the

Automated Trading Systems

307

FIGURE 6.32 Large Effective Volume and price trends for ASEI.

Active Boundaries–based strategies that were showing (see Figure 6.22) a better performance in the early portions of the trades, which allowed us to use time limit parameters that produced much stronger general performances. But, at this point, we cannot say that a strategy that takes more time than another strategy to produce its return is necessarily worse. This is especially true for funds that need more time to take or exit from trading positions: A slower-moving strategy would allow funds more time to play the stock, which would in time produce a better return than fast-moving trading strategies could allow them. In order to compare the two types of strategies, the Active Boundaries–based strategy and the supply-based strategy, I selected from Figure 6.16 three Active Boundaries strategies (see Figure 6.35), for which I tried to find the equivalent strategies in terms of supply-based methodology. The natural equivalent to the original Active Boundaries strategy was the original supply strategy, and the natural equivalent to the Active

308

TABLE 6.10

VALUE IN TIME

Return of the 10 Percent Supply Trading Strategy Combined with the Large Effective Ratio

Group of Stocks

Sharpe Ratio

Yearly Expected Return

Yearly Buy/Hold Return

Improvement over Buy/Hold Return

Average Invested Time

Laggards group Standard group Highfliers group

0.66 1.99 2.70

16.1% 34.2% 44.1%

−2.1% 13.6% 38.9%

18.2% 20.6% 5.2%

10.8% 18.1% 17.7%

Boundaries combined with the Effective Ratio strategy was the supply strategy combined with the Effective Ratio. But, in both cases, what level of supply should be used? In order to compare two trading strategies, we need to find common ground, which I choose to be the ratio of invested days to the total number of days available for trading. That is, if I am invested for the same length of time using strategy A as strategy B, then, by comparing the risk/return balance of the two strategies, I could decide which is better. The idea is, for each of the Active Boundaries trading strategies in Figure 6.35, to find for the corresponding supply-based strategy the right supply level that will give it the same ratio of invested days as the corresponding Active Boundaries trading strategy. In other words, each group of two corresponding trading strategies must display the same ratio of invested days (see Figure 6.36). Table 6.11 summarizes the correspondence between the two types of trading strategies.

FIGURE 6.33 YER for the supply trading strategy and the combined supply/Large Effective Ratio trading strategy, calculated for various supply levels.

Automated Trading Systems

309

FIGURE 6.34 Evolution of the average profit produced by each trading strategy depending on the progress of the trade from the date of initial purchase.

Figures 6.37 through 6.39 compare the two types of trading strategies, separated into three sets. As you can see, the first two sets are almost equivalent, either in terms of return (YER), performance measured using the variability of returns (Sharpe ratio), or performance measured using drawdowns (Burke ratio). For the third set (the 1.75 percent supply and the

FIGURE 6.35 YER for Active Boundaries–based trading strategies.

310

VALUE IN TIME

FIGURE 6.36 Comparison of the ratio of invested days to the number of days available, for the supply-based strategy versus the Active Boundaries–based strategy.

Active Boundaries with early profit taking, both using the Effective Ratio), the story is different:

r The Active Boundaries strategy produces a YER that is, in relative value, 35 percent higher than the returned produced by the supply strategy (see Figure 6.37). r The Active Boundaries strategy shows a performance that is 113 percent greater (more than double in performance) than the performance produced by the supply strategy (see Figure 6.38).

TABLE 6.11

Correspondance between supply and active boundaries-based strategies

Automated Trading Systems

311

FIGURE 6.37 YER comparison between the supply-based strategy and the Active Boundaries–based strategy.

r In terms of Burke ratio, the Active Boundaries strategy shows a performance that is 33 percent better than the performance of the supply strategy. This is mainly due to the fact that the Active Boundaries produces a YER that is 35 percent higher than the YER of the supply strategy. Since the Burke ratio divides the YER by the risk level (measured in terms of drawdown), this means that the risk is identical for both strategies (see Figure 6.39). If we just follow the analysis of Figures 6.36 to 6.39 we can now say that the two types of strategies are equivalent in terms of both risk and return, except when we start using time limit parameters or early profittaking tactics (for which the Active Boundaries strategy combined with the Effective Ratio and the time limit parameter performs much better). Such a conclusion is somewhat suspect, because it goes against a major difference between the two types of strategies:

r The Active Boundaries strategy issues a buy signal as long as we are close to the Lower Boundary. However, when we break through the Lower Boundary, the buy signal is inhibited since this move often means a major trend change. r However, the supply strategy issues a buy signal as soon as the supply level falls below a certain low level of supply, but also continues to

312

VALUE IN TIME

FIGURE 6.38 Sharpe ratio comparison between the supply-based strategy and the Active Boundaries–based strategy.

issue that buy signal as long as we are below that level (of course, the Effective Ratio must also be higher than its historical maximum). This means that, in theory, if we misjudge a stock, after the price has fallen so much that the supply strategy issues its buy signal, the price could still fall to the ground, even if large players at some point are heavy buyers.

FIGURE 6.39 Burke ratio comparison between the supply-based strategy and the Active Boundaries–based strategy.

Automated Trading Systems

313

Therefore, the supply strategy is inherently more risky than the Active Boundaries strategy. As can be seen in Figure 6.40, for the supply-based strategy, there is indeed a larger proportion of trades that ended with a stop loss than for the Active Boundaries–based strategy. Let me explain Figure 6.40, which is special because the horizontal axis means different things for the supply and the Active Boundaries strategies. The upper line represents the proportion of stop loss trades produced by the supply/Effective Ratio–based strategy. From the left to the right, the level of supply is decreasing, starting from 20 percent down to 2 percent. This trend corresponds to trades with increasing YER (as we saw in Figure 6.33). We can see that the ratio of stop loss trades is diminishing with lower levels of supply. A ratio of 25 percent means that one trade out of four was a bad trade, since we have been stopped out with a 10 percent loss. The lower line represents the proportion of stop loss trades produced by the Active Boundaries/Effective Ratio–based strategy. From the left to the right, the selling signal from that strategy comes more quickly, starting at the Upper Boundary on the left and progressing toward the Lower Boundary on the right (do not forget that with this strategy, we buy at the Lower Boundary; selling close to the Lower Boundary means that we sell just a few days after buying). This trend also corresponds to trades with

FIGURE 6.40 Ratio of trades terminated with stop loss protection to the total number of trades.

314

VALUE IN TIME

higher YER. We can clearly see that the Active Boundaries–based strategy is generating a much lower ratio of stop loss trades than the supply-based trading strategy. A ratio of 10 percent means that only one trade out of 10 was a poor trade. It is interesting to see that the use of a stop loss feature has no effect on the YER, as can be seen in Figure 6.41: For both types of strategies, the YER does not change significantly regardless of whether we use stop loss protection. However, let’s look at the risk itself, measured by Burke’s downside risk formula (refer to Chapter 5). Figure 6.42 shows that the use of stop loss protection significantly decreases the risk linked to the supplybased strategy (look at the down arrows), while it has almost no influence on the risk linked to the Active Boundaries strategy. Figure 6.42 is very important for those who like searching for the fallen angels or the distressed stocks: They will need to use stop loss protection if they choose a supply-based strategy. As a consequence, the size of the funds that they will be able to invest in will be limited, since in an emergency it is always more difficult to get out of a large position than a small one. The second conclusion that we may draw is that if the trading strategy is producing reliable signals, you may avoid using stop loss tactics, such as in the case of the Active Boundaries trading strategy.

FIGURE 6.41 YER comparison with or without the use of stop loss protection.

Automated Trading Systems

315

FIGURE 6.42 Burke’s downside risk comparison with or without the use of stop loss protection.

Three Pillars of Successful Trading Strategies We saw in Figure 6.26 that the three pillars of a successful trading strategy for swing trading are: 1. Have a good assessment of value so that you select a stock that is

cheap, with an upwards potential. 2. Find the right trigger to enter the trade at a good time. 3. Time management: manage the evolution of the trade, and shorten it if

possible. Let’s review the different traditional tools that could be used for the three pillars: Value Assessment As explained in Chapter 2, I refer to the assessment of value in terms of trading opportunity. This is independent from the fundamental or intrinsic value of a stock, which is usually measured in terms of price-earnings ratio, growth, cash flow, and the like.

316

VALUE IN TIME

Here are a few well-known tools for detecting trading value:

r Price. The price must be in an uptrend, both in the weekly chart and in

r

r

r

r

r

r

the daily chart. The price must be above its 50-day moving average, and the 50-day moving average must be above the 200-day moving average. Relative Strength Index (RSI). The RSI compares the recent price gains to recent price losses and converts them into a number between 0 and 100. The buying signal comes when the RSI comes back over the 30 line that indicates an oversold level (see Chapter 2). Support line. This line indicates price congestions or the price levels where many buy/sell decisions were taken in the past. When the price declines to its support line, it is prone to move back up if enough buyers appear. Price patterns. There are famous stock price patterns such as the head-and-shoulders formation or the cup-and-handle pattern. These patterns, which are abundantly described in the literature, allow us to detect when a stock is entering a zone of interest in terms of trading value. Fibonacci retracements. These are also well documented in the literature. The most commonly used numbers for retracements are 38.2 percent and 61.8 percent. If the stock is in an uptrend, the Fibonacci theory says that if the price goes down 38.2 percent of that uptrend, it will likely move back up, and even more if it pulls back down to the 61.8 percent level. Trading value is found at the Fibonacci pullback point. Specific dates. If they can assess the fundamental value correctly, some traders try to invest only before the quarterly earnings date; others want to capitalize on momentum created after earnings are reported. Active Boundaries and supply analysis. In this book, I introduced the Active Boundaries tool (Chapter 2) and the supply analysis tool (Chapter 4) for value detection.

The Trigger, or Entry Timing The timing of trade entries is often found in short-term pattern analysis or the use of oscillator-type indicators. Most of these indicators use only price, but Richard Wyckoff worked extensively on the price/volume relationship more than 85 years ago. Here are a few well-known tools for entry timing:

r Candlestick analysis for the past few trading days (entire books are dedicated to these patterns).

r Moving average convergence/divergence (MACD) or MACD histogram (MACDH). The MACD momentum oscillator was developed by Gerald Appel. It compares a fast and a slow moving average in order

Automated Trading Systems

r

r r

r

317

to detect if the price change is quicker or slower than before. It compares the acceleration (rate of change) of the fast and the slow moving averages. If the acceleration of the fast moving average is higher than the acceleration of the slow moving average, this indicates a positive momentum in the price. Price trend. Price must ride between the 5-day moving average and the 10-day moving average to be a good play; the 10-day moving average is the first stop level in case of pullback. The 20-day moving average is the “always” exit point. Price volatility. When the 50-day price volatility is above 0.4, look for pullbacks; when it is below 0.4, look for breakouts. Richard Wyckoff’s method. This is too long to explain here. I would simply refer you to the set of DVDs of David Weis, the well-known Wyckoff advocate. David Weis shows in great detail how to look at the combination of volume and price to search for the ease of movement as a buy trigger. These DVDs are available on Dr. Elder’s web site (www.elder.com). Effective Volume, Effective Ratio, and divergence analysis. I introduced in this book the Effective Volume tool (Chapter 1), the Effective Ratio and the divergence analysis tools (Chapter 3) as trading triggers.

Time Management To my knowledge, there is no specific tool to manage the time aspect of a trade. The simple idea is that some trading strategies allow you to enter just before an important move in price. In such conditions, it is often better not to wait for the full swing before selling, because the price advance is often stronger during the earlier part of the trade than during the later part. For example, I showed that the Active Boundaries–based strategies are good candidates with which to use time management in the trade, while the supply analysis–based strategies are not. Strategies that use volatility or specific earnings dates are very good candidates for fast profit taking.

WHAT WE LEARNED IN THIS CHAPTER The most important thing that we learned in this chapter is that nothing has changed since human beings started trading for a profit: We must buy value at the right time. We also learned that if we misjudge the value, we will most probably lose money, even if we believe that the timing is good.

318

VALUE IN TIME

More specifically, we focused on two trading philosophies: 1. Swing trading, for which we used a combination of Active Boundaries

and the Effective Ratio. 2. Picking distressed stocks, for which we used a combination of supply

analysis and the Effective Ratio. Both types of strategies produce much better returns than the buy-andhold benchmark, but in both cases we noticed two things: 1. Getting higher returns from trading strategies also means having to

analyze a larger number of stocks in order to find trading opportunities. 2. Knowing how the trading strategy impacts the trade’s evolution in time

is essential, because, depending on the type of strategy, the right exit timing tactics can sometimes generate the most significant returns. We then summarized the three pillars of a successful strategy: (1) the value assessment, (2) the entry trigger, and (3) the time management of the trade evolution. We also learned that the selection of the right trading strategy mainly depends on the amount of capital to invest and on the trader’s ability to automatically analyze a large number of stocks.

PART THREE

The Bonus Section

CHAPTER 7

The Market Is a Two-Way Street Shorting Strategies

horting consists of borrowing a stock and selling it in the market, hoping that the stock price will decline. After some time, the short seller will repurchase the stock in the market (at a lower price, it is hoped, than what the stock was sold for, pocketing the difference) and return the borrowed stock to the lender. Over the past few years, the New York Stock Exchange has been reporting that the short level for stocks has been situated, on average, between only 5 percent and 7 percent of the average total daily volume of trading. This relatively small figure indicates that short selling is not a common trading practice. Some trading books advocate that good traders place as many short plays as long ones, but since only 5 percent to 7 percent of the stocks exchanged are due to short traders, I cannot imagine that the remaining 93 percent to 95 percent are bad traders/investors. I know many good traders who do not short stocks. It is a perfectly acceptable strategy to be mostly in cash during a bear market and wait for good investment opportunities without actively shorting stocks.

S

THE SHORT SALE “TICK TEST” RULE Before going into the technical analysis for short trading, I would like to come back to the short sale “tick test” rule. This rule dates back to 1938. It simply stated that short sales were allowed only at a price above the last

321

322

VALUE IN TIME

sale price, or if there was no change in the last sale price and this last sale price was higher than the previous price, then short sales were allowed at the last sale price. This simply meant that it was not possible to push the price down when shorting. This rule has made things relatively difficult for me in terms of Effective Volume, because it has made it impossible for the Effective Volume tool to detect shorting activities. Indeed, the Effective Volume method records the volume that is responsible for a price change from one tick to the next. Since short traders were not allowed to push the price down but were only allowed to increase the ask price, the Effective Volume could not detect them. However, the Securities and Exchange Commission (SEC) came to the rescue of the Effective Volume method and abolished this “tick test” rule as of July 6, 2007. In other words, since July 2007, the Effective Volume method has become even more effective.

HOW TO USE THIS BOOK’S TOOLS FOR SHORT TRADING This chapter is not intended to push you into short trading. My intention is to examine whether the tools introduced in this book can be combined to create successful shorting strategies.

Effective Volume Throughout this book, I have repeatedly said that the Effective Volume flow size separation between the Large Effective Volume and the Small Effective Volume is the best way to measure funds’ activity, and that it pays to follow funds. The advantage of the Effective Volume tools is that compared to standard tools, they give you a much better view of what is happening in the markets. This is especially true for long plays, as we have seen. We may then ask the following question: When we are selling a long position in a stock, should we directly turn into shorting the stock? As you may imagine, the answer is no: A shorting trade is not a carbon copy of a buying trade. You cannot simply take the buying strategies of Chapter 6 and turn them into shorting strategies. We will better understand this after studying the example of the energy producer Reliant Energy. As you can see in Figure 7.1, at the end of January 2007, Reliant Energy broke out of its trading range and started a long uptrend. Figure 7.2 shows a nice accumulation by large players, while the price was making a higher high and a higher low, indicating a possible breakout. Let’s suppose that we buy the stock on January 24.

The Market Is a Two-Way Street

FIGURE 7.1 Reliant Energy: price trend. Source: Chart courtesy of StockCharts.com.

FIGURE 7.2 Reliant Energy: Large Effective Volume at the buying point.

323

324

VALUE IN TIME

After buying on January 24, we can see in Figure 7.3 two very strong selling patterns: 1. On March 22, at point A, there was a strong price spike that exhausted

the buyers. This spike came at the end of a profit-taking trend that was signaled by the downtrend of large players (down arrow 1). 2. On April 5, at point B, we can see that the price is forming a double top, while large players are actively liquidating positions (down arrow 2). There is no question that we need to sell either at A or at B. We have indeed seen that it is not safe to bet against large players. The next question is: Should we bet with these large players and short at point A or at point B? Of course, when you look at Figure 7.1, you already know that shorting at the end of March 2007 or early April 2007 would have been a mistake.

FIGURE 7.3 Reliant Energy: Large Effective Volume at the selling point.

The Market Is a Two-Way Street

325

The problem of Reliant Energy is the contradiction between the large players’ selling movement and the strong price uptrend. You may think that a strong selling pattern by large players is enough to break a price uptrend. However, this is usually not the case: Strong price uptrends do not easily break, the same way that strong price downtrends do not easily reverse. It is mainly a question of value and the perception of value. Strong, lasting uptrends attract new investors up to the point where there will be an important disconnect between the value and the stock price. Once a growing number of investors start noticing the overpricing of the stock, they will start selling and the uptrend will begin to reverse down. This price reversal will attract latecomers, and this pattern between late trend followers who push the price up and early sellers who push the price down will continue until all the buyers are exhausted and a new downtrend can settle in. This is when you may go short. In reality, a good short will be discovered if you are able to find a disconnect between the value of a stock and how it is priced. We already saw that the Effective Volume tool is not good for value assessment. The Effective Volume tool is mainly used as a trigger, when value has already been ascertained. I have also noticed that this trigger works much better on the buy side than on the short side. There is indeed one main reason for buying a stock: to sell it later at a higher price. However, there can be many reasons for selling: taking some profit off the table, decreasing the risk before an earnings release, rebalancing a portfolio, and so on. You cannot blindly follow large funds when you see them selling, since it is impossible to know the exact reasons for their moves. However, when large funds are heavily buying, you know that you should also buy, especially if the supply side has dried up or if you have correctly determined the value of the stock. In the case of Reliant Energy, you will notice that at the end of March, the price trend is well above the 50-day moving average (see Figure 7.1), indicating a very strong price move. This is not a place where you will find an easy short trade. Short trades are more commonly found in a weakening price trend, which is typically seen as a price moving below its 50-day moving average. In other words, the heavy selling trend from large players cannot be trusted to signal a short play if it is not confirmed by a weakening price pattern. Short-trading rule #1: The price must be below its 50-day moving average. It is very important to remember here that when selling, large funds are very careful not to trigger a downtrend, since they would incur an instant paper loss on the shares they have not yet had the time to sell. Large

326

VALUE IN TIME

funds are seldom trendsetters in a selling pattern. The selling trend will be triggered either by negative news, by a general market move down, or by a larger than usual wave of profit taking from all the players. To detect a selling wave, we therefore have to look at the weakening pattern in total Effective Volume as opposed to the analysis of the strengthening of the Large Effective Volume that is used to detect a buying pattern. The weakening pattern in total Effective Volume must be compared to the price pattern using the divergence analysis tool. It can also be compared to past trends using the total Effective Ratio signal.

Divergence Analysis The divergence analysis between the total Effective Ratio and the price rate of change (refer to Chapter 3) is more relevant than the total Effective Volume flow analysis, because the divergence analysis compares today’s selling pattern to the strength of past selling patterns, which provides a better clue to where we stand in the selling wave. We will see that by studying the case of the software company MicroStrategy. As you can see in Figure 7.4, this company experienced a healthy price uptrend between August 2006 and mid-November 2006 (point 1). Looking at Figure 7.4, it is difficult to see at what point we should short: 1, 2, 3, 4, or 5. All of these points seem to be good shorting spots, since the price was subsequently falling. The dotted line shows the resistance level that is getting stronger each time it is hit without being broken. In Figure 7.5, which shows the total Effective Volume flow, I have also plotted the five points of interest that we are now going to study.

r Point 1: This is the least likely place for a safe short. We are still in a strong uptrend, far above the 50-day average (see the length of the double arrow between point 1 and the 50-day average in Figure 7.4). We can also see in Figure 7.5 that at point 1, the total Effective Volume flow is higher than its 20-day average and getting stronger—although it collapsed from point 1. You can see in Figure 7.4 that after point 1 the price pattern breaks down below its 50-day average line on a strong selling move. Something happened. The investor’s confidence is in jeopardy. r Point 2: For the first time, we are now reaching the resistance level (at point 1, the resistance line had not yet been created). But, since the total Effective Volume flow seems to have reversed up, as shown in Figure 7.5, we may not short. r Point 3: The price (Figure 7.4), is back above the 50-day moving average, slightly above the resistance line. The total Effective Volume flow

The Market Is a Two-Way Street

327

FIGURE 7.4 MicroStrategy: price pattern. Source: Chart courtesy of StockCharts.com.

is in a slight accumulation after heavy previous selling from point 2 (Figure 7.5). This is a dangerous short. It is better to wait for the price to move back down. In Figure 7.6, I have represented the Effective Volume flow separated by size, both leading to points 2 and 3. You can see that leading to point 2, the small players and the large players are moving in opposite directions with the same strength, neutralizing each other (arrows A and B). This is hardly a strong selling pattern. Regarding point 3, notice that even if the Large Effective Volume had shown

FIGURE 7.5 MicroStrategy: total Effective Volume flow.

328

VALUE IN TIME

FIGURE 7.6 MicroStrategy: Effective Volume flow by size, February 2007.

a strong selling pattern since point 2, at the very right of Figure 7.6 (at point 3) there is some accumulation starting, after a flat Large Effective Volume pattern (arrow C). This could signal a change of mood by large players. Since we are just crossing over the resistance line, we may suspect that this new wave of buyers could break the resistance at point 3 (Figure 7.4). This is not a safe short. r Point 4: We are again back to the resistance line (Figure 7.4), while the total Effective Volume shows a very strong selling pattern (Figure 7.5). This is a much safer short play than the possible short of point 3. Figure 7.7 details the Effective Volume flow separated by size, both leading to points 3 and 4. You can see that even if the price trends leading to points 3 and 4 (arrows D and F) are almost identical in strength, large players leading to point 3 (arrow C) were slightly positive buyers, while you can see strong large sellers leading to point 4 (steeper down arrow E). Point 4 is therefore a better short.

The Market Is a Two-Way Street

329

FIGURE 7.7 MicroStrategy: Effective Volume flow by size, March 2007.

r Point 5: I have to say that this is my favorite type of short. At point 5, we are in a price trading range, while the total Effective Volume is heavily down. This pattern appeared after the price gap down of April 11 pushed the price below the 50-day moving average, shattering the hopes that the stock could come back above its resistance level. You can see in Figure 7.8, which shows the Effective Volume flow separated by size, that large players were strong sellers (down arrow G) during the price trading range (flat arrow H). For me, shorting at point 5 is safer than shorting at point 4, because at point 5 you do not have to fight against a price trend, and you are below the 50-day moving average. Short-trading rule #2: The total Effective Volume flow must be below its 20-day moving average.

330

VALUE IN TIME

FIGURE 7.8 MicroStrategy: Effective Volume flow by size, April 2007.

If we now turn over to the divergence analysis tool (see Figure 7.9), we can see that the signal was indicating possible shorts close to points 3, 4, and 5. The divergence analysis indicator gives more reliable signals than does the Effective Volume, but we can see that both are difficult to use without an assessment of value. The only indicators of value that I used for the example of the MicroStrategy case study were the resistance line and the 50-day moving average. Now let’s see if the Active Boundaries indicator is a good complementary tool for assessing value in case of short selling. Short-trading rule #3: The divergence between the total Effective Ratio and the price rate of change must be below the sell divergence limit.

The Market Is a Two-Way Street

331

FIGURE 7.9 MicroStrategy: divergence analysis.

Active Boundaries Since it is an excellent indicator for trend monitoring, the Active Boundaries indicator can be used to find shorting opportunities, especially when combined with the Effective Volume or the divergence analysis tools. Starbucks, the world-famous coffee shop chain, offers an interesting case study for an analysis of shorting opportunities using a combination of indicators. We can see in Figure 7.10 that uptrend A, which occurred between August and November 2006, was followed by downtrend B. I marked five points (1 through 5) of interest for possible shorting opportunities. These points have also been marked in Figure 7.11, which represents both the Active Boundaries on the upper panel (the value), and the total Effective Volume on the lower panel (the trigger).

332

VALUE IN TIME

FIGURE 7.10 Starbucks: price pattern. Source: Chart courtesy of StockCharts.com.

FIGURE 7.11 Starbucks: Active Boundaries and total Effective Volume.

The Market Is a Two-Way Street

333

Let’s examine points 1 to 4:

r Point 1 is set on October 6, 2006, which is located far above the 50-day average (Figure 7.10). The sudden price increase pushed the stock into a new high, on heavy positive total Effective Volume flow. At point 1, the Active Boundaries indicator reaches the high level of 20 percent, but this type of average profit is very commonly seen and is not a sufficient reason to short the stock (although it is a good enough reason to take profit out if we are invested). r Point 2 is set just after the price gap down of November 16, 2006. This gap down came as a surprise to all the investors who had been actively buying shares since point 1, indicated by the increasing trend in total Effective Volume flow between points 1 and 2. Many of those who bought between points 1 and 2 lost a few percentage points on their investments. It usually is not enough to trigger an overall selling spree, but at point 2, these shareholders doubt that their decision to buy was correct. The price gap down of November 16 shattered their confidence, and they will now start selling. Furthermore, those who bought before October 6 and did not sell after the price surge of October 6 now regret that they did not, since on average at that time they were enjoying a 20 percent profit. Many of these are also going to sell to protect what is left of their profit. In other words, a small gap down after a trading range that is topping a price uptrend could be strong enough to trigger a sell-off. This sell-off can be strong enough to change the mood of the majority of the active shareholders, which will translate into a new trend and new Active Boundaries. It is therefore a good idea to short on the sell-off that was triggered by the gap down at point 2. r Point 3 has been set two days after a small price gap up that occurred on December 5, 2006, which was followed by renewed selling (decreasing Effective Volume flow). Note on the lower panel of Figure 7.11 that at point 3, the 20-day total Effective Volume average is starting to trend down. It is also interesting to note that on December 5, 2006, the price gap up was so weak that the Active Boundaries returned only up to the 0 percent level. From that point, we can already see the total Effective Volume flow decreasing together with the price decrease. Why would shareholders continue selling from an average return of 0 percent? Because they lost hope! The failure of the price gap up on December 5 to bring the Active Boundaries above 0 percent is an indication of renewed selling. We can also see that a new Upper Boundary was created at point 3. Since the Lower Boundary is still at –15 percent, we are now in a confirmed –7.5 percent downtrend (the midpoint between 0 percent and –15 percent), just at the Upper Boundary, where the share price is expensive. It is therefore permissible to short at point 3.

334

VALUE IN TIME

r Point 4 is set on the price spike of January 18, 2007, very close to the new Upper Boundary of 0 percent. We can see in the lower panel of Figure 7.11 that at point 4 the total Effective Volume flow was positive and above its 20-day average, indicating an accumulation of shares. This accumulation is detailed in Figure 7.12, where you can see that at point 4, large players were still accumulating shares. Is point 4 a good shorting point? No, because you absolutely may not bet against large players, even if you think that you are right and they are wrong. They have the purchasing power, which gives them the strength to move the market in their direction. Point 4 is therefore not a good shorting point. You have to wait until point 5, which shows a strong selling trend by large players. Short-trading rule #4: When close to the Upper Boundary, short only when the downtrend is confirmed.

FIGURE 7.12 Starbucks: Effective Volume flow by size, January 2007.

The Market Is a Two-Way Street

335

The Supply Analysis Tool In Chapter 4, we studied the level of supply with the objective of finding disequilibrium points between a very low level of supply and an increasing demand. At such points, we saw that a surge in demand, indicated by an increasing Large Effective Volume flow, could trigger a new price uptrend. We also saw that a supply level of 45 percent is not very different from a supply level of 35 percent. It simply means that in both cases, a buyer will easily find shares to buy. Figure 7.13 represents the supply level for

FIGURE 7.13 Starbucks: supply analysis.

336

VALUE IN TIME

Starbucks, as measured using the supply analysis tool presented in Chapter 4. You can clearly see that for all potential shorting points 1 to 5 the level of supply was quite high, indicating that there were, statistically speaking, enough shares available for sale. The only information of value that the supply analysis tool is offering is the potential zones of short squeeze, which are the zones of very low supply levels. It is very dangerous to short a stock that is indicating very low supply levels, because after we short, we still need other shareholders to sell their shares in order for us to gain from a lower stock price. The probability of finding sellers in a low supply level is quite small. Furthermore, we saw that even a modest surge in the buyers’ interest will push the price up, and could quickly force short players to cover, fueling the price surge that will squeeze still more short players and force them to cover. A low supply level indicates a situation that we can qualify most of the time as static: The price decrease has locked many shareholders into their positions, and these shareholders are waiting for a price increase, which could take some time to occur. However, what you are interested in as a short seller is not specifically the level of the supply itself, but the dynamic reaction of the shareholders when the supply level is still high. Indeed, you want to know how shareholders will react to the most recent price action: Will they keep or sell their shares? If you expect a sell-off, then it is a good idea to short ahead of the sell-off. One of the best methods for anticipating a sell-off is to study the repartition of the shares by profit level at a point in time (the evolution of the volume histogram), and see how this repartition evolves as a reaction to a price change. Let’s come back to the Starbucks case study. Figure 7.14 shows the share price pattern up to October 6, 2006, just after a price increase of more than 15 percent that occurred after October 2. We want to see whether between October 2 and October 6 shareholders will be willing to keep their

FIGURE 7.14 Starbucks: price trend up to October 2006.

The Market Is a Two-Way Street

337

shares or they will prefer to sell them. In Figure 7.14, I identified two groups of shareholders: 1. Group A is the regretful group of shareholders. This group bought

before August 2006 and therefore experienced the price drop that occurred in August. Some of them sold after the drop, but those who did not sell experienced a sentiment of failure, which was followed by the great relief of September, when the stock price rebounded closer to their purchasing price. 2. Group B is the winner group of shareholders, since most of them purchased their shares at a price lower than the October 4 price, just before the price gap up of October 5. We want to see how these two groups of shareholders could possibly react to the price gap up of October 5. In order to analyze the change, I have represented in the upper panel of Figure 7.15 the volume histogram of

FIGURE 7.15 Starbucks: volume histogram, October 2006.

338

VALUE IN TIME

these two groups of shareholders, in profit/loss percentage terms measured as of October 4, 2006 (before the price gap). The lower panel represents the same volume histogram, but as of October 6, 2006 (after the price gap). On October 4, many shareholders in group A were ready to sell in relief (since they first experienced a paper loss on a price drop that was later followed by a price rebound), while some shareholders in group B were ready to take their profit. On October 6, you can see that all shareholders in group A are now enjoying a small profit. This profit has come so quickly that most would be inclined to wait for a continued price appreciation. On the contrary, shareholders in group B are experiencing a very healthy profit and could decide to take money off the table. Then you have the newcomers who create the new group C (at the left of the lower panel of Figure 7.15), formed by the shareholders who bought on October 5 and 6. These shareholders have come in with high expectations and are not likely to sell. Shorting after the price gap of October 5 would therefore not be a good idea: The number of hopeful shareholders is too high compared to the possible number of profit takers. Let’s now turn to November 16 and 17. As shown in Figure 7.16, a small price gap down occurred on November 17, which triggered a sell-off in the following days. Since October 5, when the previous gap up occurred, the number of shareholders entering into group C has been increasing, while the number entering into group A has been decreasing. The upper panel of Figure 7.17 shows the volume histogram of the three groups of shareholders just before the price gap of November 17. As can be seen, most of the shareholders in group B are experiencing a slight profit. Their expectation for a further price increase is still strong. This expectation is being reinforced by the small price uptrend leading to November 17. Those in group

FIGURE 7.16 Starbucks: price trend up to November 2006.

The Market Is a Two-Way Street

339

FIGURE 7.17 Starbucks: volume histogram, November 2006.

A, who did not sell earlier, are now experiencing a relatively good profit (do not forget that this was a group of losers just two months earlier). Group B is the most profitable, and a large number of these shareholders are actively selling for profit taking. How is each of these groups going to react to the price gap down of November 17?

r The A group is the die-hard group of long-term holders who already went through one disappointment at the beginning of August. Those who did not sell at that time and still hold the stock may be slow to sell at this time. This group has become immune to pain. r The B group that is enjoying a good profit (in the center of Figure 7.16) is already in a selling mood. The small drop on November 17 could accelerate their selling tendency.

340

VALUE IN TIME

r Group C is now experiencing a slight loss. It is the group that came with the highest and most recent expectation for a price increase. This group is certainly disappointed now, and many of them will sell their stock in order to limit their losses. Looking at the lower panel of Figure 7.17, we can see that the total of the sellers (groups B and C) is much larger than the total number of nonsellers, meaning that if new buyers do not come in force, the selling movement will accelerate. We can see with this example that the supply analysis tool does not give clear signals of when to short. However, it will show when shorting is too dangerous because of the low supply levels. What I like, however, regarding the supply analysis tool is that when we separate shareholders by type (winners, losers, recent holders, less recent holders), the evolution of the volume histogram gives very good clues as to how actual shareholders may react in the event of a price increase or price decrease. This analysis is useful for both long and short plays.

WHAT WE LEARNED IN THIS CHAPTER We learned that shorting is possible using combinations of the tools that were introduced in the first section of this book. The four trading rules that I use for shorting are: Short-trading rule #1: The price must be below its 50-day moving average. Short-trading rule #2: The total Effective Volume flow must be below its 20-day moving average. Short-trading rule #3: The divergence between the total Effective Ratio and the price rate of change must be below the sell divergence limit. Short-trading rule #4: When close to the Upper Boundary, short only when the downtrend is confirmed. However, since I have not done any back-testing of shorting strategies using these rules, it is impossible to evaluate their true performance.

CHAPTER 8

Market and Sector Analysis

hat is the market influence on the performance of a stock? What is the influence of the sector performance on the performance of a stock? I do not know. However, I know that if I invest in a stock, my investment will perform better if the stock is in a trendy sector and if the markets are in a positive trend. The corollary is that when a sector or the markets are sharply down, even good stocks will go down in concert. Market analysis and sector analysis are so important that I did not want to finish the book without at least scratching the surface of these very large topics. I will not talk about such items as the evolution of consumer spending, commodity prices, interest rates, and the like, although they are critical for both market and sector analysis. However, I have been wondering if, on the technical side only, my tools could be used to offer broader sector and market views. This type of proposition looks both very attractive and dangerous: How can I apply some tools to a usage for which they were not developed? If you consider, for example, the cyclical pattern of the Active Boundaries or the comparison of the divergence analysis signal to historical references, both are based on a very fundamental hypothesis: The composition of the group of traders invested in a specific stock changes very slowly compared to the cycle of the stock price itself. The same traders will trade the same stock again and again at different times; they will use the same analysis method to take their trading decisions, forming a recurring pattern of trends and reversals. It is this pattern that is captured by both the Active Boundaries and the divergence analysis tools.

W

341

342

VALUE IN TIME

Can we also find such a cyclical price pattern in both the sectors and the market? Can we adapt the toolbox? The answer is yes, we can adapt the tools, and the preliminary results presented in this chapter are very interesting, although not complete. This chapter is divided in two sections: The first section applies both the Active Boundaries indicator and the divergence analysis to market analysis, while the second section applies the divergence analysis to an analysis of the different sectors.

WHEN IS THE MARKET BECOMING EXPENSIVE? This question is a difficult one to answer. So many scholars and Nobel Prize winners have been answering this question that I feel that my small contribution may not add anything of interest. The problem with that question is that there is no reference level against which to measure the expensiveness of the market. Let’s review some of the references that we could use, supposing that the stock market is well represented by the S&P 500 index:

r Monetary unit. Around June 2007, the S&P 500 was breaking a new high in U.S. dollars, but when we adjusted its value to the euro, then it was still far from the highs of 2000. Was the S&P 500 therefore considered cheap or expensive? r Purchasing power. As of June 2007, the S&P 500 had been growing since August 2002. Did it grow at the same pace as the gross domestic product (GDP) growth? Is it expensive compared to the growth of the wealth produced by the nation? Do we measure the wealth in terms of individual income growth or in terms of corporate income growth? If the purchasing power of both consumers and corporations has increased more quickly than the S&P 500 increase, may we conclude that in June 2007 the stock market was still cheap? r Future earnings growth. Can the growth in purchasing power continue? Since the S&P 500 represents the index of the 500 largest U.S. companies, what could affect their earnings growth? If we believe that the lack of availability of labor and cheap natural resources will have a negative effect on the future earnings growth, then the stock market could already be expensive. In Chapter 2, I evaluated the value of a stock in terms of the expectation of its current shareholders, and I said that shareholders’ expectation is inversely proportional to their gains. By measuring the average profit/loss of a group of shareholders (the shareholders who own the

Market and Sector Analysis

343

Active Float of the stock), we obtain a cyclical signal that I called the Active Boundaries. The Active Boundaries evolves between a Lower Boundary that indicates a cheap price and an Upper Boundary that indicates an expensive price. Unfortunately, this way of thinking may not be easily applied to an index analysis such as the S&P 500, because the S&P 500 is not one single company. Shareholders who have invested in these 500 different companies have profit expectations linked to their specific investments, not to the gain/loss of the index itself. Therefore, the hypothesis of a united group of shareholders who have a cyclical decisional pattern cannot be applied to a wide variety of shareholders who invested in companies that follow different cycles of their own. I had another idea: Since the market is “made” by large funds, and since these funds are evaluated on a yearly basis, why not use the same 12-month time period for an Active Boundaries calculation on the S&P 500 index? We know how many shares are traded daily of the companies forming the S&P 500. We could then calculate the average profit of all the shares related to the S&P 500 index that were exchanged during the past year and see how this figure evolved day by day. Such results are shown in Figure 8.1a. You can first see that the Active Boundaries catch the market value much better than the index trend itself or even the index rate of change over the same period (see Figure 8.2). For example, you can see that since 1987, the S&P 500 has been evolving between an Upper Boundary of 20 percent (UB1) and a Lower Boundary of –20 percent (LB1). You may notice that the October 1987 market crash is more apparent in the lower panel of Figure 8.1 than in the upper panel. The October 1987 crash moved the market almost overnight from the expensive Upper Boundary 1 to the cheap Lower Boundary 1. In Figure 8.1a, I have indicated three zones to which a set of boundaries each correspond: 1. Zone A shows the 1995–2000 uptrend, which is well captured by UB2

(+20 percent) and LB2 (0 percent). 2. Zone B covers the bubble crash that occurred between 2000 and 2003,

and is captured by UB3 (0 percent) and LB3 (–20 percent). 3. Zone C relates to the most recent uptrend, which started in 2003. It is

captured by UB4 (20 percent) and LB4 (0 percent). As I write this update on January 18, 2008, the market has evolved considerably, as can be seen in Figure 8.1b. As a reference, I also included in Figure 8.2 the price rate of change using a 12-month window. We can see that Figure 8.2 does not produce as

344

VALUE IN TIME

FIGURE 8.1a S&P 500: index price pattern and Active Boundaries.

regular a pattern as the one in Figure 8.1a. One of the reasons for this is that the price rate of change signal changes not only when a price change occurs, but also when that change is moving out of the 12-month analysis window. For example, we can see in Figure 8.2 that the 1987 crash produced a mirror image (which is shown by a large increase in the price rate of change), simply for a mathematical reason: The mirror image occurs when the price gap down of the 1987 crash leaves the analysis window (you may refer to the explanations of Figure 3.4 in Chapter 3, related to the analysis window). This problem can be corrected using some weighting methods (linear or exponential), but these corrections serve only to massage the data to obtain desirable results.

Market and Sector Analysis

345

FIGURE 8.1b S&P 500: updated index price pattern and Active Boundaries.

The Active Boundaries indicator is much stronger, because it measures the evolution of the average profit of shareholders, which is the main factor behind shareholders’ selling decisions. The strength of a value indicator such as the Active Boundaries is that at the right side of the figure, we can instantaneously see if we are close to one of the boundaries, and how close we are to it. For example, at the right of the lower panel of Figure 8.1a, the Active Boundaries indicator is at 11.3 percent, corresponding to an S&P 500 value of 1,536. (The method used is identical to the method described in Chapter 2 for finding price targets at the boundaries.) The simulation shows that the 20 percent UB4 level will be reached for an S&P value of

346

VALUE IN TIME

FIGURE 8.2 S&P 500: index rate of change.

1,650; at that point, the S&P will definitively be very expensive. On the other hand, the 0 percent LB4 level will be reached for an S&P value of 1,380. If we go down below 1,380, it will be the sign of a confirmed bear market. The calculation of the S&P values that corresponds to the different levels of the Active Boundaries must be reexecuted every day, because the trading activity during the day has an impact on the Active Boundaries indicator and therefore on the S&P value calculated through simulation. In other words, if on July 30 the S&P 500 value corresponding to UB4 is 1,650, at the end of July 31, this value will be slightly higher or lower depending on (1) the number of shares that have been exchanged and (2) the price change between July 30 and July 31. I wrote the previous paragraph at the end of August 2007. Since then the S&P 500 has gone decisively through its LB4 Lower Boundary (see Figure 8.1b.) The next Lower Boundary level is LB3, which indicates an average loss of –20 percent. That will happen for an S&P 500 value reaching 1,150. If ever we go there, Figure 8.1a indicates that we will snap back up. Another interesting research path regarding the application of the Active Boundaries to a measurement of the market’s strength could be to calculate the S&P 500 Active Boundaries using the Active Boundaries calculations operated for each stock in the S&P 500. The idea is the following: Knowing that each stock moves within its own price cycle, we could assign a number with a value between –1 and +1 to each stock’s calculated Active Boundaries. This value would depend on the position of the Active Boundaries signal within the boundaries (–1 when the signal is at the Lower Boundary, +1 when the signal is at the Upper Boundary, and between –1

Market and Sector Analysis

347

and +1 depending on the situation of the signal between boundaries). If we weight this number by the S&P 500 weight, I believe that we could build a new Active Boundaries indicator that is closer to the value of each of the components of the S&P 500, which would be more accurate than the Active Boundaries indicators in Figures 8.1a and 8.1b. The real weakness of the Active Boundaries indicator is that it cannot tell us if we will rebound at the boundary or break it. Indeed, whenever we touch one of the boundaries, there is no way to predict the future movement. When we analyzed a single stock, it was quite natural to rely on the Large Effective Volume as the trigger that would indicate the direction of the next probable move of the stock price. But, for the S&P 500, there are 500 different stocks and, since each stock has a different face value, we cannot simply add up the Large Effective Volume flow of each stock hoping to obtain a total Large Effective Volume flow for the whole index. (As a matter of fact, the Active Boundaries calculation on the S&P 500 itself is flawed, since a stock that is exchanged in a company with a share price of $100 carries the same weight in the calculation as one with a share price of $10.) The best approach is to use the concept of money flow, which allows us to measure the market evolution using the following six-step procedure: 1. For each stock of the index, calculate the minute money flow, which

2. 3.

4.

5.

is the minute-by-minute product of the Large Effective Volume and the average stock price during each minute. Select an analysis period—for example, 20 days. During this analysis period, calculate the Large Effective Money flow, which may be obtained by adding all the minute money flows calculated in step 1. During the same analysis period, calculate the total money flow, which may be obtained by tallying for each trading minute the total volume multiplied by the average stock price. (Please note that the total money flow is always positive, because it represents all the money that is invested in the stock during the analysis period.) However, the Large Effective Money flow is moving up and down, because the Large Effective Volume could be positive or negative depending on whether the price inflection between the previous minute and the current minute was positive or negative (see the Effective Volume definition in Chapter 1). For each stock, we divide the result of step 3 by the result of step 4. We then obtain for each stock what I call the “20-day large players’ strength” expressed as a percentage of the money invested in each

348

VALUE IN TIME

stock during that period. We now have a list of strength percentages corresponding to each stock of the index. 6. Each of these percentages must be weighted by the stock’s market

value, since it is how the S&P 500 weights the 500 different stocks that build the index. We can then obtain what I call the “market money flow,” which is a minute-by-minute evolution of the flow of money that moves in or out of the stocks of the S&P 500 as part of large players’ activity. I am pretty sure that if we study the evolution of this money flow at critical positions of the Active Boundaries for the S&P 500, we will be able to create a tool that could possibly predict the next move of the index. I believe that applying this type of idea to different indexes could lead to interesting results. As an example of such a calculation, I took 254 reference stocks from 36 different sectors (these are in fact some of the stocks that I follow on a daily basis, and for which I have sufficient data for this example). Figure 8.3 shows the evolution since August 1, 2006, of the price change of this group of 254 reference stocks, compared to the evolution of the S&P 500. We can see that until September 2007, both evolutions were closely related—my reference list indeed includes fewer financial and more basic material stocks than what is included in the S&P 500. We therefore may not conclude that the market strength analysis performed on these 254 reference stocks is a good indication of the global market strength. However, Figure 8.4 shows interesting results that a further study—presently underway—for the S&P 500 could confirm.

FIGURE 8.3 Price change for the 254 reference stocks compared to the S&P (between August 1, 2006, and March 6, 2008).

Market and Sector Analysis

349

The upper panel of Figure 8.4 shows the results of the market money flow calculation for these 254 reference stocks. Here are a few observations about Figure 8.4:

r The five negative divergences noted N1 to N5 were showing consecutive lower highs in the total money flow, while the corresponding average stock price was increasing to a new high. This pointed to a weakening market that was soon followed by a significant price drop. r Correspondingly, the two positive divergences noted P1 and P2 indicated that the money flow dropped to a higher low, while the price dropped to a lower low. This indicated an oversold market that soon snapped back up.

FIGURE 8.4 Price change compared to large players’ strength evolution (between August 1, 2006, and January 17, 2008).

350

VALUE IN TIME

r From July 6, 2007, the “tick test” rule changed (see Chapter 7), which may have influenced the pattern of the signal from that point on. (Note: this analysis was pointed out to me by Tim Ord, founder of www. ordoracle.com. Tim has an excellent daily newsletter specialized in the gold market, that also applies price/volume techniques).

SECTOR ANALYSIS There are two ways in which the sector analysis has been traditionally used: 1. Stock strength comparison within a sector If the price of a specific

stock is lagging its peer group, we can conclude that this stock is nonperforming and therefore sell it. If an important company that is part of a sector declares an earnings warning, we may conclude that the whole sector is under the same negative economic situation and will also react negatively. In anticipation, we would sell the other companies’ stocks that belong to the same sector. 2. Fund reallocation between sectors. It is well known that sectors have different cycles. For example, the telecommunications sector will not be likely to move in phase with the retail or housing sectors. Therefore, traditional defensive funds such as pension funds will reallocate their funds between sectors on a regular basis, depending on measures such as the relative sector price performance or the change of interest rates. The question is: Can the Effective Volume method improve how sector analysis has traditionally been performed? Remember that the Effective Volume, when it is split into the Large and the Small Effective Volume, allows us to detect when funds are accumulating or distributing shares of a certain company. Let’s review how we can use that method:

r To detect the relative strength of stocks in terms of Large Effective Volume (it could indeed be interesting to discover that a specific stock in the sector is under heavy accumulation compared to the other stocks belonging to the same sector). r To calculate the total money flow entering or exiting a sector (on that basis, we could then perhaps predict the future price direction in which the sector will move).

351

Market and Sector Analysis

Strength Comparison within a Sector As usual, I prefer working on examples that we could eventually extend to a general approach. I will therefore first use the oil drilling sector and more specifically the sector of “oil sea drilling.” Indeed, since land and sea drilling use different techniques and equipment, it is wiser to separate the two fields and compare companies that are working in the same field. Table 8.1 gives the list of companies active within the oil sea drilling group that I will study from now on. The “capitalization” column represents the market capitalization of each company, while the rightmost column represents in monetary terms the total amount of money that was exchanged in order for the daily average number of shares to be traded. For example, we can see that for the company Transocean, an amount of $573.7 million in company shares is exchanged daily. Table 8.2 shows the proportion for each column compared to the total of the column. We can easily see that:

r The first four companies (ATW, THE, OII, and RDC) are rather small compared to the last three (NE, GSF, and RIG). (GSF was later merged with RIG) r For every stock except Rowan Companies, the proportion of the daily stock transactions is very close to the proportion of the capitalization. In Figure 8.5, I have represented the profit/loss starting on June 19, 2006. Although the black-and-white figure does not allow us to distinguish each company clearly, we can easily notice that each stock’s profit/loss follows the same cycle: They go up and down in unison, probably following the changes in the underlying commodity price: oil. In Figure 8.5, I have singled out with a thicker line the company Todco (THE), which was bought out on March 19, 2007. by Hercules Offshore, Inc. TABLE 8.1 Group of Oil Sea Drilling Companies Amount Linked to Daily Stock Transactions (Million $)

Company

Symbol

Capitalization (Million $)

Atwood Oceanics Todco Oceaneering International Rowan Companies Noble Corporation GlobalSantaFe Transocean

ATW THE OII RDC NE GSF RIG

2,050 2,790 3,050 3,330 12,290 15,520 28,200

35.0 72.3 40.6 128.6 260.4 313.7 573.7

Total

67,230

1,424.3

352

VALUE IN TIME

TABLE 8.2 Group of Oil Sea Drilling Companies, Proportional Figures

Company

Symbol

Proportion of Capitalization

Atwood Oceanics Todco Oceaneering International Rowan Companies Noble Corporation GlobalSantaFe Transocean

ATW THE OII RDC NE GSF RIG Total

3.0% 4.1% 4.5% 5.0% 18.3% 23.1% 42.0% 100.0%

Proportion of Daily Stock Transactions

2.5% 5.1% 2.8% 9.0% 18.3% 22.0% 40.3% 100.0%

Looking at Figure 8.5, it is difficult to decide which stock is performing best. We can see that since November 2006, Rowan Companies (RDC) has been the worst performer in the group. Let’s now compare the signal of the Large Effective Volume flow for each of the companies. Since each company stock is traded at a different price, comparing the Large Effective Volume flow does not provide a good indication as to where the money is flowing. In order to see how much money is comparatively invested in each stock, it is better to multiply the Large Effective Volume flow by the stock price. Not surprisingly, the companies with larger capitalizations attract more money than the companies with lower capitalizations, as shown in Figure 8.6. If we compare Transocean (RIG) to Atwood Oceanics Inc. (ATW) (at the right of

FIGURE 8.5 Oil sea drillers: profit/loss comparison.

Market and Sector Analysis

353

FIGURE 8.6 Oil sea drillers: large players’ money flow.

Figure 8.6), we can see that RIG shows a positive money flow by large players of about $1.4 billion during the last year, while ATW shows a positive money flow of about $133 million. In other words, RIG is attracting 10.5 times more money from large players than ATW is. But we saw in Table 8.1 that RIG in general was attracting $573.7 million per trading day, which is 16.4 times that of ATW, which was attracting only $35 million per trading day. This means that if large players active in RIG had been accumulating as much as in ATW (compared to the total money that the company is attracting per day), they would have had to accumulate $1.4 billion × (16.4/10.5) = $ 2.18 billion. In other words, RIG large players have been 36% weaker than ATW’s large players (10.5 compared to 16.4). If we now rebalance the large players’ money flow by the proportion of the total daily funds that each company attracts (as stated in the second column of Table 8.2), we can build a completely different figure for the money flow of large players (see Figure 8.7). As we can see in Figure 8.7, Todco (THE) has been the company with the strongest rebalanced large players’ money flow before being acquired. Of course, this is a pure coincidence. But, I still believe that if I wanted to invest in a sector by spreading the risk among different companies, I would comparatively buy more shares from the companies that attract the most buying money from large players.

Reallocation of Funds between Sectors The great principle of sector reallocation is to increase one’s positions in sectors that are gaining momentum and reduce one’s positions in sectors that are starting a downward trend. A sensible question would then be to

354

VALUE IN TIME

FIGURE 8.7 Oil sea drillers: rebalanced large players’ money flow.

wonder if the Effective Volume analysis applied to the stocks of one sector can help predict the future movements of that sector, just like for the 238 reference stocks of Figure 8.4. The test can be done in four simple steps, which must be performed daily: 1. We first need to take for each stock in the sector a number of days dur-

ing which we will count the money flow (Large Effective Volume flow multiplied by the stock price) related to large players. For example, let’s take 50 days. (For Figure 8.4, I took instead 20 days, which moves more quickly than the 50 days figure.) 2. The second step is to calculate for each stock the ratio between the 50-day money flow related to large players and the total money that was invested in the corresponding stock during the same period of 50 days. This gives us the strength of accumulation/distribution by large players. 3. The third step is to weight this strength figure by the weight of the stock within the sector. 4. The final step is to sum up all the strength figures that we obtained

for all the stocks in the sector. We thus obtain the accumulation/ distribution strength related to the sector. The results are represented in the upper panel of Figure 8.8. This is an image of the average strength used by large players to move the sector’s

Market and Sector Analysis

355

FIGURE 8.8 Oil sea drillers: money flow–profit comparison.

profit up or down (the sector profit is defined as the change of the weighted price of all the components of the sector since the start of the analysis period, which is June 19, 2006, in Figure 8.8). The lower panel of Figure 8.8 shows the actual movements of the sector. We can first see on the upper panel of Figure 8.8 that the sector’s large players’ strength is sometimes close to 10 percent. This means that large players have been very keen buyers of the stocks of the sector (there was no heavy selling). If you remember Chapter 3, when we studied the Effective Ratio, we saw that this Effective Ratio signal itself is usually rather low

356

VALUE IN TIME

(below 6 percent). This is due to the fine equilibrium between large buyers and large sellers: The difference between them is only a few percentage points of the total volume that was exchanged over the prior 50 days. With this 6 percent figure as a reference from my personal experience, 10 percent looks very strong. The second important point about Figure 8.8 is that large players’ strength seems to be a good predictor of the sector’s movements:

r Point A comes before the top in sector profit (the large players’ 50-day strength trend turned down some days before the sector profit trend).

r Point B shows that strength is already growing before the sector profit starts its new uptrend.

r Point C shows that large players started to become weak again, although that is not reflected in the sector profit. I said large players’ strength seems to be a good predictor, because in reality, large players’ strength is not predicting anything! It is just showing when large players are getting stronger or weaker. Therefore, the way to time your entry into a sector is not by looking at the large players’ moves, but by looking at the sector’s moves and by complementing that information with large players’ strength shifts. It is therefore useful to summarize five sector trading rules, and see how they can be applied to a few examples: 1. Buy if the sector’s trend is flat or moving up from a bottom and if it was

preceded by a strength increase by large players. 2. Do not buy when large players are showing weakness. 3. Sell when you believe that you have enough profit. 4. Sell when the sector’s trend is turning down on weakness by large

players. 5. Do not sell yet if the sector’s trend is still positive, even if large players

are showing weakness, because funds take more time to get out of positions and they start doing it while the price is still in an uptrend. Software Companies Figure 8.9 shows the price evolution—shown as the percentage profit/loss—of each company listed in Table 8.3. These companies are by no means a true representation of the software sector, but I have been following them for some time. (The fact that we cannot distinguish the different companies in Figure 8.9 is unimportant. We may notice, however, that stocks in this sector do not follow uniform movements like those of the oil sea drillers.)

357

Market and Sector Analysis

FIGURE 8.9 Software: profit/loss comparison.

Let’s examine points A through D at the lower panel of Figure 8.10, which shows the comparison between the large players’ strength and the profit evolution of the sector:

r Point A: The sector is turning up, while large players’ strength is firmly positive. It is a good idea to buy.

r Point B: The sector is turning flat after a steep climb, with large players’ strength slightly negative. We may wait or sell, but not buy.

r Point C: The sector is turning negative, with large players’ strength strongly negative. We must sell.

r Point D: The sector is turning positive after a three-month decrease, while large players’ strength has been strongly positive for about two months. We may buy. TABLE 8.3 List of Software Companies Company

Ticker

Cadence Design Systems Adobe Systems Electronic Arts BMC Software Oracle Symantec Corp. Verisign

CDNS ADBE ERTS BMC ORCL SYMC VRSN

358

VALUE IN TIME

FIGURE 8.10 Software: money flow–profit comparison.

Real Estate Figures 8.11 and 8.12 show the individual rebalanced money flow and the percentage profit/loss, respectively, for each of the home builders (that I also refer to as the real estate list) listed in Table 8.4. Figure 8.11 is difficult to use, because it doesn’t visually catch the accumulation/distribution movements by large players. The change of large players’ strength as shown in the upper panel of Figure 8.13 is easier to understand. We can indeed see that at the left of the curve, the large players’ strength was the strongest when the profit was the lowest (the

Market and Sector Analysis

359

FIGURE 8.11 Real estate: rebalanced large players’ money flow.

profit is calculated as a price change since June 8, 2006). This accumulation weakened together with the rise in profit. A weaker accumulation simply means that less money was moving in the sector, even if the sector’s profit move was still positive. Note the increase in large players’ strength that started about two weeks before point A. This renewed strength reversed

FIGURE 8.12 Real estate: profit/loss comparison.

360

VALUE IN TIME

TABLE 8.4 List of Home Builders Company

Ticker

Centex Corp. KB Home Lennar Corp. Pulte Homes, Inc. Toll Brothers

CTX KBH LEN PHM TOL

FIGURE 8.13 Real estate: money flow–profit comparison.

Market and Sector Analysis

361

down at the end of April 2007, when some home builders published earnings lower than expected. At point B, the profit is again in a downtrend, but this time backed by much weaker large players.

WHAT WE LEARNED IN THIS CHAPTER We learned that it is entirely possible to adapt the Active Boundaries and the Effective Volume tools to study both the market and the sector evolution. At this stage, more data and more computing power are necessary to come up with the publication of real funds’ accumulation/distribution figures linked to all of the existing sectors. I believe that such data could be useful for deciding how to reallocate funds between sectors. Since I wrote this text, I executed a more thorough research in the filed. Recent results of sector and market analysis can be found on the web site: www.effectivevolume.eu.

Conclusion

n a speech given at the Monetary Economics Workshop of the National Bureau of Economic Research Summer Institute in Cambridge, Massachusetts, on Tuesday afternoon, July 10, 2007, the Federal Reserve chairman, Ben Bernanke, stated: “Undoubtedly, the state of inflation expectations greatly influences actual inflation and thus the central bank’s ability to achieve price stability.” In other words, the Fed is monitoring inflation expectations as much as inflation itself. This is also what the stock market is all about: expectation. In writing this book, my goal was to explain how the market’s expectation evolves using tools that stay as close as possible to fundamental market forces: the motives of market players that may explain why they buy and sell stocks at a specific moment, and the supply/demand balance.

I

A SHORT REVIEW Having taken these principles as the basis for my research, I detailed in each chapter at least one new concept (sometimes several), and then summarized it (or them) at the end of the chapter. After rereading these summaries chapter by chapter, I noticed that while factual, they do not include my own opinion of what works and what doesn’t. I will now briefly review these chapters and give you my opinion, which derives from my experience in using the results of my own research. First of all, I need to say that in my real, live trading, I do not produce the exceptional returns displayed in Chapter 7. I hate losing the family money, and therefore use very tight stop loss levels of 2 percent to 4 percent. This allows me to pick up a few good opportunities, but I miss many others. Overall, the returns have been excellent, while keeping me out of danger during the most turbulent periods—for example, during August 2007 and January 2008.

363

364

CONCLUSION

Large/Small Effective Volume Flow (Chapter 1) This tool is excellent for detecting funds’ accumulation/distribution of shares. However, used in the short term, I found it misleading: Indeed, markets are very efficient, meaning that strong buying is accompanied by an almost equally strong selling force (refer to Chapter 4). The Effective Volume tool measures the fine balance between buyers and sellers, which, as we saw, quite often represents only a few percentage points of the total exchanged volume. This means that in the short term, one fund could decide to sell more quickly than other funds are accumulating. This could result in a small one- or two-day Effective Volume pattern that could eventually contradict a previously established stronger pattern and that could eventually reverse. This is sometimes disturbing for traders accustomed to trading mainly on the price pattern. My advice when receiving contrarian signals is to look at the long-term trend in Large Effective Volume flow (the 40-day or even 60-day trends). This long-term trend in Effective Volume does not fail to show the underlying support of the price trend.

Active Boundaries (Chapter 2) This is one of my favorite tools, because it greatly helps in assessing the short-term value of a stock. My preference is for stocks that are part of a well-known stock-picking list (such as the Investor’s Business Daily list compiled by William J. O’Neil), entering buy orders on stocks that reach the Lower Boundary during a price pullback. You may remember that I explained in Chapter 2 that the adjustment of the Active Float to define the initial Active Boundaries is a difficult process, one that requires consecutive, painstaking refining steps. This is true if you do it manually, but this adjustment of the Active Float can be totally computer-generated.

Large Effective Ratio (Chapter 3) The Large Effective Ratio is certainly a more useful tool than the Effective Volume, for two reasons: 1. The Large Effective Ratio measures the accumulation/distribution over

a period of three to five days; it filters out the short term’s spikes in Effective Volume.

Conclusion

365

2. The current level can be compared to the historical levels of the Large

Effective Ratio, allowing us to judge if the buying/selling is important by historical standards.

Divergence Analysis Tool (Chapter 3) The divergence analysis tool measures for a fixed period of time the difference in strength between the Effective Volume trend and the price trend. The size of this fixed period of time was set in order to adjust for the volatility between price and volume, with the objective of producing a price signal that is stronger than the volume signal. This tool therefore generates three different types of trading signals: 1. When the price is in a strong uptrend that is stronger than—or in

the opposite direction of—the Effective Volume trend, this divergence analysis gives the trader the feeling that the price trend is not sustainable, whereas in most cases it is. We saw many instances where strong price uptrends or downtrends do not reverse easily. In such a case, the divergence analysis signal could make the trader miss the possibility of entering a long price uptrend in the middle of the trend, when there is still lots of upside left. 2. When the price is in a strong downtrend that is stronger than—or in the opposite direction of—the Effective Volume trend, the divergence analysis gives the trader the feeling that the price will soon reverse up, because of the volume accumulation that is taking place. Unfortunately, most of the time this reversal does not materialize, because in a steep downtrend, the first accumulation signs are often short traders who are covering their positions. When short traders have finished, the price may just continue its downtrend. This trading signal could thus be misleading. 3. In my experience, the divergence analysis tool produces its best trading signals when the price is in a trading range. In such a case, a strong divergence simply indicates that strong accumulation or distribution is taking place and that the price trading range would probably break in the direction of the Effective Volume trend. In Figures 6.27 and 6.28 (Chapter 6), I briefly compared the results of the divergence analysis tool to the Effective Ratio tool when they are used as buying triggers. It is clear from these charts that the divergence analysis tool produces lower returns and generates fewer signals.

366

CONCLUSION

For these reasons, I use it as an alert signal that informs me when a strong divergence is taking place; it may be worthwhile to monitor, but I do not use it as a trading signal.

The Supply Analysis Tool (Chapter 4) The supply tool shows when a stock has been so strongly sold off that the probability of finding new sellers becomes increasingly low. Chapter 6 has shown that this tool, combined with the Effective Ratio tool, provides for very good trading opportunities. Because the Effective Ratio tool does not distinguish between the buying that results from short covering and the buying that results from real buyers, I mainly use the supply tool in combination with a support line: I buy when the stock price forms a second or third bottom in an increasing Effective Ratio trend. The supply tool also renders other services that I did not elaborate on in Chapter 4. Indeed, since it is based on a mathematical model of the readiness of shareholders to sell their shares, depending on the timing and the price of their purchase, it can be used to operate a sensitivity analysis, more commonly known as a what-if analysis: What if the price increases or decreases by 5 percent? Will this attract more or fewer sellers? This type of analysis, which may eventually be included in a subsequent book, is rather useful during the planning phase of a trade.

The Yearly Expected Return and the Monthly Loss Transferred (Chapter 5) The yearly expected return (YER) and the monthly loss transferred (MLT) to the portfolio, which measure the return and risk, respectively, of a trading strategy, are very simple concepts. The YER is very useful for evaluating a trading strategy independent of stock-picking skills. The MLT is closer to the risk, because it uses the real drawdown of each trade. Once the trading strategy is selected, the YER and the MLT are of no use. I do not use them for live trading.

Automated Trading Systems (Chapter 6) This chapter is separated into two parts: the alert screens and the automated trading systems. For my everyday trading, the alert screens are extremely useful, because they automatically bring up the best trading opportunities. However, I believe that the part on automated trading systems is more valuable in the long term, because through the use of different combinations of a new set of tools, we discover trading common sense: “Buy value at the right time.”

Conclusion

367

The Bonus Section (Chapters 7 and 8) Out of these two chapters, I mainly use Chapter 8 for my everyday trading. The most important for me is Figure 8.4, since it gives a good picture of the underlying accumulation of funds in regard to the overall market price movements.

WHY I DISCLOSED MY METHODS Some of my early readers thought that I disclose too much in the book. However, I think that it would be a mistake not to disclose my findings. The stock market has greatly changed in recent years: decimalization, highspeed communication, automatic program trading, and computing power at the PC level have brought more changes to the stock market in the past six years than during the previous 60. Traditional trading tools will need to be adapted to this new reality. Within a few years, technical analysis will be very different from the technical analysis we have been accustomed to seeing in the preceding century—and that century ended just eight years ago. Many more tools than those I have presented here will soon become available. For me, it is more important to be a part of that change than to make money behind a computer screen.

MARKET MANIPULATIONS? When I started my research, I had the strong feeling that markets were heavily manipulated:

r I started my research at the time when the accounting improprieties of Enron and WorldCom were unfolding.

r In general, companies did not disclose the various stock options incentives that they offered, and these were often not included in the earnings statement, which meant that these companies’ value was often misrepresented—since part of that value had been promised to third parties through the options incentive programs. r Insiders were continuously trading ahead of the news. r Naked short selling was authorized. (Naked short selling is the act of selling a stock short without first borrowing it from another shareholder.) r Funds had the power to control the stock price in order to accumulate a position on the cheap.

368

CONCLUSION

r As of December 2007, problems with the subprime loan market are unfolding. Debt rating agencies had been generously labeling certain debt packages as “investment-grade.” These debt packages, which we now know are not investment-grade, were revealed to be carrying more risk than was suggested by the rating. Now, however, I feel more confident about the markets:

r In Chapter 4, Figure 4.17 shows that markets are very efficient. This renders control of a stock price virtually impossible—at least for welltraded stocks. r I now have the tools to trade ahead of the news. I will not be the last one to know when something is happening in the market.

WHAT’S NEXT? After reading this book, the question that you should ask yourself is: What’s next?

r If you intend to buy a trading platform, it may be wise to wait until new platforms come out that include modern technical analysis tools.

r In Chapter 6, I list the different traditional technical analysis tools that can be used to measure a stock’s value as well as the trading triggers. You may try to adapt my trading principles using these traditional tools. r After due consideration, I have decided to give free access to the Effective Volume tool introduced in Chapter 1. More information is available on the web site www.willain.com. I also intend to open a discussion group on the subject of volume analysis and might gradually disclose my other tools to participants. If you want to participate, drop me a line. Knowledge is not a “zero sum game”: I believe that sharing knowledge will generate many new ideas. My dream is that, someday, this discussion group evolves into a community of traders and researchers that will collectively bring my work to another level that will benefit everyone.

Conclusion

369

THE LAST WORD Before closing this book, I have three important messages: 1. What counts is value: To make money, you must buy value at the right

time. Therefore, if you have some free time to study the market, please spend it by studying the value of the equity you buy. I did not write about fundamental analysis because it is a well-known subject, but value detection is generally how the money is made. It is sometimes wiser to buy a base metal mining stock that trades at a price-earnings ratio of 3 than to find technical patterns that will net you a quick 10 percent. 2. The market does not care about your opinion. Even if you think that it should move in one direction, the market may not agree with you. It is best not to have any opinion, but to stay open and keep your alert level high to the changes that are about to come. Stay humble and listen to the market. 3. Generating a profit should not be a goal of your trading activity, but rather a result of the improvements that you make in your trading methods. Do not be afraid to learn from others, but also to share your ideas. Sharing knowledge brings knowledge! My wife Michiko has told me on several occasions that she wants me back in the family. She told her friends that over the past few years she had become a “computer widow.” With this long research and writing period finished, it’s now time for me to focus again on my family. Trading the market is always a pleasure, but where does it fall on my list of priorities? Where does it fall on your priority list?

Data Providers

he United States is well-known for its openness in terms of supply of data. The suppliers listed below offer minute data that can be imported into MS Excel. Historical Quotes Downloader (HQD) is a software package sold for $49.95 by the company Ashkon Software L.L.C. (www.ashkon.com). HQD allows downloading minute data from up to 1,000 stock symbols, dating back 20 days. Data is 20 minutes delayed, and sometimes you need to wait 8 hours after the market closes in order for HQD to collect the exact volume data from the different exchanges. The company announced at the end of December 2007 that although the software sales activities will not be interrupted, no further technical maintenance will be offered, since the company is changing its business model. I have been using HQD for many years. Tickdata (www.tickdata.com), a service offered by Nexa Technologies, Inc., a provider of online and direct access trading platforms and electronic order routing solutions, offers historical intraday data for up to 16 stock exchanges. The database dates back to many years. I have used Tickdata’s data for the backtesting of Chapter 6. I found their data to be very good quality, although the importing and reformatting process is quite long. Historical data costs $18 per year and per stock symbol. Additional fees and stock exchange fees must be added for life data. IQfeed (www.iqfeed.net) is a service offered by DTN Market Access, a provider of web content and data feeds for a variety of companies in the agricultural, energy, and financial industries. IQfeed provides historical and life intraday data for nine different stock exchanges. Minute data is available for the last eight months. The service is charged $55/month, to which a monthly stock exchange fee must be added. The service must be programmed in order to provide a continuous data feed in the requested format. I am in the process of testing this service.

T

371

372

DATA PROVIDERS

Opentick Corporation (www.opentick.com) offers free real-time and historical market data for U.S. exchanges. Minute data is available for the last five years. Users, however, must pay the monthly stock exchange fees. The service must be programmed in order to provide a continuous data feed in the requested format. I am also in the process of testing this service.

Sources

Appel, Gerald. Technical Analysis: Power Tools for Active Investors. Upper Saddle River, NJ: Financial Times Prentice Hall/Pearson Education, 2005. Elder, Alexander. Come into My Trading Room. New York: John Wiley & Sons, 2002. Elder, Alexander. Entries & Exits: Visits to Sixteen Trading Rooms. Hoboken, NJ: John Wiley & Sons, 2006. Elder, Alexander. Trading for a Living. New York: John Wiley & Sons, 1993. Elder, Alexander. Sell and Sell Short. Hoboken, NJ: John Wiley & Sons, 2008. Harris, Larry. Trading & Exchange: Market Microstructure for Practitioners. New York: Oxford University Press, 2003. Lhabitant, Franc¸ois-Serge. Hedge Funds: Quantitative Insights. Hoboken, NJ: John Wiley & Sons, 2004. O’Neil, William J. Investor’s Business Daily. www.investors.com. Ord, Tim. The Secret Science of Price and Volume: Techniques for Spotting Market Trends, Hot Sectors, and the Best Stocks. New York: John Wiley & Sons, 2008. Prigogine, Ilya, and Isabelle Stengers. La nouvelle Alliance. Paris: Gallimard, 1979. Vince, Ralph. Portfolio Management Formulas: Mathematical Trading Methods for the Futures, Options, and Stock Markets. New York: John Wiley & Sons, 1990. Weis, David H. Catching Trend Reversals (DVD). www.Elder.com. Williams, Larry. Long-Term Secrets to Short-Term Trading. New York: John Wiley & Sons, 1990.

373

About the Author

ascal Willain is a private trader who trades only his personal funds, and he uses the very tools presented in this book. He graduated in 1983 with an engineering degree in telecommunications from the Universite´ Catholique in Louvain, Belgium. Pascal then went on to earn a master’s degree in applied mathematics in 1987 from the University of Electro-Communications in Tokyo, Japan. Two years later, he earned an additional degree in business from the Universite´ Libre in Brussels, Belgium. Dr. Alexander Elder featured some of the tools presented here in a chapter of his own book, Entries & Exits: Visits to Sixteen Trading Rooms. More about the Effective Volume tools can be found on the web sites at www.willain.com and www.effectivevolume.com. The author can be contacted by e-mail at [email protected]. Before becoming a trader, Pascal created three companies in such varied fields as consulting, computer-voice interfaces, and parking-garage systems. These companies are now managed by partners. With his wife Michiko, Pascal also created the Nello and Patrasche Foundation, which is dedicated to helping handicapped orphans. For more information, go to www.multilines.be/np.

P

375

Index

Account churning, 15 Accumulation/distribution tactics and methods, 19, 24–32, 33, 64, 66, 85, 116, 130–32, 206, 270, 274, 302, See also Tactical moves Active and passive funds, 211, See also Institutions Active and passive sellers, 213, 216 Active Boundaries, 5, 82, 75–82, 82, 67–111, 266, 272, 274, 276–302, 316, 304–16, 331–34, 341–48, 364 Lower Boundary, 5, 272, 287, 311, 313 new Active Boundaries, 167, 184 Upper Boundary, 5, 164, 272, 287–89 Active Boundaries, how to calculate, 109 Active Boundaries, strategy, 310 Active buying to accumulate a position, 64, 127, 213 Active sell market orders, 185 Aite Group, 65 Alert screen, 6, 366 Alert signal, false early alert signal, 271, 366 Alexander Elder Dr., 2–3, 30, 32, 50–52, 61, 151, 153, 302, 317 American Science & Engineering (ASEI), 305 Amount of capital, 298 AMR Corporation (AMR), 153 Analysis window, 120, 127, 128, 129, 133, 137, 177, 180, 177–82, 214, 273, 344 Analysts, 211

Ariba, Inc. (ARBA), 47, 159 Ask-Bid, 16, 65, 66, 135, 186, 188, 206, 212, 213, 232 Attractive price, 187 Automated computer trading, 2, 65, See Program trading Automated trading systems, 6, 51, 69, 264, 266, 366 Average annual returns, 252 Average downside deviation, 248, 249 Average holding time, 252 Average invested time, 279, 284, 290 Average logarithmic daily return, 230 Average maximum drawdown, 249–62 Average profit per winning trade, 239 Average ROI, 74–76, See also Float ROI Average Separation Method, 52, See Effective Volume:the separation volume Bargain hunters, 179, 184, 194, 276 Bargain investors, 192–94, See also Bargain hunters Bear/bull equilibrium, 60 Becton Dickinson (BDX), 78–80, 162–63 Ben Bernanke, 362 Benchmark, 246, 275 Beta value of a stock, 223 Blue-chip companies, 211 Burke ratio, 262, 287, 290, 295, 309, 314, See also Risk:risk-adjusted performance Buy value at the right time, 302, 317, 366

377

378 Buy/sell balance, 63, 129, 132, 186, 188 Buy/sell operations, 292 Buy-and-hold strategy, 222, 223, 232, 247, 250, 275, 298 Buyers/Sellers equilibrium, 213, See also Buy/Sell balance Buying momentum, 180 Buying Pattern Analysis, 135 Buying trigger, 6, 295, 301, 303, 315 Calculating the Large and the Small Effective Volume flow, 122 Candlestick analysis, 1, 316 Casino games, 227 Catastrophic losses, 164 Catastrophic situation, 193, See also Catastrophic losses Celgene Corporation (CELG), 151 Chaikin money flow indicator, 132–34, See also Trading:traditional analysis tools Cheap uptrend reversal, 276 Chesapeake Energy (CHK), 116 Chico’s FAS (CHS), 102–4, 207, 210 Churn rate, 235 Cognizant Technology Solutions (CTSH), 49 Combining Divergence and Active Boundaries, 155, 167 Comparing the divergence signal to the price trend, 184 Comparing the present price to a past price, 192 Compounding effect, 230 Computing time, 267, 274, 297 Consecutive occurrence of losses, 254, See also Catastrophic losses Convertible debenture, 76 Correct buy signals, 201 Darden Restaurants Inc. (DRI), 124, 118–29, 136–44, 177, 178, 179, 214 Dead Cat Bounce, 107, See also Downtrends and Active Boundaries Decimalization, 1–2, 7, 16, 17, 15–19, 54, 64, 187–91, 212, 217–18, 367

INDEX

Demand/supply equilibrium, 12, 34, 101, 114, 135, 186 Departing shareholders, 103, 194 Detect a change trigger, 275, See also Large Effective Volume, See also Large Effective Ratio Detecting tops, 135 Determination of the overall trend, 38 Deutsche Bank AG, 207, 210 Diluted number of shares, 76, See also Float:available float Discovery of value, 6 Distressed stocks, 305–6, 314, 318 Distribution of returns, 246 Divergence buy zone limit, 142, 164, 166 divergence analysis examples, 144 divergence peaks, 138, 153, 155 divergence troughs, 142 historical maxima, 273, See also Divergence:buy zone limit historical minima, 273, See also Divergence:sell zone limit sell zone limit, 142 selling pattern analysis, 139 Divergence Analysis, 5, 58, 86, 113, 114, 127, 266, 273–74, 300, 326–31, 365 Divergences comparison of past divergences, 135, 184 Diversification, 207, 210, 297 Don Worden, 65 Downside frequency, 248 Downside risk, 247 Downtrends and Active Boundaries, 102–8, 114, 137 Downtrends and divergence analysis, 176, 201 Downtrends and Effective Volume, 202 Drawdowns, 250, 252, 286, 295, 309 Dried-up demand, 194 Dynamic method to measure the supply/demand equilibrium, 191 Early profit taking, 287 Early profit-taking tactics, 311

Index

Earnings per share (EPS), 76 Earnings surprise, 67, 88 Earnings-related insider move, 47 Effective Ratio, 58, 114, 127, 128, 365–66 Large Effective Ratio, 178, 181, 203, 273, 282, 300, 364 total Effective Ratio, 129, 142, 143, 159, 177, 326, 330, 340 Effective Volume, 40–41, 54, 66, 114, 129, 131, 216, 266, 296, 326, 321–26, 364 Large Effective Volume, 35, 40, 266, 268, 274, 280, 300 Large Effective Volume flow, 282 separation volume, 34, 38, 52, 53, 56, 124 Small Effective Volume, 35 Small Effective Volume, 214 total Effective Volume, 266 Effective Volume at key turning points, 79 Effective Volume flow, 121 Effective Volume trigger, 277 Elimination of the Price Trend Condition, 284 Elliott wave analysis, 1 Empty trading minutes, 183 End-of-day data, 12, 26, 29, 59, 64, 132, See also Minutes data End-of-day indicators, 20, 63, 64, See also Trading:traditional analysis tools Entry strategies, 275 Envoy Communications Group, 107, 108 Equi-Power Separation Method, 54, 56, See Effective Volume:the separation volume Excel add-on, 267 Exit strategies, 275 Expectation, 272, See Traders:trader’s expectation Expensive starting downtrend, 276 Exponential average function, 120, 121, 139 Extreme volumes, 177

379 Fallen angels See also Distressed stocks Falling Knife, 105, See also Downtrends and Active Boundaries Federated Investors Inc. (FII), 41, 155 Fibonacci retracements, 1, 316, See also Trading:traditional tools Financial disaster, 199 Finding bottoms, 135 Finding long-term value, 275, See also Active Boundaries, See also Supply Analysis Finisar Corporation, 101 Fixed Separation Method, 52, See Effective Volume:the separation volume Float active float, 76–101, 101, 104, 109, 343, 364 available float, 76, 77, 199, 201, 303, 305 float ROI, 74, 77 Force Index, 61, See also Trading:traditional analysis tools, See also Alexander Elder Dr. Franc¸ois-Serge Lhabitant, 221, 262 Fund managers, 4, 114, 128, 135, 206, 210, 245, 249, 266, 267 Fund reallocation between sectors, 350, 353 Fundamental analysis, 11, 39, 211, 369 Funds, 3, 4, 5, 6, 72, 135, 204, 276, 285, 293, 298, 301, 307, See also Institutions Future earnings estimates, 67 Gerald Appel, 51, 316, See also Moving Average Convergence/Divergence (MACD) Gibbons Burke, 262, See also Burke ratio Good trading strategy, 247 Grandmother analogy, 27 Group of shareholders, 274

380 Hedge funds, 19, 211 Herd behavior, 58 Highfliers group, 224 Historical divergences, 136, 138 Historical reference levels, 273, 274, 365 How do funds make money, 211 How to antagonize all the long-term shareholders., 193 How to turn a profit, 180 Illiquid markets, 205, See also Funds Ilya Prigogine, 13, 20 IMAX Corporation, 88, 89, 90, 164–70, 192, 193, 194, 195 Imposing a five-day time limit, 295 Incorrect buy signals, 201 Incremental calculations, 267 Insider trading, 47, 117, 212 Insiders, 3, 5, 7, 43, 66, 159, 367, See also Insiders trading analysis of different types of news, 44 Instantaneous trade execution, 188 Institutional activity, 207 Institutional investors, 1, 3, 19, 20, 63, 66, 76, 106, See Institutions stock accumulation by, 64, 206, 285 Institutions, 22, 33, 57, 207, 210, See Institutional players, See Institutional investors Insurance policy against bad trades, 243 Investment investment style, 87 longterm investment, 87 Investment opportunities, 221, 279, 284, 290, 291, 292, 320 Investor’s Business Daily, 364 Investors long-term value investors, 86, 87, 252 short-term investors, 252 KB Home (KBH), 271, 360 Lack of shares, 201 Laggards group, 224

INDEX

Large drawdowns a measure of, 252, 254, 262 Large Effective Money flow, 347 Larry Williams, 24, 25, 26, 27, 30, 31, 32 Laszlo Biriny, 65 Leaving a losing position, 210 LEV, 270, 271, See Large Effective Volume Lexmark (LXK), 281 Limit orders, 127, 188 Limiting the duration of the trade, 298 Liquidity, 16, 18, 19, 76, 84, 204 illiquid environment, 5 Look into the wallets of potential buyers, 191 Losing trades number of, 238, 242 Loss of opportunity, 239 Louis Pasteur, 1 Low supply limit, 201, See also Supply level Manipulation, 4, 17, 18, 26, 54, 127, 186, 212, 218, See also Price:price manipulation Market analysis, 342–50 Market forces, 5 Market maker, 22 Market orders, 188 Market players, 2 Market testing, 191 Market trend, 7 Market visibility, 1, 16, 17, 54, 187, 212, 217 Mathematical expectation, 227, 228 Mathematical operations necessary for modern trading tools, 75 Maximum drawdowns, 252 Memories of past trades, 264 Meridian Resource Corporation (TMR), 92, 94, 114 MicroStrategy (MSTR), 326, 327, 328, 329, 330 Midsize volume, 213 Milking cows, 222 Milking equipment, 247 Model, 58, 60, 64, 177, 180, 198, 223, 273, 303, 305, 306, 366

Index

Momentum oscillator, 51 Money Flow Index, 65 Monthly loss transferred (MLT), 6, 257, 366 MONTHLY LOSS TRANSFERRED (MLT), 6, 222, 260, 286, See also Monthly loss transferred Morningstar, Inc, 261 Most recent buyers, 197 Moving average (MA), 69, See also Trading:traditional analysis tools MOVING AVERAGE (MA), 69 Moving Average Convergence/ Divergence (MACD), 1, 41, 51, 58, 59, 316 Naked short selling, 367 NASDAQ, 14, 107, 207 News positive or negative, 3, 88, 105, 130, 132, 159, 164, 194, 326 Nonconsenting market, 210 Nontrading Minutes, 181, 182 Normal distributions of returns, 246 Number of outstanding shares, 207 Offering price, 72 Oil drilling sector, 351 On-balance volume, 64, See also Trading:traditional analysis tools Openwave Systems (OPWV), 97, 101, 201, 207 Optimizing a trading strategy, 225, 230, 248, 275 Optimum, 228 Order book, 16, 18, 127, 186, 188, 217 Order fragmentation, 19, 65 Order-placing algorithms, 22 Overbought or oversold stocks, 12, 41, 51, 68, 69, 316 P/E ratio, 67, See Price-earnings (P/E) Paper gain, 72 Paper loss, 71, 325 Passive buying to accumulate a position, 64, 135, 217

381 Past results, 224 Past share accumulations, 117, 203 Pension funds, 207, 210, 350 Percentage of institutional holdings, 207, See also Institutions Perception of the reality, 70, See Traders:trader’s expectation Performance ratio, 227 Pessimistic return ratio, 227 PetroQuest Energy, Inc. (PQUE), 45 Pillar of successful trading, 12, See also Trading:trading pillars Playground Analogy, 38 Pool of traders, 180, 190 Portfolio, 221, 249, 260, 287, 292 Position size, 206 Positive or negative trades, 228, 229, 230, 239, 255, 257 Positive or negative trading days, 228 Potential sellers, 199 Predefined algorithms, 65, See Program trading Present share accumulation, 203 Price catastrophic price drop, 201, 202 cheap stocks, 3 current price, 1 overlapping price zones, 198 price adjustments, 19 price bounce, 107, 185 price breakout, 153, 268, 321 price cycle, 77, 170, 346 price evolution, 1, 5 price fluctuations, 1 price gap corrections, 137, 166, See also Price:Price Gaps price Gaps, 78, 79, 84, 88–91, 92, 103, 120, 132, 136–44, 164, 167, 254, 344, See also Price Gap Corrections price histogram, 189, 191 price inflections, 21, 22, 35, 36, 40, 52, 55, 114, 123–27, 127, 214, 347 price manipulation, 2, 6, 17, 63, 212, 217, 218 price patterns, 316 price precedence trading rule, 14

382 Price (Continued) price rate of change (ROC), 119, 120, 121, 136, 137, 136–43, 159, 176–84, 213, 266, 273, 326, 330, 340, 343, 344 price repartition of volume, 27 price reversal, 194, 305, See also Price:price bounce price spread, 24, 26, 59, 64, 212, 213 price volatility, 138 price/volume based indicators, 58, 59, 64, 132 price-based technical indicators, 58, 59, 69, See Trading:traditional analysis tools, See Trading:classic technical tools Price-earnings (P/E), 67, 76, 193 trend above the 9-day average, 145, 147, 153, 166, 184 Price/Volume relative strength, 113 Priceline (PCLN), 270 Privileged information, 41, 159 Probability of making a profit, 164 Probability of selling, 197 Production Screen, 271 Profit early profit taking, 289, 292 evolution of the average profit, 294, 345 in relation to expectation, 74 linearly increasing average profit, 294 profit growth, 11 profit optimization, 226 profit target, 6, 275 profit targets, 223, 235 Profitability of the trade, 275 Program trading, 17, 77, 111, 367 PRR, 242, See Pessimistic return ratio Public order precedence, 14 Pullbacks, 85, 112, See also Trend:trend retracement Quarterly earnings dates, 316 Ralph Vince, 227, 239 Ratio of invested days, 308

INDEX

Ratio of stop loss trades, 313, See Stop loss Readiness of shareholders to sell their shares, 366 Real estate sector, 358 Relative sector price performance, 350 Relative Strength Index (RSI):, 41, 51, 58, 59, 68, 316 Reliant Energy (RRI), 105, 106, 189, 190, 199, 200, 321–26 Resistance to change, 185, 212, 214, 217 Retail investor, 145, 206, 211, See also Traders:retail players Return consistency, 245 Richard Wyckoff, 59, 316, 317, See also Price:price/volume relationshipbased tools Risk bankruptcy risk, 249, 252, 254 risk control, 210 risk of a portfolio, 222 risk of a trading strategy, 222, 366 risk-adjusted performance, 260 rsik management policy, 73, 87 stock market risk, 222 trader’s behavior risk, 245 trader’s wrong analysis, 224, 245 Risk/return balance, 6, 153, 218, 221, 245, 249–62, 308 Robustness of the trading strategy, 245, 247 Same purchasing or selling power, 38 Sample strategy, 226 Scanning systems, 268–74 Sector analysis, 7, 341, 361 Securities and Exchange Commission (SEC), 321 Selling activity by institutions, 207 Selling momentum, 180 Selling parameters, 278, 280, 284, 287, 288, 300 Selling pattern, 326 Selling pressure, 91 Selling reasons, 226, 325 Sell-off, 192, 194, 333, 336, 338

Index

Sensitivity analysis, 199, 200, 366 Separation volume, 137, See also Effective Volume:the separation volume, See also Effective Volume:separation volume Share availability problem, 201 Shares shares accumulation, 5, 24, 268 shares availability, 72 shares distribution, 5, 24 Sharpe ratio, 260, 261, 287, 290, 309, See also Risk:risk-adjusted performance Short or long-term trade, 266 Shorts covering a short position, 203 short plays, 49, 72, 269, 331 shorting strategies, 7, 86, 304, 320–40 Sierra Health Services, Inc. (SIE), 147 Significant position taken by funds, 23, 199, 201, See also Funds Slippage, 86, 188, 232, 276, 292 Small drawdowns a measure of, 254 Software sector, 356 Spread cost, 1, 16, 18 Standard deviation, 246 Standard group, 224, 250 Starbucks (SBUX), 331, 332, 334, 335, 336, 337, 338, 339 Static resistance, 214 Static supply of shares, 127 Stochastics, 1 Stock strength comparison within a sector, 350, 351 Stock trading new vision of, 2 Stock-picking skills, 221 Stocks That Need to Rest, 102, See also Active Boundaries Stop loss, 6, 71, 85, 159, 207, 245, 257, 313, 314 Stop Loss stop loss levels, 223, 235 Stop-loss limit order, 68 Straight average function, 120 Strategic buying decisions how to catch, 64

383 Strategic moves, 19, 65 String of small losses, 257 Strong accumulation limit, 132, See also Effective Ratio, See also Effective Ratio Strong distribution limit, 132, See also Effective Ratio Successful trading strategies, 299 Supply Analysis, 5, 191, 198, 267, 273, 302–14, 316, 335–40, 366 Supply and Demand, 3, 5, 59, 60, 86, 117, 185, 188, 191, 195, 301 Supply indicator, 274, See also Supply analysis Supply level, 199, 201, 303 Supply of shares, 23, 62, 63, 64, 73, 106, 107, 116, 127, 128, 185, 186, 191, 199, 206, 214, 218, 273, 305 Supply/Demand Equilibrium, 189, See also Supply and demand, See also Dynamic method to measure the supply/demand equilibrium Supply-based strategy, 302–14 Support/Resistance Lines, 51, 59, 68, 189, 190, 304, 316, 366 Tactical moves, 2, 5, 65, 66, 128 Technical analysis tools, 1, 2, 3, 5, 12, 13, 19, 39, 60, 119, 211, 213, 320, 367, 368 Tellabs (TLAB), 27, 34–37, 52, 53, 81, 82, 83, 90, 91, 92, 131–34, 167–76, 196, 198, 201, 202, 203, 215, 216, 217 Temporary illiquid situations, 211 The first trading minute of the day, 137 TIAA-CREF Investment Management LLC, 210 Tick test rule, 7, 321, 349 Tick volume Analysis, 20, 58 Tick Volume Analysis, 65, See also Trading:traditional analysis tools Time frame evaluation, 64, 118, 275 one-minute time range, 22, 64 Time limit trading parameter, 6, 151, 223, 242–45, 260, 305

384 Time precedence trading rule, 14, 16 Time repartition of volume, 28 Timing detection, 199 Todco (THE), 276 Total exchanged volume, 19, 124, 128, 213, 364 Trade time management, 299, 315, 317 trade duration, 239, 250 trade evolution, 293 trades production systems, 268–74 trigger, 299 Traders active traders, 5, 11, 20, 76, 77–93, 104, 110, 186, 198, 276 beginning traders, 2 confirmed traders, 2 how to determine the behavior of traders, 59 individual traders, 4, 267, 297, See also Traders:retail players investors VS traders, 76 latecomers, 179 locked in trader, 71, 302 long-term traders, 275, See also Traders:investors VS traders manipulators, 7 momentum players, 58 passive traders, 128, 185, 213, 214 position trader, 64, 111 professional traders, 1 profit takers, 82, 338 retail players, 53, See Institutional investors short-sellers, 203 short-term traders, 69, 228, 275, 293 skilled traders, 4 speculator, 69 swing trader, 86, 293, See Traders:position trader team of traders, 267 traders erratic decisions, 245 trader’s expectation, 3, 5, 11, 12, 20, 70–89, 104, 107, 186, 188, 342, 362, See also Float:active float

INDEX

traders’ exuberance, 107, 108, 173, See also Uptrends and Active Boundaries traders’ movements, 2 traders’ optimism, 245 traders’ pain, 245 traders’ personalities, 77 traders’ will or intent, 188 trading artists, 2 trend followers, 58 trendsetters, 5, 58, 66, 130, 179, 326 Trading back-test trading ideas, 266 best trades ranking tool, 269 buy/hold trading method, 6 classic technical tools, 27, 50, See Trading:traditional analysis tools day-trading, 188 position size, 6 production screen, 6 swing trading, 86, 87, 315, 318, See also Traders:swing trader trade evolution, 6 trading day, 21, 59, 63, 137, 189, 196, 197, 228, 247, 353 trading opportunities, 6, 108, 184, 274, 290, 304, 318, 366 trading patterns, 1, 2 trading platform, 1, 2, 267, 368 trading range, 17, 40, 41, 91, 112 trading rules, 5, 94, 167, 184, 201, 267, 268, 274, 277, 280, 284, 288, 295, 300, 276–317, 340, 356 trading session, 12 trading strategies, 4, 6, 223, 275–317, 366 trading system, 222 trading tools, 2, 3, 4, 5, 6, 7 trading transactions, 186 traditional analysis tools, 12, 64, 120, See also Traditional chartists transactional level, 20, 64, 65 use common sense, 107, 216, 275, 366 Traditional chartists, 2, 4

Index

Transaction costs, 235, 236, 237, See also Commissions, See also Slippage Transactional data, 65 Trends betting agaisnt a price trend, 112 comparing trend direction, 115 comparing trend strength, 115 Effective Ratio trends, 131 Effective Volume Trend, 121 long-lasting trend, 85 price trends, 3, 5, 67, 118, 166, 267 trend confirmation signal, 270 trend line, 68, 69 trend retracement, 78, 79 trend reversals, 12, 61, 79, 89, 91, 95, 96, 97, 100, 177, 341 Tuning parameters, 223, 226–63 Uptrends and Active Boundaries, 81, 82, 84, 86, 103 Value, 5, 67, 70, 89, 199, 298, 299, 305, 315, 325, 364, 368 Value investors, 192, 193, 194 Volatility, 5, 19, 29, 52–55, 63, 65, 101, 114, 120, 176–89, 211, 222, 223, 246, 260, 317, 365 volatility at the one-minute bar level, 65 volume volatility, 65, 138

385 Volume volume data, 5 volume histogram, 190 volume spikes, 65, 177, See also Volatility volume Weighted by the Price Spread, 63, See also Trading:traditional analysis tools volume-based tools, 59 Weakness Index, 62 Welles Wilder, 51 Westlake Chemical (WLK), 145 What-if analysis, 366 When to buy, 199 Where does the supply come from, 195 Why trends exist, 67, 112, See also Active Boundaries William J. O’Neil, 364 William Sharpe, 223, 260, See also Sharpe ratio Winners and losers, 196 Winning trades number of, 238 Winning/losing duration ratio, 241, 242, 278 Winning/losing trades ratio, 242 Yearly expected return (YER), 6, 222, 229, 285, 290, 293, 295, 299, 304, 306, 313, 36