Preview only show first 10 pages with watermark. For full document please download

Demark, Tom - The New Science Of Technical Analysis

DeMark, Tom - The New Science of Technical Analysis




TRADING APPLICATIONS OF JAPANESE CANDLESTICK CHARTING Gary S. Wagner and Brad L. Matheny FRACTAL MARKET ANALYSIS: APPLYING CHAOS THEORY TO INVESTMENT AND ECONOMICS Edgar E. Peters UNDERSTANDING SWAPS John F. Marshall and Kenneth R. Kapner GENETIC ALGORITHMS AND INVESTMENT STRATEGIES Richard J. Bauer, Jr. TRADER VIC II—PRINCIPLES OF PROFESSIONAL SPECULATION Victor Sperandeo THE NEW TECHNICAL TRADER Tushar S. Chande and Stanley Kroll FORECASTING FINANCIAL AN D ECONOMIC CYCLES Michael P. Niemira and Philip A. Klein TRADING ON THE EDGE Guido J. Deboeck GLOBAL ASSET ALLOCATION: TECHNIQUES FOR OPTIMIZING PORTFOLIO MANAGEMENT The New Science of Technical Analysis Jess Lederman and Robert Klein, Editors Thomas R. DeMark John Wiley & Sons, Inc. New York • Chichester • Brisbane • Toronto • S ingapo re I dedicate this book to those individuals who have contributed emo tionally, physically, spiritually, inspirationally, intellectually, and professionally to my investment career. Without their influence on my life, this endeavor would never have been anything more than a mere fantasy— To my children T.J., Carrie, Meghan, Rocke, Evan, and Dominic, for the time they sacrificed and the patience they exhibited to allow me to complete this project; To my wife Nancy for her willingness to support my efforts and to prevent distractions; This text is printed on acid-free paper. Copyright © 1994 by Thomas R. DeMark Published by John Wiley & Sons, Inc. All rights reserved. Published simultaneously in Canada. Reproduction or translati on of any part of this work beyond tha t permitt ed by Secti on 107 or 108 of the 1976 United States Copyright A ct without t he permission of the copyright owner is unlawful. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold with the understanding that the publisher and the author are not engaged in rendering legal. accounting, or other professional services. If legal advice or other expert assistance is required, the services of a competent professio nal person should be sought. Library of Con gress Cataloging in Publication Data: DeMark, Thomas R., 1947The new science of technical analysis / Thomas R. DeMark. p. cm. — (Wiley finan ce editions) ISBN 0-471-03548-3 (acid-free paper) 1. Investment analysis . 2. Stock price forecasting. 3. Financial instruments—Prices—Forecasting. I. Title. II. Series. HG4529.D46 1994 332.6—dc20 94-18145 Printed in the United 10987 654 States of America To my father Louis and my mother Carmilla, for the direction they provided and the values they instilled; To William R. Johnson and A. Keith Johnson, for the employ ment opportunity of a lifetime; To Paul Tudor Jones and Peter Borish, for their recognition and appreciation of profess ional money management; To Van Hoisington, for the fortitude and foresight to invest in the future; To Larry Williams, for his encouragement to create a lenge myself and for his friendship; nd to chal To Charlie ("D") DiFrancesca and John DiFrancesca, for the in spiration to excel and to accept adversity; To David Baker, Gibbons Burke, Ed Cicosz, Leon Copperman, Joe Generalis, Thomas Henson, Anthony Kolton, Jack Kunkel, John Miller, Brian Pedersen, Dr. Guenther Pfister, John Snyder, Ron Williams, and Bernard C. Ziegler III, for their contributions to my investment career. Preface If you ar e seeking a pa nace a for all your trad ing ills, t his boo k is not your cure. There is, in fact, no infallible investment approach. Extrane ous items outside the realm of market research also con tribute to a trader's performance. Specifically, considerations such as sound money management principles, including both cap ital preservation and strict trading discipline, are key components of trading success. The parameters of this book are limited to re search, and any discussion of money management techniques and of market psychology is incidental to the presentation of market timing tools and methods. This is not to imply that a trader could not survive profitably using solely market timing models, but rather to stress the fact that other variables not addressed in this book play a vital role in distinguishing between a mediocre and an accomplished trader. Throughout my career, I have been fortunate to have been asso ciated as a partner, a consultant, and an employee with many of this generation's most notable investment luminaries and compa nies. I have advised the companies of individuals such as George Soros, Michael Steinhardt, Leon Cooperman, and Laurence Tisch; have worked as executive vice president for Paul Tudor Jones; have established a trading company with Charlie ("D") DiFrancesca; managed a futures fund with Van Hoisington, and created market timing systems with Larry Williams. In addition, I have served as a consultant to the key decision makers at various investment gi ants—Goldman Citibank, Morgan Bank, Discount ration of New Sachs, York, IBM Pension, Minnesota Mining Corpo Pension, Atlantic Richfield Pension, Trust Company of the West, New York Life, and Criterion Fund, among many others—often commanding fees in excess of $100,000 per year from each. Although these viii Preface Preface clients respected my advice, the most successful among them were, almost without exception, those who were able to blend this infor mation with the money management skills and market experience they had acquired throughout years of professional practice. They will acknowledge that they utilize various market timing strategies but ultimately their decisions a re subject to their own predilections. Unanimously, were you to ask these profess ionals to describe their investment style, they would say it was eclectic, relying on the infor mation they deemed important at the time. Your question would be like an athlete's asking basketball legend Michael Jordan to de scribe how he was able to reverse dunk in midair. This talent is not something that is necessarily taught or learned: it's innate an d non transferable. J us t as a composite of skills contributes to an athlete's ability, so does a composite of knowledge, instincts, and experience determine a money manager's capacity to perform. reviewing srcinal work published by others; most information in the public domain was redunda nt and incomplete. I offer to you a response my friend Larry Williams provided me years ago. Regardless of the importance of the market timing infor mation you may give someone, invariably most individuals quickly become disenchante d or distracted, or they fai l to possess the disci pline required to effectively apply it. I would hope that you will not fall victim to these mental lapses and that your success in applying my trading techniques will prove this statement false. I am pleased to share with you the products I have developed during my many years of market research. While I may have dis cussed some of these ideas at various semina rs or workshops in the past, this book prese nts a compilation of my work f or the first time. I alone am responsible for the research discussed. I think you will find this material srcinal and timely even though some of the ideas Were it sufficiently simple for trading success to be translated into a series of formulas, trading would no longer be the challenge it is. It is my ambitious goal to provide you with the methodology re quired to systematize market timing techniques. My hope is that you will be able to integrate my research ideas and experience into your trading protocol. Many of the timing techniques t hat you may have previously perceived either to be deficient or to have been ren dered obsolete may now become valuable to you with the enhance ments I will provide. Why have I chosen to share my research and ideas with you, given their reputed value? Let me cite an incident that occurred a few years ago. At the behest of my partner, I visited the Houston of fice of an individual who promoted market timing systems. Immedi ately upon my arrival, he engaged in a calculated campaign to sell me his services. To convince me of his allege d mark et timing exper tise, he claimed authorship of two timing systems that I had created and my partner had offered for public sale the previous year! It was a shock to me since I was the one who had created these systems. Rather than argue with him regarding his representations, I de parted questioning his legitimacy and his integrity. Since that inci dent, I have witnessed other individuals claiming ownership rights to other techniques I have created. By publishing thi s book, I intend to place my signature on t he ideas presented, to protect my rights as creator and as owner. I believe it is important that I make a significant contribution to the body of market timing information. Unfortunately, at the time I conducted my market research, I did not have the luxury of were conceived and researched years ago. Hopefully, you will be as delighted with the book as I was in its preparation. ix THOMAS R. DEMARK Although the author and the publisher believe the information, data, and contents presented are accurate, they neither guarantee their accuracy or completeness nor assume any liability. It should not be assumed that the methods, techniques, or indicators pre sented in this book will be profitable or that they will not result in losses. Trading involves the risk of loss as well as profit. Past per formance is not a guarantee of future results. Trademarks Foreword The following are trademarks held by Thomas R. DeMark: Countdown D-Wave Daily Range Projections DeMarker Magnet Price Price Countdown Price Intersector Price Setup Range Expansion Breakout Range Expansion Index (REI) REBO Sequential Setup TD Breakout Qualifiers TD Channel TD Demand lin e TD Dollar Rated Option Ratio TD Line Breakout TD Line Value TD Lines TD New High-New Lo w Index TD Points TD Price Points TD Price Projector TD Rate of Change TD Retracement Arc TD Retracement Qualifier TD Supply Line TD Supply Points Trend Factors When I first received a copy of Tom DeMark's new book on technical analysis, I was a bit puzzled by the title. Af ter all, technical an alysis isn't "new." I've even written a couple of books on the subject myself. After reading through the material, however, I quickly realized that the key word in the title was "science." Technical analysis has al ways had more a rt than science to it. T wo charti sts could loo k at the same ch art of any gi ven stock, and the same group of technical indi cators, only to come up with two completely different conclusions. Much of technical analysis is truly "in the eye of the beholder." For many reasons, this book is particularly timely. For one thing, technical analysis has never been more popular. The prolifer ation of powerful computers, supported by inexpensive software programs, put the arc ane worl d of technical analysis at the fingertips of even the smallest investor and trader. With the growing pop ularity of futures and options trading—and the expansion of those trading vehicles to include stock indexes, Treasury bonds and for eign currencies—traders have been forced increasingly to fall back on technical methods to cope with such* fast moving markets. Intermarket linkages between the four market sectors—commodities, bonds, currencies, and stocks—forced traders to follow a much wider universe of market s. Global linkages between financial mar kets also forced traders to adopt methods that required lightningquick responses to rapid market movements—namely technical analysis. Another major contribution to the growin g popularity of tech nical analysis comes from the television screen. Daily business cov erage on CNBC, which is seen all over the world, includes a heavy dose of technical analysis. Never before have so many people been exposed to daily explanations and analysis utilizing technical Contents Introduction 1 1 Trendlines 5 2 Retracements 59 3 Overbought/Oversold 85 4 Wave Ana lysi s 101 5 Accumulation/Distribution 109 6 Moving Avera ges 129 7 Sequential™ 135 8 Gaps 183 Daily Range Projections 193 9 10 Rate of Change 197 11 Equities 215 12 Options 227 13 "Wald o" Patt erns Conclusion 233 241 Index 243 Introduction It's amazing how precise the formulas and the contents labels for health p roducts m ust be and how clo sely they are monitored to e n sure the physical safety of consumers. On the other hand, it is dis turbing that no such prescription or safeguard exists to guarantee the financial safety of investors' assets by requiring thorough test ing and analysis of the techniques employed to trade their funds. It's a sad commen tary to see supposedly wel l-educated, trained in dividuals risk huge sums of money applying artistic and totally subjective methods to arrive at trading decisions. Surely enough historical and profitable observations have been made to justify the use of various m arket timing approaches. Unfortunately, most cases ar e identifie d retros pective ly; when the h uma n eye is trai ned to select only those examples that work best and to overlook the many unprofitable ones. Too often, analysts fail to dissect techniques in order to ferret out what is important and what can contribute to their trading suc cess. Throughout many years, I have isolated a number of the key components of various market timing approaches. With the help of cha rts, data, and other empirical observations, I wi ll iden tify these important elements, as well as many others. This book should en able you to become a successful trader. Although it is not essential that you accept all the techniques and suggestions presented in or der to improve your trading capability and versatility, proficiency in a few should help considerably. 2 Introduction There are three distinct approaches to chart anal ysis. The first is casual and subjective and is based on the interpretation of a price chart using "gut feel" or, more precisely, guesswork. Most traders operate on this simplest level because it requires no rigorous analysis or justification. Unfortunately, for the sake of expediency they sacrifice consistency and logic. The second approach creates market timing indi cators that identify the price levels generally associ ated with overbought/oversold zones. Although a number of traders subscribe at least partially to this type of analysis, typically these individuals not only limit the scope of their research to various widely fol lowed indicators but also practice generally accepted method s of interpretation. In other words, they totally lack creativity and fail to attempt either to develop their own stable of indicators or to make improve ments to existing indicators. Furthermore, they often possess inflated expectations regarding the value of these indicators and they fail to appreciate their limi tations. The most effec tive an d most valuable ap proach is the development of systems that actually generate buy and sell signals. Few analysts possess the back ground, the experience, and the willingness to devote the time and energy necessary to acquire this exper tise. I intend to discuss the evolution from the artis tic first level to the sophisticated, mechanical third level, and I will highlight examples describing the benefits of applying both proven systems and disci plined techniques. President Herbert Hoover used to remark on his relentless search for a one-armed economist who was incapable of qualif ying his forecasts with a state ment tha t bega n, "On the other han d " The implied complaint w ould apply to most market a nalys ts today. I have often observed that many analysts have a unique ability to talk out of both sides of their mouths. With the techniques I will share with you in thi s book, the opportunity to equivocate no longer will exist. Ide ally, not only will you cease to be dependent on others' advice but also you will be equipped to assume total responsibility for all your trading activities and deci sions. In other words, the proverbial "pointed finger" will be directed toward yourself. The procedures and Introduction 3 rules that are critical to trading independence will be presented and explained in detail. For vague and poorly defined trading techniques, I will substitute definitive and clear steps to market timing success. No longer will you be afforded the luxury and the ex cuse of trading with a rearview mirror. The simplistic and risk-free app roach of retrospectively and subjec tively identifying buy and sell levels will be replaced with the skill and know-how required to evaluate en try and exit points that are prospectively and me chanically ideal. The maturation process from "chart artist" (chartist) to chart scientist will have begun in earnest. I recommend that, as you read this book, you con centrate on and introduce into your trading regimen only those elements with which you feel comfortable and which are compatible with your trading style. Most of the techniques and ideas presented in the fol lowing chapters reinforce on e another, but they are so dissimilar that you may elect to study and perfect only a few at a time. Keep in mind that these ideas evolved over a period of more than 23 years spent in market research, as both a vocati on and an avocation. Conse quently, mastery of these topics should not be ex pected immediately; they will require your undivided attention and total concentration. I suggest that you maintain a reasonable pace studying the numerous techniques and concepts presented and that you not be discouraged by an inability to totally grasp all the details and nuances of the subject matter immedi ately. The format in which the various topics are pre sented allows for both intensive and comprehensive study. At the same time, because of the diverse natu re of the chapters, you are afforded the opportunity to concentrate only on a reas t hat are of specific interest, without the necessity of referring to and understand ing other unrelated information. Most of the ideas and concepts presented through out this book are unconventional, unorthodox, and for eign to what most traders have learned and practiced in the past. They are srcinal, fresh, and they cover many area s of the discipl ine of market timing analy sis. In some circles, I expect, I will be characterized as a trading iconoclast who is shattering many time-worn practices and norms. My only wish is that readers 4 Introduction accept these novel approaches in the context in which they are both intended and presented. They are new and exciting investment timing tools designed to sup plement, to upgrade, and to complement the current group of trading methods. For beginning traders, the book will provide a solid and valid foundation on which to develop a tr ading rese arch back ground. The techniques described in this book have been prepared and designed for a trading audience. My ex perience confirms that most of the concepts discussed have universal application with equal success to o ther fields wherein any series of data or graphic presenta tions are readily available and studied. I believe that almost any discipline that can be quantified and that lends itself to trend analysis is a potential candidate for this type of research discussion and application. Specifically, I have calculated retracements, projec tions, and objectives for data in diverse areas ranging from interest rates and other economic statistics to forecasts of the migration tre nds of birds. I encourage you to examine thoroughly and critically my trading techniques and, should you desir e, to experiment with and explore the possibilities of their application to other fie lds. At the same time, I challenge you to make enhancements to my research. My biggest complaint has always been that most traders are like the split ends on football teams rather than the quarterbacks: they are capable only of receiving information, not of supplying it. I have witnessed the evolution of market timing research from a simple Bowmar calculator to the cur rent preoccupation with such exotic, high-technology analyses as artificial intelligence, chaos theory, opti mization mo dels, neur al networks, and so on. The on slaught of this advanced mathematical theory and of elaborate computer capabilities has fostered a disin terest in uncomplicated, basic, "blue-collar" market timing techniques and devices. However, even as these sophisticated approaches have instilled a false sense of trading security, they have failed to reward their advocates with markedly improved performance re sults. Consequently, I predict a return to the simple, pure analytical approaches and hope that this book and the trading suggestions contained in it serve as a catalyst to expedite this revival. Trendlines Whether a trader is a practitioner of fundamental or of technical analysis, invariably, at on e time or another, he ha s relied on trendlines to make his forecasts. Although trendlines are universally used, it is surprising how dissimilar they are in construction and interpretation, and how subjectively they are applied. Not only is it commonplace for different analysts to draw different trendlines representing the same data during the same time period, but the same individual, on separate occasions, will also draw two totally different trendlines based on the identical information, depending on his inclination each time. Consisten cy and uniformity are totally lacking. Not all the tre ndlines ca n be correct— only one is. Through exhaustive, pain staking research and years of exper ience and appli cation, I have arrived at an effective method to select the two critical points that are essential to the proper construction of a trendline. Once learned and applied, analysis is no longer subjective; instead, it becomes totallytrendline mechanical. Trendline breakouts are precisely defined and price objectives can be easily calculated: sys tems can actually be created. Price gaps and large price range moves assume a significance never before imagined. 6 Trendlines Selection of TD Points™ and Construct ion of TD Lines™ 7 Selection of TD Points™ and Construction of TD Lines™ Supply and demand dictate price movement. Specifi cally, should demand exceed supply, price advances; conversely, should supply exceed demand, price de clines. These are basic economic tenets accepted by all economists. In order to illustrate this phenomenon pictorially, analysts construct a descending line to represent supply and an ascending line to represent demand (see Figures 1.1 and 1.2). The difficulty, when creating these lines, involves the specific points to select and connect (see Figure 1.3). As it often does, human nature interferes in the Source: Logical Inform ation Machines, Inc. (LIM}, Chicago, IL. Figure 1.2 Observe the ascending price movement as depicted by the upsloping "demand" line as well as the series of both higher price highs and lows. proper const ructi on of these lines. For example, we ar e accustomed to review the historical price activity of a market—from the past to the present, with the dates reading from left to right. As a result, the demand and supply lines are drawn and extended from the left side of the ch ar t to the right. Intuitively, th is is incorrect . Recent price activity is more significant than histori cal movement. In other words, precision and accuracy demand that the lines be extended from right to left, Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figur e 1.1 Mote the declining price movement as defined by the downsloping "supply" line, as well as the pattern of both lower price highs and lows. with the most recentthis date appearing the right side the chart. Initially, may appear at unorthodox but, of in actuality, my experience and numerous observa tions confirm this approach. Simplicity and ease of construction should never serve as substitutes for 8 Selection of TD Points™ and Construc tion of TD Lines™ Trendlines 9 changed my analytical life. Being fledgling traders, both he and I were consumed by the activity of the markets. Not only was the analysis of price behavior our profes sion, but it had also become our total obses sion. Subsequent to leaving the office and returning home, we would preoccupy our evenings discussing interesting char t price pat tern s over the telephone. On one such occasion, we discussed at length the inter play of a series of price trends. We each drew the trendlines on our own charts. When we arrived in the office the next day and compared our respective charts, however, the lines on the charts did not even come close to resembling one another. This bothered me greatly. It was as if we had been speaking two en tirely different languages to one another. I was deter Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.3 Figure 1.3 It's obvious that many lines can be drawn to to establish establish the the price price trend. The key elements are to select the two critical points, construct the correct trendline, and ignore the many others. logic and accuracy. Imprecision and total disregard for detail are reflected in the common practice of con structing multiple trendlines as well as in the typi cally cavalier attitude of analysts who believe that one of these lines will accurat ely define the tren d. Suc cess in using trendlines requires both an attention to detail and a pattern of consistency. Rather th an merely presenting a set of rules de signed to establish the proper method for selecting points and then connecting these points to construct a supply or a demand line, I would like to share with you a frustrating but professionally pivotal experience I had with a business colleague approximately 20 years ago. This episode proved to be a catalyst that mined to prevent this from ever happening again. It was essential that we both understood and commun icated with the same vocabulary and definitions in order to avoid any further confusion and misunder standing. Specifically, I embarked on a journe y to cat alog and standardize widely used market timing techniques, to improve on them, and to create my own. To this day, I strive to accomplish this goal. Trendlines were my first project. For purposes of discussion and illustration, I will generally refer to daily charts and data when in fact any other time period can be easily substituted. My reasons for selecting daily information are threefold: 1. It is the most readily available and has been the most widely used time series for decades; 2. It not only relieves the trade r of the necess ity of constantly following the market on an intraday basis, but it also reduces considerably the risk of price revisions such as those that plague intraday data bases; 3. It increases the chance that when market sig nals are generated based on this information, the price fills will in actuality be executed. 10 Trendlines Early on, I concluded that important supply price pivot points were identified once a high was recorded that was not exceeded on the upside the day immediately before as well as the day immedi ately after (see Figures 1.4a,b}. Conversely, to define demand price pivot points, just the opposite ap proach was employed: a low was recorded that was not exceeded on the downside the day immediately before as well as the day immediately after (see Fig ure 1.5). This made sense to me; these were critical days that proved to be trend turning points in price activity. Supply overcame demand and price declined in Figures 1.4a,b, and demand overcame supply and Selection of TD Points™ and Construction of TD Lines™ 11 price advanced in Figure 1.5. I have labeled these key price points as TD Points. Since my research un covered these price points, I identify them with my initials. In Figures 1.4a,b, the two most recent peak de scending TD Points™ were identified and then con nected to construct the supply line (hereinafter referred to as a TD SupplyLine™); in Figure 1.5, the two most recent ascending low pivot points were identi fied to construct the demand line (here inafter referred to as a TD De mand Line™). It was t ha t simple. No more excuses that I ha d selected the wr ong points. The tech nique was now rigid and objective. Furthermore, the Source: Logical Informa tion Machines, Inc. (LIM), Chicago, IL. Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Figure 1.4(a) Note the highs are circled whenever that particular day's high is preceded the day before and succeede d the day after by a lower high. The supply price pivot points (TD Supply Points) are key levels since price was incapable of exceeding the resistance due to supply. Figure 1.4(b) Supply price pivot points (TD Supply Points) are resista nce points that are defined by a high that is both preceded and succeeded, on the days immediately before and after, by lower highs. These TD Price Points are identified on the chart. 12 Trendlines Refinements to TD Point Selection 13 Source: Logical Information Machines. Inc. (LIM), Chicago, 1L. Source: Logical Information Machines, I nc. (LIM), Chicago, IL. Figure 1.5 Demand price p ivot points (TD Demand Points) are support points that are defined by a daily price low that is both preceded and suc ceeded, on the day immediately before and the day immediately after, by higher lows. The TD Price Points are identifie d on the chart . Figure 1.6 Four potential descending TD Supply Points are identifi ed: A-B is the first supply line. Once TD Supply Point C is formed, however, a new supply line is constructed: B-C. Finally, when point D is defined, the supply line is revised to C-D . As you can se e, the supply/demand ba lance is in a con stant state of flux. Consequently, the supply line adapts to reflect these changes. real attrac tion of this m etho d of point selection was tha t it was dynam ic. In other words, the mar ket itself announced any changes in the supply-demand equi librium equation by continuously resetting TD Points. Consequently, TD Lines are constantly being revised as more recent TD Points are being formed (see Figurofe 1.6) Oncet aga t he Point imp ort (1) the selection the .mos recentin,TD an anc d itse of connec tion to the second most recent TD Point as well as of (2) the construction of the TD Line itself becomes apparent. Refinements to TD Point Selection I have found two modifications to the ID Point selec tion process to be helpful in some instances. Although they are not critical to your success in correctly select ing TD Points, they are presented for your consider ation as well as for the sake of completeness. An important factor when selecting TD Points re lates to both the closes two days before the pivot high 14 Refinements to TD Point Selection Trendlines and the pivot low. In the case of the formation of a TD Point low: 1. Not only must the lows the da y before and th e day after be greater than the lowest low—the low in between—but the pivot low must also be less than the close two days before the low. In other words, should a price gap separate the low one day before the lowest low and the close the day before it, t hat close cannot be less tha n or equal to the lowest low (see Figure 1.7). Source: Logical Information Mach ines. Inc. {LIM }, Chicago, 1L . Figure 1.7 If the price gap defined as the distance fro m A to B—the price low (B) and the previous day's close (A)—is filled in, the low on the following day is no longer a demand point because the low is not less than the true low the previous day. 15 Conversely, in order to identify a TD Point high properly: 2. Not only must the hig hs the day befo re and the day after be less than the highest high—the high in between—but the pivot high must also be greater than the close two days before the high. In other words, should a price gap separate the high one day before the highest high and the close the day before it, that close cannot be greater than or equal to the highest high (see Figure 1.8). As a matter of practi ce, some mark et timers differ entiate between the highs and the lows that appear on a price chart and those that would appear if one were to fill in the price gaps. I coined the phrase for the former as "chart" highs and lows; and charting convention has labeled the latter as "true" highs and "true" lows. Having worked with TD Lines for a number of years, I was able to anticipate when the TD Points se lected might prove invalid. My ability to subjectively validate these points was never defined or translated into decision rules until recently. By isolating the good examples from the bad, I was able to establish prerequisites that enabled me to perfect TD Point se lection. This validation process involved the relation ship between the most recent pivot point low or high and the close the day immediately following it. Specifi cally, if the close the day after the most recent pivot point low is below the calculated value of the TD Line rate of advance, the validity of that low is suspect (see Figure 1.9). Conversely, if the close the day after the most recent pivot day high is above the calculated TD Line rate of decline for that day, the legitimacy of that high is questionable as well (see Figure 1.10). These refinements reduce the frequency of TD Points and , conseque ntly, TD Lines. At selection the sam e time, however, they serve to of validate both the of the TD Points and the utility of the TD Lines in iden tifying support and resistance levels as well as in fa cilitating the process of calculating price projections. 16 Trendlines Benefits Derivedfrom Proper TD Point and TD Line Selection 17 Japanese Yen Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.8 Were the price gap between the hig h and the previous day's close considered, the supply point would not exist. Benefits Derived from Proper TD Point and TD Line Selection Source: Logical Information Machines, Inc. (LIM), Chicago, IL, Figure 1.9 In this instance, the close on the day immediately a fter the most recent pivot point low is below the rate of ascent as defined by the demand line connecting the two most recent ascending demand points. This occurs two times on the chart: lines A-B and A-C. Although it does not invalidate the demand lines, it raises questions regarding their reliability. I was soon to learn that many benefits to this ap proac h had been derived as a result of bot h the proper identification an d the application of TD Point s and of TD Lines. Inexplicable price gaps that had previously appeared out of the blue took on a special meaning and significance. Often, price vaulted above aTD Sup TD Line, it became appar ent that, in the ensuing price activity, a natural rhythm was dominant and was of ten predictable. For example, the extent of the price movement beneath a TD Line is often reflected in a comparable price movement above the TD Line {see ply Line1.11). precisely at its intersection with price Figure Conversely, this phenomenon was (see ob served when price gapped below a TD Demand Line (see Figure 1.12). In addition, once price exceeded a Figure 1.13). Similarly, the degree of the price move ment above a TD Line is often repeated beneath the TD Line (see Figure 1.14). The following discussion de scribes this technique in more detail. 18 Trendlines Benefits Derived from Proper TD Point and TD Line Selection 19 Source: Logical Information Machine s. Inc. (LIM), Chicago, IL. Source: Logical Informatio n Machi nes, Inc. (LIM), Chicago, IL. Figure 1.10 If the supply line construc ted by connecting supp ly points A and B is extended, it intersects the day after supply point B, below that day's close. As you can see, a holiday gap appears on the chart and shifts price activity somewhat, but even if the chart is adjusted to accommodate this fact, the close one day after point B exceeds the line. Consequently, the value of that particular line is questionable. Figure 1.11 See how price gapped upside above the T D Supply Line on the opening, and remained above all day. 20 Trendlines Benef its Derived from Proper TD Point and TD Line Select ion 21 Source: Logical Information Machines , Inc. (LIM), Chicago, IL. Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Figure 1.12 Figure 1.13 The price movement beneat h the TD Supply Line A-B repe ate d only in reverse, once price exceeded the TD Line upside. The difference in price between the lowest price beneath the TD Line and the TD Line value on the same day is added to the breakout to arrive at a price objective. Specifi cally, price X on the TD Supply Line A-B is precisely above point Y, which is the lowest price recorded beneath the A-B Supply Line. By adding that value to the breakout above the A-B Supply Line, price projection Z is calculated. Observe how price opened b elow the TD Demand L ine. 22 Trendlines Source: Logi cal Information Machines, Inc. (LIM), Chicago, IL. Source: Logical Inform ation Machines, Inc. (LIM), Chicago, IL. Figure 1.14 Price movement above TD Demand Line A-D is repeate d in re verse on the downside, once the TD Line is penetra ted. By calcula ting the dif ference between point Y—the highest price above the TD Line—and Point X—the specific price on the TD Line immediately beneath it-—and by sub tracting that value from the breakout below the A- B Demand Line, price pro jectio n Z is de term ined . Figure 1.15 The price activity is attracted to the line that bisects the price movement. At the time I began to experiment with drawing trendlines on charts, I observed that if the price activity continued to pursue the magnetic properties of this line. Having reviewed a considerable number of charts, I isolated as many common denominators as I could identify. You might say that this effort was both a precursor and a "back door" introduction to TD Lines. I mention t his only to illustrate how I uncovered the unique property of symmetry in price movement. Earlier, I discussed the selection of TD Points and the construction of TD Lines. Once you are comfort able with the proce dures required to perform these ex presented ntireamount, ch art were bisected a line separating,onbythe ane equal extreme priceswith above and below that line, prices would often be drawn to and be repelled by that line (see Figure 1.15). I was fascinated. With a limited knowledge of trendlines, I ercises, the phenomenon of price symmetry becomes apparent. Careful inspection reveals that the differ ences between extreme price points immediately above a TD Demand Line and the TD Demand Line it self, as well as immediately below a TD Supply Line Price Projections 24 Trendlines and the TD Supply Line itself, replicate themselves once the TD Line is penetrated (see the discussion of TD Breakout Qualifiers in the last section of this chapter). Although the pat tern itself is never precisely repeated, the extent of the movement both above and below the TD Line often is, an d thi s behavior is wha t I describe as price symmetry. TD Price Projectors There are three distinct methods to calculate price projections once a trendline is penetrated validly; I call them TD Price Projectors. The particular tech nique selected is a function of the degree of precision and accuracy required by the user. TD Price Projector 1, the least precise and the easiest to calculate, is as follows: when price advances above a declining TD Line, usually price continues to advance to at least a price level equivalent to the dis tance between the lowest price value beneath the TD Line and th e TD Line value direct ly above it, a dded to the TD Line val ue on the day of the breakou t to the up side. What may sound complex when described in words is very simple when viewed on the chart (see Figure 1.16). Conversely, the identical symmetry is apparent when price declines below an ascending TD Line. Usu ally, price continues to decline to at least a price level equivalent to the d istance fro m the hig hest price value above the TD Line to the TD Line value directly be neath it, subtracted from the TD value on the day of the breakout to the downside (see Figure 1.17). Often, visual inspection will allow the user to forecast ap proximate price objectives; most traders require more precision, however. Basic arithmetic will enable the userdividing to calculate the rate ofbetween chan ge the of aTD si mply by the difference two Line TD Points by the number of days between them (excluding nontrading days). Further, by multiplying the additional numbe r of trading days from the most recent TD Point Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.16 See how price gapped above A-B TD Supply Line and how by adding the difference between X—the value of the TD Line immediately above Y—and Y—the lowest price beneath the A-B Line—the price projection Z is determined. to that precise point at which the TD Line is pene trated by the rate of change, the exact breakout value can be calculated (see Table 1.1). To that breakout value, the difference between the TD Line and the trough/peak immediately below/above—depending on whether it is a buy or a sell-can be added or sub tracted to arrive at a price objective. Once again, what may appear complicated becomes much clearer when displayed on the charts (see Figures 1.18 and 1.19). TD Price Projector 2 is somewhat more complex. For example, in the case in which price exceeds a de clining TD Line , instead of selecting the lowes t price 26 TD Price Projectors Trendlines Table 1.1 27 Rate of Change Calculation ofTD Line 1. Count the number o f days from the mo st recent TD Point to the second most recent TD Point. 2. Calculate the price differ ence between the two TD Points. 3. Divide the difference betwe en the two TD Points by the number of trading days separating the two TD Points to arrive at the daily rate of advance (rate of decline). 4. Multiply the rate of advance (rate of decline) by the number of trading days to the breakout price in order to calculate the exact breakout price for purposes of arriving at a precise price objective. 5. Identify both the highest high above an ascending TD Line (the lowest low below a descending TD Line) and the value of the TD Line immediately below the highest high (above the lowest low). Next, calculate the difference between the that highest high (lowest low) the TDprice. Line, and subtract value from (add to) theand breakout beneath the TD Line and adding that value to the breakout point, there is a slight variation. The intraday low price on the d ay of the lowest closing price be neath the TD Line is selected. Often, the lowest intrad ay low is recor ded on the same day as the lowest close, so there exists no difference between TD Price 1.22). As you can readily see, using the intraday low as the referen ce point, regard less of the close that day relative to all other closes beneath the TD line, as TD Price Projector 1 requires, is considerably more liberal and simpler t o calculate. On the o ther hand , TD Price Projector 2, using the intraday low of the lowest close day, is a little more difficult in arriving at a price ob jecti ve once a decl ining TD Line is excee ded ups ide . Conversely, to arrive at downside price projections when an ascending TD Line is br oken, th e reverse pro cedure is employed (see Figures 1.23 and 1.24). In this case, the key day to concentrate on is the highest close day or, more precisely, the intraday high that particu lar day. Although it may appear that TD Price Projec tor 2 is more precise and conservative than Projector 1, this is not always the case. For example, if the rate of advance or decli ne is particu larly steep a nd the low est close in the case of a downtrend or the highest close in the case of an uptrend—the reference day for Projectors 1 and 2; but, in those instances when the lowest close day and the lowest low day are not coinci dent, the adjustment is made. Some examples wil l illustrate the distinc tions be tween the two methods (see Figures 1.20 through Projector the intraday or intra day high,2—occurs then the before price objective for low Projector 2 is greater. Conversely, if the low or the high close day be neath or above the trendline occurs subsequent to recording the intraday low or high, then Projector 2 is Source: Logical Information Machin es. Inc. (LIM). Chicago, IL. Figure 1.17 In an uptrending market, TD Pr ice Projector 1 conce ntrates on the extreme price high above the TD Demand Line. In this instance, price opened beneath A-B TD Demand Line and proceeded to satisfy price objec tive Z, which is calculated by subtracting the difference between the highest price above A-B Demand Line (Y) and the A-B value (X) on that same day from the price breakout. 28 Trendlines TD Price Projectors 29 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.18 In a declining market, TD P rice Projector 1 concentr ates on the extreme low price ben eath the TD Supply Line . In order to arrive at a price objective Z, calculate the difference between X—the value on the A-B TD Supply Line—and Y—the lowest price recorded beneath A-B Supply Line prior to the price breakout—and add that value to the price breakout to arrive at the price objective. In strong markets, secondary price projections can also be made by multiplying the difference by 2. Figure 1.19 Once the A-B TD Demand Line is constructed, TD Price Pro jec tor 1 requ ire s th at the price objective be c alcu late d b y s ubtr act ing value X-—the price on the TD Demand Line on the day the peak price (Y) above the Demand Line is recorded—from Y and then subtracting the resulting value from the breakout to arrive at 2 price objective. 30 Trendlines Source: Logical Information Machines, Source: Logical Information Machines, Inc. (LIM), Chicago. IL. Figure 1.20 Mote, in this example, TD Price Projector 2 is the same as TD Price Projector 1 because the lowest close day (Y) recorded below TD Supply Line A-B is on the same day as the lowest low (Y). Inc. (LIM), Chicago, IL. Figure 1.21 TD Price Projector 2 is not the sam e as TD Price Projector 1 in this example because the lowest close (Y) below TD Supply Line A-B is not on the sa me day as th e lowest low (Y'). Even though the breakout point is th e same value, the difference between Y and the value immediately above it on Supply Line A-B is different from the difference between Y' and the value im mediately above it on Supply Line A-B. Consequently, the price objectives are not the same. 32 Trendlines TD Price Projectors 33 Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL Figure 1.22 The price objectives (Z and Z'), arrived at by subtrac ting the lowest low (Y') beneath Supply Line A-B and the low on the lowest close (Y) beneath Supply Line A-B from the X and X' values on the A-B line, are not the same in this c ase be caus e the lowest close and the lowest l ow do not occur on the same day. Thus the distinction between TD Price Projectors 1 and 2. Figu re 1.23 TD Demand Line A-B defines two different downside price ob jec tiv es (Z' a nd Z) bec ause the h ighest high day a bove th e De mand Line (Y1) is not the sam e day as the highest close day (Y). This differentiates TD Price Projectors 1 and 2. . 34 Trendlines TD Price Projectors 35 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.24 In both instan ces on the chart, lines A -B define the TD De mand Line, but the highest close above the TD Line(Y) and the highest intraday high above the TD Line (Y') do not occur on the s ame day. Consequently, the price objectives for TD Pri ce Projectors 1 and 2 are not the same . Source: Logical Information Machines . Inc. (LIM) , Chicago, IL. Figure 1.25 TD Price Projector 3 is conservative, and its price ob jective is generally realized prior to fulfillment of the price objectives for TD Price Pro ject ors 1 and 2. This exam ple dem onst rat es the select ion of po ints for TD Price Projector 3. Note the close on the lowest day (Y) is near to the Supply Line A-B and, consequently, a lower target is gene rated once price brea ks out above the TD Supply Line. less. Personally, I prefer Projector 1 to Projector 2, but I believ e th at Projector 2 is valid and is a viable option. TD Price Projector 3 is more conservative gener ally than the prior two methods. To calculate the price projection value in the in stanc e of a declining TD Line, merely calculate the difference between the TD Line an d the CLOSE of the lowest intr ada y low in projecting prices because it usually yields the smallest returns. In order to reach the price targets established by TD Price Projector 1, the price objec tive generated by TD Price Projector 3 must have also been hit. Generally, this same observation applies to day immediately beneat h it, NOT the intrada y lo w it self {see Figures 1.25 th rou gh 1.27 ). The major dis tinction I wish to make relates to the intraday low and t he close on the day of th e intr ada y low. By defi nition, this approach is likely the most accurate TD Price Projector 2, that is, the TD Price Projector 3 price target is realized first. Conversely, to project downside tar gets given a penetratio n of an asce nding TD Line, the opposite procedure is conducted. The difference betwee n the closi ng price of the hi ghes t 36 Trendlines TD Price Projectors 37 Source: Logical Information Machines, I nc. (LIM), Chicago, IL. Source: Logical Inform ation Machines, Inc. (LIM), Chicago, Figure 1.26 Once again, the upside price objective is muted because of the high close on the lowest low day relative to the Supply Line. Figure 1.27 This chart highlights numerous instances in wh ich TD Price Projector 3 yields lower price targets than TD Price Projectors 1 and 2 be cause the close of the intraday low beneath the TD Supply Line (A-B) is used. IL. intraday high day above the TD Line and the TD Line value immediately beneath it is calculated (see Figures 1.28 and 1.29). Once again, I emphasize the intraday high and t he close on the day of the i ntra day high. Of the three approaches presented to make price projections, TD Price Projector 3 is the most precise and the most conservative. Through experimentation, you should be able to select the approach with which you are most comfortable. I highly recommend that, target's being realized. Specifically, when breaking an upsloping TD Line, subtract one tick from the high or the close, depending on the method used, and add one tick to the TD Line. Conversely, when breaking a downsloping TD Line, subtract one tick from the TD Line, an d add one tick to the low or the close, depend ing on the method used. regar dless of whi ch one you might select, you shave one price tick from the high, low, and TD Line when calculating the price projection, to compensate for roun din g off an d to ensu re the likelihood of the price To fully comprehend the differences among the three TD Price Projectors, you should study Table 1.2, which summarizes the various similarities and differences.  38 What Could Go Wrong? Trendlines 39 Source: Logical Informati on Machines, Inc. {LIM ), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.28 Unfortunately, by using TD Price Pr ojector 3, conservative price projections are made. As in this ex ample, should the close on the lowest intraday low day beneath the TD Supply Line be positioned in the area of the high for the day, the price objective is less than if the close were positioned closer to the low for that day. What Could Go Wrong? No technique is perfect. Forecasting price movements, after all, cannot be that easy. What unexpect ed situa tions could arise? Three potential developments might occur, all of whi ch can be deal t wi th simply: 1. A cont radi ctory signa l could b e generate d by a penetration of an opposing TD Line. This would effectively nullify the active TD Line Figure 1.29 TD Price Projec tor 3 requires that the diff erence between the close of the highest intraday high day above the A-B Line (Y) and the value of the A-B point (X) on that same day be subtracted from the downside break out to arrive at the price objective. Breakout and instate in its place, as the pre vailing trend, the new TD Line breakout in the opposite direction. This is the most common way for a price trend to be terminated and for a price objective to be negate d (see Figure 1.30). 2. The initial indica tion of a valid TD Line bre ak out may be just plain false or may be reversed by an unexpected news event that dramati cally shifts the supply-demand balance. This becomes immediately apparent when the open 40 What Could Go Wrong? Trendlines 41 Table 1.2 TD Price Projector s Price Projector 1: Buy Signal—Calculate the difference between the lowest price low below the descending TD Line and the TD Line value immediately above it, and add that value to the breakout price. Sell Signal—Calculate the difference between the highest price high above the ascending TD Line and the TD Line value immediately below it, and subtract that value from the breakout price. Price Projector 2: Buy Signal—Calculate the difference between the intraday price low on the day in which the lowest close below the descending TD Line is recorded and the TD Line value immediately above it, and add that value to the break price. out Sell Signal—Calculate the difference between the intraday price high on the day in which the highest close above the ascending TD Line is recorded and t he TD Line value immediately below it, and subtract that value from the breakout price. Price Projector 3: Buy Signal—Calculate the difference between the close of the lowest intraday low below the descending TD Line and the TD Line value immediately above it, and add that value to the breakout price. Sell Signal—Calculate the difference between the close of the highest intraday high above the ascending TD Line ahd the TD Line value immediately below it, and subtract that value from the breakout price. for the next day is recorded: and, on the open ing, it either exceeds downside the value of the active descending TD Line, which was previ ously penetrated, and price continues to de cline; or it gaps downside on the opening and, on the close, breaks below the descending TD Line. Conversely, the breakout is suspect when, on the next day, either the open or the close is recorded and either exceeds upside wit h a gap the value of the as cen ding TD Line Source: Logical Informati on Machines. Inc. (LIM), Chicago, IL. Figur e 1.30 Note that the objective of Price Projector 1 had not been ful filled for the breakout below TD Line A-B by the time an upside breakout of TD Supply Line C-D took place. Consequently, the downside price objective based on the penetration of the Demand Line A-B is no longer active. it previously penetrated, and price continues to advance (see Figures 1.31 and 1.32). To re duce the financial ri sk associated with just such unforeseen events, a stop loss can be in stalled once price opens on the ensuing day. 3. The fulfi llment of a price objective defi ned by the breakout above or below a TD Line could disqualify an active trend. Such occurrences were discussed thoroughly earlier. ) 42 TD Trendlines Cocoa Source: Logical Information Machine s. Inc. (LIM), Chicago, IL. Figu re 1.31 Althoug h the TD Supply Line A-B was exceeded, on the day following the breakout the opening price was below the breakout day's close and it proceeded to immediately decline below the extended A-B Line. This price activity invalidates the breakout. TD Lines of a Higher Magnitude The TD Lines described an d disc ussed above are of a level 1 magnitude, that is, each TD Point used to con struct them required no more than three days to be defined—a high immediately preceded and succeeded by a lower high (or a low immediately preceded and succeeded by a higher low). The TD Line created by connecting two TD Points is of short duration, due to the fact that construction of as few as five highs (in Fin. Times Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figu re 1.32 See how price failed to decline on the day following the pene tration of the TD Demand Line A-B. In fact, price opened the following day unchanged and continued to advance from that price level above the ex tended A-B Line, thus nullifying the breakout. the c ase of a TD Supply Line) or five lows (in the c ase of a TD Demand Line) could require two pivot points and immediately surrounding days. A trader may wish a longer-term perspective. To satisfy this requirement, I experimented with higher-level TD Points and Lines, and realized worthwhile results. To draw a TD Line of level 2 magnitude, a mini mum of five days is required to identify each TD Point—a high immediately surrounded on both sides by two lower highs (or a low immediately surrounded 44 Trend lines TD Lines of a Higher Magn itud e 45 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Source: Logical Information Machines , Inc. (LIM), Chicago, IL. Figure 1.33 As you can see, level 3 magnitude TD Points are ide ntified with a circle surrounding the high (low), [By definition, these are also level 1 magnitude TD Points as well. Those highs and lows marked with an "X" are level 1 points but do not qualify as level 3 because the highs (lows) all three days immediately before and immediately after are not lower highs (higher lows).] Figure 1.34 The distinc tion between level 1 magn itud e TD Points ("X") and level 2 magnitude TD Points (circles) are apparent on this chart. Whereas level 1 requires merely a lower high (higher low) the days immediately before and after, level 2 requires two lower highs (higher lows) the two days immedi ately before and after. On both sid es by two higher lows). Similarly, a TD Line of level 3 magnitude would require a total of at least seven days for each, and so on for TD Lines of higher levels. It is correct to say tha t all TD Points of a magni tude greater than level 1 are also level 1 TD Points; but all do not qualify as active level 1 TD Line Points be cause, as discussed earlier, only the two most recent points are considered valid. To visualize this distinc tion, refer to Figures 1.33 and 1.34. Regardless which TD Line level of magnitude is selected, the same requirements exist as are de scribed above for both TD level 1 Points and Lines. The only exception is the number of days required to define the TD Points. Similarly, the identical TD Price Projectors are used. It has generally been my preference, however, to follow level 1 magnitude TD Lines. 46 TD Lines of a Higher Magn itude Trendlines 47 Source: Logical Informa tion Machines, Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.35 By awaiting the formation of a TD Point of a higher level of magnit ude, lower-level TD Lines that ar e valid are forfeited. This chart illus trates two examples of valid breakouts that would have been identified had the points selected been level 1 or level 2 and not level 3 (see line A-B), and one example that would have been valid had level 1 been selected and not level 2 (see A'-B'). I have two primary reasons for concentrating on the basic TD Line: 1. A s the level of mag nit ude incr ease s, often the breakout occurs before the most recent TD Point is completely formed, and the trading op por tun it y is fo rfeite d; consequent ly, the exer cise becomes a game of "beat the clock" {see Figure 1.35). Figure 1.36 TD Demand Line A-B (level 2) is active u ntil price excee ds TD Supply Line A-B (level 2) upside. (Note how the breakout above the first TD Supply Line a-b correctly predicted the subsequent price movement and price objective.) 2. As the level of magn itud e increa ses, th e likeli hood of realizing the price objective before a contradictory signal occurs is reduced propor tionately. Occasionally, I will examine a higher-level TD Line to determine the trend defined by the TD Line and thereby to confirm that the TD Line I am using is consistent; in other words, I will refer to it to confirm my marke t outlook (see Figure 1.36). 48 Trendlines Revolutionary Breakthro ugh 49 Revolutionary Breakthrough: Validation of Intraday Price Breakouts After the TD Points have been properly selected, the TD Line has been correctly drawn from right to left, the price objective has been calculated, and the three possible outcomes—(1) reversal signal, (2) dramatic shift in supply-demand equation, and (3) price ful fillment—have been addressed, there is one addi tional factor to consider: validation of intraday price breakouts. This element is significant. It is a major contribution to the stud y of marke t timing analysis. Furthermore, it has application to other techniques as well. It's not surprising to hear that traders have taken positions on presumed trendline breakouts only to witness price fail and reverse and to incur significant losses. What is hard to understand, however, is that these same traders will continue to repeat this futile exercise and never question what causes it to occur. The incidence of false breakouts has always been high. They have been the nemesis of traders for years and have often been the excuse for totally abandoning the use of tre ndlines. The creation of TD Lines all evi ates this problem somewhat, but invalid breakouts do occur occasionally. Heretofore, no method has been devised to differentiate between valid and invalid price breakouts. Many years ago, I was involved in similar situa tions, became frustrated, a nd was determined to de velop rules that would qualify TD Line breakouts. I was convinced that the TD Lines drawn were valid. I searched for a common denominator associated with both the good and the bad signals. It was not an easy task. The conclusions I made were startling and, at the same time, logical and simple. These were my findings. I discovered three TD Breakout Qualifiers—two patterns occurring the day before a suspected break out and the other pattern occurring the day of the breakout. Specifically, I concluded that if a particular Source: Logical Informat ion Machines, I nc. (LIM), Chicago, IL. Figure 1.37 Note that the closing price on the day immediately before the breakout to the upside was a down close versus the previous day's close. This pattern suggests an oversold condition prior to the breakout, which is a posi tive formation. market or index is oversold/overbought the day before a breakout, the chances are increased that the buying pressure/selling pressure would not be dissipated subsequent to the breakout, thus merely creating the illusion of continued strength/weakness. I experimented with numerou s conditions prece dent to an a breakout and found if the the day before upside breakout is that down, the close likelihood is increased that the intraday breakout will be valid and intraday entry is warranted—TD Breakout Qualifier 1 (see Figure 1.37). Furth er, if the close of 50 Trendlines Revolutionar y Breakthro ugh Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Source: Logi cal Information Machines , Inc. (LIM), Chicago, IL. Figure 1.38 Observe that the close on the day prior to the upside breakout was an up close, thus indicating an overbought condition and the likelihood of a breakout failure. the day prior to the upside breakout is up, the possi bility of a false move exists {see Figure 1.38). Con versely, if th e day before a downs ide bre ako ut is up , the likelihood is increased that the intraday breakout is valid and intraday entry is warranted (see Figure 1.3 9). Furthe r, i f the close the day prior to the down side brea kou t is down, th e likelihood of a false move exists (see Figure 1.40). An exception to the requirement that the close prior to an upside breakout be down and the close prior Figure 1.39 Note that the close on the day prior to the downside breakout was up, thus indicating a short-term overbought state and the likelihood of a valid breakout downside. to a downside breakout be up was uncovered when an analysis of successful breakouts showed that not only would an oversold/overbought close qualify entry, but so would an open above a declin ing TD Line or an open below an ascending TD Line—TD Breakout Qualifier 2 (see Figures 1.41 and 1.42). This occurrence would suggest excessive strength/weakness and would jus tify entry at tha t opening pri ce re gardless of the previ ous day's close disqualifying the entry. 51 52 Trendlines Revolutionary Breakthrough 53 Source: Logical Informati on Machi nes, Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.40 Notice that the close p rior to the downside penetration of the A-B TD Demand Line was down, thus suggesting a false breakdown. Figure 1.41 Observe that the opening price exceeded the TD Supply Line, thus validating a breakout. TD Qualifier 3 is similar to TD Qualifier 1 to the extent tha t it is based on price activit y durin g the day prior to a breakout. In this case, however, the differ ence between the high and the close on the day prior to a trendline downside penetration is subtracted from that day's close to arrive at a supply value. The differ ence between the close and the low on the day prior to a trendline upside penetration is added to that day's close to arrive at a dem and value (see Figure s 1.43 and 1.44 ). If the a scen ding tren dline is below the supply action is warrante d. Examples of TD Brea kout Quali fier 3 are pre sente d in Figures 1.45 and 1.46. Table 1.3 further describes the Qualifiers. As presented in this chapter, TD Points are objec tively defined and, when properly connected, they cre ate TD Lines. When TD Breakout Qualifiers are introduced and the TD Lines are penetrated, legiti mate price breakouts are identified and price targets can be derived. It has never been so simple. Guess value and price exceeds the trendline, the price de cline should accelerate and intraday action is war ran ted . Conversely, if the descend ing trendl ine is above the demand value and price exceeds the trendline, the price advance should accelerate and int raday work and lack of consistency are eliminated totally. Uniformity in construction, in application, and in in terpretation has been accomplished. The successful identification of trends and of trend reversal points is complete. 54 Trendlines Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Figure 1.42 See how price gapped on the open below the TD Demand Lin e, confirming a breakout. Revolutionary Breakthr ough 55 Source: Logical Informati on Machines, Inc. (LIM ), Chicago, IL. Figure 1.43 By calculating the diff erence between the close on the day prior to an upside breakout and that same day's low (or the previous day's close, whichever is less) and adding that difference to the close prior to the breakout, validation of the breakout is determined. If the difference added to the close is less than the breakout price, a valid breakout is identified. If the difference is greater than the breakout price, a false breakout is likely to oc cur. Specifically, in this example, the difference between the close and the low on the day prior to the breakout above TD Supply Line A-B is calculated and it is less, thus qualifying the breakout. TD Demand Line A'-B' is also drawn, and the same concept in reverse qualifies the downside breakout (see Figure 1.45). 56 Trendlines Source; Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.44 The difference between the close and the l ow on the day prior to the upside breakout of A-B Supp ly Line added to the close on the day prior to the breakout is less than the breakout price. Consequently, a valid break out has been confirmed. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 1.45 By subtracting the dif ference between the hi gh one day before a downside breakout (or the close two days before it, if it is greater than the high one day before the breakout) and the close the day before the breakout from the close that same day, the breakout can be validated. In this instance, the A-B Demand Line was less and the breakout was validated. 58 Trendlines Retracements Source: Logical Informa tion Machines, Inc . (LIM), Chicago, IL. Figu re 1.46 The difference betw een the close and the low the day before the breakout added to the breakout is less than the breakout price, thus vali dating the breakout. Table 1.3 TD Breakout Qualifi ers TD Breakout Qualifier 1: To validate a buy signal, the close the day before a buy signal is a down close. To validate a sell signal, the close the day before a sell signal is an up close. TD Breakout Qualifier 2: To validate a buy signal, a price open greater than the breakout price must occur. To validate a sell signal, a price open less than the breakout price must occur. TD Breakout-Qualifier3: To validate a buy signal, the value arrived at by adding the difference between the close and the low the day prior to a breakout or the close 2 days before whichever is less to that same day's close must be less than the breakout price. To validate a sell signal, the value arrived at by subtracting the difference between the high or the close 2 days before whichever is greater and the close the day prior to a breakout from that same day's close must be greater than the breakout price. Price advances and decli nes do not continue uninte rrupt ed. C ontratrend rallies and trend reversals occur all the time. Predicting the extent and the duration of these mov es is a preoccupation of marke t analysts. Techniques to project these retracement levels, however, are no t only inexact but also often haphazar dly applied. Generally, little preparation and forethought are given to the selection of the points critical to the calculation of both support and resistance lev els. Through trial and error, I have identified proper price selection techniques, as well as ratios, that can be universally applied to all markets. I have replaced the practice of random selection and guesswork with an objective, mechanical approach that offers a se ries of explanations and includes exampl es to justify my techniques. Selection of Points and Retracement Ratios In the summ er of 1973, one of m y senior business associates gave a compelling speech forecasting a steep retracement rally for the stock market to be completed by early fall. When quizzed regarding the upside potential, he flatly stated that he expected an advance Retracements Selection of Points and Retrace ment Ratios 61 approximating th ree-eighths to five-eight hs of the pre vious decline. When asked to be more precise, he was unable to do so. When asked how he arrived at these figures, he flippantly responded that most rallies in a bear market expire at one of these two levels. He could not supply any additional information other than the fact that he had read of similar ratios in a market let ter and suggested I contact the writer to learn of the significance and srcin of these numbers. When I called, the writer made reference to a market analyst by the na me of R. N. Elliott and an article he had writ ten in Financial World many years earlier. At the li brary, I was able to retrieve the article from the archives. What I read regarding Fibonacci numbers was fascinating and informative. Unfortunately, it 1970s I was the first to apply .382 (1.00 - .618) retracements as well. The retracements increase from .382 and .618 to unit y or "magnet price" (not 1. 00, as most bel ieve, b ut the high/low d ay's close, depend ing on whether the correction is up or down). From the full retraceme nt to the magnet price, they continue 1.382, 1.618, and 2.236 (1.618 + .618), and in the case of a marke t recording all -time, record highs, the actua l ra tios .618 and .382 times the absolute price high. The series of retracement levels are described below (see Table 2.1). More important than the ratios themselves is the selection of the price points required to calculate the retracements. For most analysts, such a process is a moving target. Once again, in order to maintain con proved to be woe of fully incomplete. an an e xhaus tive examination price activity toI beg uncover a precise, objective approach to calculating retracements. My goal was to identify a method that was mechanical and ha d application to all mark ets. The conclusions of my researc h are presented in the foll owing discussion. Rather than devote a lengthy amount of time and space to a discussion of Fibonacci numbers and their presence and dominance in nature and in surround ings, I recommend, should you desir e further informa tion, that you consult the numerous recent textbooks and articles dealing with the subject matter and the "golden mean." Briefly, I highlight the significance of the Fibonacci time series starting with the numbers 1, 2, 3, 5, 8, 13, 21, 34, 55, . . . and so on. Essen tially, these numbers are obtained by summing two consecutive numbers in the series to arrive at the ne xt number. As the numbers increase in size, the ratio ob tained by dividing a preceding number by a succeed ing number approaches .618, and, conversely, the ratio obtained by dividing a succeeding number by a preceding number is 1.618. This is a characteristic peculiar only to this specific time series. sistency and uniformity, I refined and defined these price levels precisely and objectively. My research sug gested that th e best results were obtained by applying the following procedures. Assume a recent lo w has been recorded. To e stab lish reference points for retracement price objectives, extend an imaginary horizontal line from that recent low toward the left side of the chart to the last time a lower low had been recorded (see Figure 2.1). Next, refer to the highest price between these two points; that is the "critical price." By subtracting the recent low from that value, price retracement levels can be projected (see Figure 2.1). Conversely, to arrive at I apply these numbers and ratios to much of my work. Specifically, in the case of retracements, I use what has become the standard, the "mother of all retracement ratios": .618. I believe that in the early Table 2.1 Retracement Factors .382 .618 Magnet price (high/low clay's close) 1.382 1.618 2.236 2.618 3.618 62 Retracements Selection of Source: Logical Informati on Machi nes, Inc. (LIM), Chicago, IL. Figu re 2.1 Once price level B has been defined, refer, on the left side of the chart, to the last time a lower low than that registered on day B was recorded (see price C). The highest high between these two points (A) is the critical price and is an important reference level when calculating upside re tracements. By calculating the price difference between points A and B and multiplying by retracement factors, upside projections can be made. Notice how price hit an important obstacle at point A's close and not at the intraday high (refer to the discussion regarding magnet price and to Figure 2.5). retr acem ent levels o ff a price peak, the sa me exercis e is performed in reverse. Specifically, extend an imagi nary line from the recent price high toward the left side of the c ha rt to the last time a higher high wa s made (see Figure 2.2). Identify the lowest low between these two points; that is the critical price. By sub tracting the critical price from the recent high, price retracement levels can be projected (see Figure 2.2). Not only is this approach simple but it ensures that the Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 2.2 To calculate the down side retracement lev els, extend an imagi nary line toward the left sid e of the chart to the la st time a price high exceeded the high of point B (see price C). The lowest low between these two points (A) is the critical price and a key level when calculating downside retracements. By calculati ng the price differenc e between points A and B and multiplying by retracement factors, downside price projections can be made. calculation of retracements will always be clear, welldefined, and uniform. Experience has proven that the initial retracement levels of .382 and .618 are valid and predictable when the critical price point is properly identified (see Figures 2.3 and 2.4). However, most analysts who had even accidentally selected the correct points were mistaken when they assumed that, once price had exceeded both the .382 and .618 levels, the next 64 Selection of P Retracements Source: Logical Information Machines. Inc. (LIM). Chicago, 1L . Figure 2. 3 By identifying the lowest low between poi nts B and C—the recent high (B) and the last time a higher high was recorded (C)—downside retracement levels can be calculated. Marked on the chart are levels 1 and 2, which identify .382 and .618 retracements of the move from point A to point B. Source: Logical Informa tion Machines, Inc. (LIM), Chicago, IL. Figure 2.4 By identifying the highest high between points B and C—the re cent low (B) and the last time a lower low was recorded (C)—upside retracement levels can be calculated. Marked on the chart are levels 1 and 2, which correspond with .382 and .618 retracement levels of the move from point A to point B. objective was the critical price day's high in the case of a rally and the c ritical price day 's low in the ca se of a decline. In actuality, this widespread expectation has served for years as merely a decoy for unsuspect ing traders (see Figure 2.5). My resea rch with retracement s many year s ago uncovered a more cruci al level, which served as a resistance/support level. Conse accelerate through a critical price and immediately proceed to the 1.382 level, just at the time when everyone expects support at the intraday low or re sistance at the intraday high. Unexpected movements occur at th is price leve l when one is n ot properly pre pared. quently, I refer price penetration as the "magnet price." How often have to youthis awaited of a critical price high/low to liquidate or to enter a trade , only to witness price reverse direction early without a fill? In other instances, it is not uncommon to see price There are exceptions to the retracement calcula tions presented above. Experience suggests that the previous examples are associated with trading range markets. Once price advances to record all-time highs, however, the selection of a "critical point" (low) 66 Retracements Selection of Points and Retracement Ratios 67 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Source: Logi cal Information Machines. Inc. (LIM), Chicago. IL. Figure 2. 5 Observe how price retraced to levels 1 and 2 and eventually de clined to the low day's range exceeding that day's close. How many traders were expecting a test of the low price at point A and were fooled? The key is the "magnet price"—low day's close. Figure 2.6(a) If one were to extend an imaginary line fr om point B to the last time price recorded a higher high to iden tify a reference low in between, he would be unable, because price had never been higher. In these rare in stances, valid downside price projections can be calculated by merely multi plying the absolute high b y .618 and by .382. In this example, price met good support at level 1, which approximates .618 times the high. is impossible, because no other high exists farther left on the chart. In these instances, I have found multi plying the absolute peak price by .618 and by .382 works well. Two instances are indelibly etched upon my memory. The first is a forecast I made to an audi ence in New York prior to the 1987 stock market the market of approximately 1697 (see Figure 2.6a). I presented that level to the audience as my downside price projection. At the same time, I presented a TD Line breakout objective of 1650 {see the trendline discussion in Chapter 1). Although mechanically a nd objectively derived in both instances, it was surpris crash. At that time, the August price high for the Dow Jones Industrial Average was approximately 2747—a record. By multiplying that value by .618, I was able to calculate a support (retracement) level for ing how both reinforced one another and lent credibil ity and conviction to my forecast. Another episode involving the application of this retracement approach relates to a forecast I made 68 Retracements TD Retracement Arcs Nikkei weekly 69 TD Retracement Arcs Another method I developed many years ago projects retracement levels incorporating both price and time. I had never seen anything quite like it prior to my re search and have never seen anything similar since. Instead of identifying one retracement price objective as most other techniques do, this approach has a Source: Logical Informati on Machin es, Inc. (LIM), Chicago, IL. Figure 2.6(b ) objective. Using the DeMark technique to predict the dow nside price subsequent to the Nikkei Average recording its alltime high at 38957. I had been sponsored by a large brokerage house and was invited to make a series of speeches in Japan regarding various Japanese markets. Invariably, the topic would be directed to the Japanese stock market. Due to the fact that a record high had been realized and that no "critical price" (low) could be identified, I applied my approach to the Nikkei Average. As you can see on Figure 2.6b, Figure 2.7 This unique approach incorporates both price and time. The ref two erence points are identified the same as before, but once identified the the technique accurately predicted the downside price objective below 15,000. At the time the forecast was made, skepticism and ridicule were common. Once price had accomplished this downside objective, however, respect for the technique was pervasive. points are connected with aprice) straight-edge ruler on andthis theline. retracement (.382, .618, and the magnet are identified Then, by levels using point B as a fulcrum, an arc is extended to the right of the chart; once a price close exceeds the arc in the future, it proceeds to the next TD Retraceme nt Arc level. This assumes that the points are properly anchored to avoid any changes in the shape in the arc when either the price or the calendar scale is cha nged. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. 70 floating price objective that adjusts to the passage of time. More specifically, the price objective changes from day to day. The specific price resides on an arc. The arc is constructed by drawing a line that connects the critical price and the recent low/high, depending on whether an advance or a decline is ex pected. Once the .382 and the . 618 points on the line itself are identifi ed, th e rec ent low/h igh serve s as a fulcrum or a pivot point for an arc that extends from those points into the future. The curve that appears describes retracement projections (see Figures 2.7 and 2.8). As price declines below/above the pivot point, a new arc must be drawn. Retracement Qualifiers Important elements of my TD Lines are the TD Break out Qualifiers (see the discussion of TD Line in Chapter 1). They are extraordinary filters that predetermine whether an intraday breakout is valid or invalid . So too, moves exceeding .382 and .618 retracements can be prescreened as legitimate or not. My research sug gests that these same TD Breakout Qualifiers apply Microsoft Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 2 .9 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 2.8 Note the numerous TD Retracemen t Arcs formed by lines ex tending from points B-A, C-B, and D-E. The arc formed by B-A is reversed by the arc defined by C-B, and the price breakout upside above the latter arc is confirmed by a similar breakout above the larger TD Retracement Arc D-E. The close on the day be fore point B is a down close {circled), which satisfiesthe TDprospects Retracement Qualifierentry 1, indicates an oversold condition, and improves for intraday and a successful trade to the retracement levels. In this instance, the .382 and the .618 retracement levels were fulfilled the same day, In fact, the move the following day extended to the mag net price (D)—point A close—and did no t, as most trade rs would ex pect, reach point A day's high. 72 Retracements Retracement Qualifiers equally well to validating retracements. To remain consistent with the topic, however, I refer to these qualifiers as TD Retracement Qualifiers 1 2 and 3. It might be worthwhile to reiterate their respective defi nitions. Specifically, TD Retracement Qualifier 1 re quires that the close on the day prior to an advance above the retracement level be down in order to qualify for intraday entry and the expectation that the ad vance will conti nue (see Figure 2.9). Conversely, it pro vides that the close on the day before a decline below the retracement level be up in order to qualify for in traday entry and the expectation that the decline will continue (see Figure 2.10). Should this qualifier in , Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 2.1 0 The close the day before the .618 retrace ment level was an up close (circled) and implied a valid retracement, as well as justified intraday entry. TD Retracement Qualifier 1 has been satisfied. 73 either situation be lacking, the potential of the ad vance or the decline reversing, rather than accelerat ing, at those precise points increases significantly. Another possibility exists, and that is TD Re tracement Qualifier 2. In this instance, both an upside and a downside penetration can qualify, provided price ex ceeds the retracement level on the opening (see Fig ures 2.11 and 2.12). This occurrence suggests that the price dynamics are so strong that entry is vali dated and that the continuation of the move is immi nent. One additional qualifier (TD Qualifier 3) can be incorporated into one's trading arsenal. The concept Source: Logical Inform ation Machines , Inc. (L1M), Chicago, IL. Figure 2.11 See the opening gap downside (X) below the .382 retracement level. This fulfills the requirement of TD Retracement Qualifier 2 and negates the fact that the previous day's close was a down close (see circled close). 74 Retracements Source.- Logical Infor mation Machine s, Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 2.1 2 Price gapped above the .618 retracem ent level and fulfilled Retracement Qualifier 2 in the process. Figure 2. 13 By calculating the difference between the high the day before penetra tion of a retracem ent level (or the close the previous day, whichever is greater) and the close the day before the breakout, and subtracting that value from that close a nd then comparing it with the retracement level, it can be determ ined whether price wil l continue to th e next lower retrace ment level or reverse. If the retraceme nt price is below the value, then price shoul d con tinue to decline; if the retracement price is above that value, then price should attemp t to rally. The differe nce between the high and the close the day before X retracement level is above the .382 retracement point; conse quently, price continued to decline. TD Qualifier 3 has been fulfilled. of overbought/oversold is employed in this instance ju st as wi th TD Qua lifi er 1, bu t t he ap pr oa ch is dif ferent. Specifically, the expression of supply or de mand is calculated on the day prior to the breakout. The difference between the previous day's high and that day's close is subtracted from that day's close to arrive at a supply value; conver sely, the differ ence be tween the close and the low on the day prior to the breakout is added to that day's close to determine a demand value (see Figures 2.13 and 2.14). Just as when the TD Line on a particular day is below the supply value or above the demand value and a legiti mate breakout is recorded by price exceeding the TD Line, so too th e same approach applies to retraceme nt points. In fact, in those exceptional instances when both the TD Line and the re traceme nt levels are quali fied, they appear to reinforce one another and trans late into successful trades. As display ed in Table 2.1, there exists a series of retracement levels; when one is exceeded, price gravi tates typically toward the next level. Not included in Table 2.1 nor in the earlier discussion, however, is a 76 Retracements Source: Logical Information Machines . Inc. (L1M), Chicago, IL . Figure 2.14 Price failed at the retracement level because the difference be tween the previous day's close and low added to that same day's close was greater than the retracement level. hybrid situation that refines retracement expecta tions even further. Despite th e fact that price may ex ceed a retracement level and that a retracement qualifier may have been satisfied, a situation may arise in which the closing pri ce is unabl e to remain in excess of the retracement value. Provided that the closing price does not fail to exceed the previous day's close, the only adjustment required is a minor one. In stead of the nex t retraceme nt level being the one listed in Tabl e 2.1 , it is actually one-half of the dista nce be tween the two levels. For example, if price exceeds the .382 retracement level, is properly qualified, and closes above the previous day's close but fails to close above the .382 retracement level, the expectation level is no longer a .618 retracement objective; rather it is Trend Factors™ 77 the 50 percen t level. Furthe rmore, should all prerequi sites be satisfied at the .618 level but the closing price fail, the revised retracement level would be one-half the distan ce from .618 to th e critical price or the mag net price, whichever is less. The same process occurs, in reverse, when price declines. A final variation of retracements incorporates both price and time, as does the TD Retracement Arc. Rather than identify on an arc the time and the price retracemen t level, this approach applies a meter to the amount of time allotted to accomplish penetration of the retracement level. For example, if the number of days from a peak to a low is counted and this value is multiplied by .382 , an expiration time in which to ful fill a retracement is defined. The same concept can be applied to larger retracements, such as .618. During the mid-1970s, in order to simplify the process of calculating pri ce retraceme nts, I purchase d a proportional divider. I ordered this device from Teledyne-Post and it had to be sent from overseas. Since that time, I have shared this tool with many other market analysts and subsequently they have purc hase d their own. I recommend t hat if you work with price c ha rts a nd if you quickly want to approxi mate retra cemen t levels, th en pu rcha se one fr om your local art or drafting supply store. The discussion above describes method s to objec  tively def ine the pa ramete rs for calculating an d quali fying retrace ments. The be auty of these approaches is that they are totally objective. No allowance for subjec tive interference or for personal bias is permitted. In the final analysis, no excuses can be attributed to faulty suppositions. Everything considered is hard and fast, with clearly defined rules for application and interpretation. Trend Factors™ Often, both traders and investors preoccupy them selves with mark et rhetoric. As far as I am concerned, however, one distinction they fail to recognize or to 78 Retracements acknowledge is the difference between the duration and the degree of price moves. Invariably, were I to ask for definitions to widely used market terms such as short, inte rmediate, and long term, I woul d wager tha t most, if not all, would respond in temporal terms. Al though the responses might vary somewhat, I would expect to hea r tha t short term relates to moves of less than one month's duration, intermediate term to moves longer than one month but shorter than six months, and long term to moves lasting more than six months. These labels may have been appropriate prior to the 1980s, bu t because of increased volatil ity in the financial markets, they have become outdated. Moves that in th e past consumed weeks or month s are being fulfilled in days or hours. Because of market il- Trend Factors™ 79 identify the correct reference value—whether high, low, or close—were critical to defining the price trend and the anticipated price movement. In order to define it, a reference low must first be qualified. Because the min imum Trend Factor ratio is .0556, a qualified low exists once price has declined to a value at most .9444 from a previous qualified refer ence high day's close (see Figure 2.15). Conversely, a qualified high exists once price has advanced the minimum Trend Factor 1.0556 from a previous quali fied reference low day's close (see Figure 2.16). It's t ha t US T Bonds liquidity, the speed of news dissemination, the herd instinct of fund managers, and other factors, this trend continues. Consequently, I apply the descrip tions of short, intermediate, and long term to percent age price movements rather than to specific time intervals. For example, I consider moves of less than 5 perc ent to be short term, moves of 5 to 15 per cent to be intermediate term, and moves greater than 15 percent to be long term. These definitions apply regardless of the time required to accomplish these moves. In this context, I rely on an analytical tool I developed many years ago to forecast the inception of trend moves of an intermediate to long-term variety. I refer to the set of specific ratios I use as Trend Factors™. In the early 1970s, I observed a dominant tend ency for various markets to exhibit support and resistance at price intervals defined by percentage re tracements from preceding price peaks and troughs. For example, upside resistance levels were projected by multiplying a recent low by a series of prescribed ratios. Conversely, downside support levels were cal culated by multiplying a recent price peak by the inverse upside ratios. a tedious and-err of or the process, I was ableThrough to approximate thetrialideal ratio values. I established prerequisites to qualify the reference highs and lows and to ensure consistency and uniformity. The selection criteria necessary to Source: Logical Information Machines, Inc. (LIM), Chicago, 1L. Figure 2.15 From point B to point A, price declined more than .0556—in other words, the value of point A is no more than .9444 times the value of point A. This validates point A as a Trend Factor low. Price X is the 1.0556 level and price Y is the 1.112 level. Trend Factors™ 80 81 levels. Figures 2.15 and 2.16 demonstrate this phe nomenon. As identified earlier, the essential ingredi ents are the qualification and the selection of the reference point and then the Trend Factor ratios. To determine th e first le vel of resistance upside, the refer ence point must be multiplied by 1.0056; to arrive at the second level of resistance upside, the reference point must be multiplied by 1.112 (2 X .0556); and to calculate the third level of resistance, the reference point must be multiplied by 1.14 (2 ,5 X .0556). On the other hand, to determine the first level of support downside, the reference point must be multiplied by .9444; to arrive at the second level of support down side, the first downside target must be multiplied by .9444 again (this is different from the upside process); Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 2. 16 Point A is no more than .9444 point B and is qualified as a ref erence point. Price levels X and Y correspond to 1.0556 and 1.112 Trend Factor objectives. Point C qualifies as a Trend Factor high reference point beca use price advanced from point A by more than 1.0556. X' and Y' identif y Trend Factor objectives, as does Z. simple to select price points that qualify. Before get ting too specific, I'd like to introduce you to the con cept of Trend Factors™. There appears to be a critical jun ctu re in price movement, at which point, once price momentum sur passes on a closing basis key support or resistance levels, it is able to proceed to th e next critica l ju nc tur e. The mark et itself legitimi zes those price moves. An alert trade r can capitalize on the message the price movement sends once it exceeds these critical price and to calculate the third level of support downside, the second level must be multiplied by .9722 (1 /2 of value between .9444 and 1.000). It is importa nt to select the proper reference price value, to ensure that the support and the resistance levels are defined accurately. My research has con firmed that specific patterns identify what price to use to calculate these threshold levels. Specifically, if (1) the reference day's close is less than the close one day earlier, and (2) the close one day before the refer ence day's close is less than the close two days earlier, and (3) the close one day after the reference day's low is greater than the reference day's close, then the up side resistance level and projection values are deter mined by multiplying the reference day's low by the ratios. If either the reference day's close or the close one day before is an up close or an unchanged close, then the reference price used to calculate the first re sistance level is the reference day's close. The only ex ception arise s when the low the day a fter the reference day exceeds the close of the reference day. In this in stance, the first level of resistance is determined by multiplying the close that day thelevels ratio of 1.0556 (see Figure 2.17). Theon second and by third resist ance are calculated by using the reference day's low. When defining the first level of support, as well as the next two levels of downside price projections, a 82 Trend Factors™ 83 Source: Logical Information Machi nes, Inc. (LIM), Chicago, IL. Figure 2.17 A and A' identify days after ref erence lows in which price gaps exist—the low fails to intersect the previous day's close. Consequently, the Trend Factors are multiplied by the closes on these days. similar evaluation of closing price relationships must be done. If th e close of the reference day a s well as the day prior to the reference day are up closes, then the high of the reference day is used to project the critical support level and the second support level is projected by multiplying the first support level by .9444. The thi rd supp ort lev el is .9722 times th e second. I f eith er the close the day before the reference day or the close the d ay of the reference day is a down close, t hen t he same ratio is multiplied by the reference day's close. The only exception arises when the high the day after the reference day is less than the reference day's close. In this instance, the first level of support is deter- Source: Logical Information Machines, Inc. Close of sec 7 units ago AND Low of sec 2 units ago > Close of sec 8 units ago AND Low of sec > High of sec 5 units ago AND Low of sec > High of sec 6 units ago then 0 Else 1 Endif RETURN ( num_zero * num_zero2 * varl ) + ( var2 * num_zero * num_zero2 ) ENDMACRO COLUMN MACRO AbsDailyVal ( VARS SECURITY sec } var3 var4 INITIALIZE var3 := AbsVal ( High of sec - High of sec 2 units ago ) AND var4 := AbsVal ( Low of sec - Low of sec 2 units ago ) RETURN var3 * var4 ENDMACRO RETURN TimePeriod sum of sub_values ( sec 1 ENDMACRO / TimePeriod sum of AbsDailyVal ( sec } I recommend t hat you experiment with both longand sh ort-term versions of the indicators. By using long-term parameters, you can get a fix on the longterm trend or market environment. By using a shortterm indicator to enter a trade at a low-risk entry point, you can fine-tune your entry and be confident that the trad e is in the context of the marke t's trend. I have included the REI and the DeMarker Indica tor to illustrate how easy it is to create your own pro prietary indicators. Along with a modicum of creativity, it takes a genuine desire to rise above the trading crowd. When I first attempted to accomplish such goals, there were no personal computers or soft ware for me to rely on. This is definitely not the case today. In short , if you are determ ined to become suc cessful as a trader, no such excuse e xist s for you not to perform such functions. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 3.7 Note how movements above 70 and subsequently below in the DeMarker Index coincide with price peaks and, conversely, how index move ment below 30 and then above 30 identifies price lows. 100 Overbought/Oversold Wave Analysis Source: Logical Inform ation Machines, Inc. (LIM), Chicago, IL. Figure 3.8 Note that, once the indicator declinesbelow 30 and above 70, potential turning points are identified. Table 3.3 DeMarker 13 Days ATTR MACRO DeMarker (SECURITY Sec, PERIOD TimePeriod ) RETURN MovingAvg ( If High of Sec > High of Sec then High of Sec - High of Sec 1 Endif, TimePeriod ) / MovingAvg ( if High of Sec > High of Sec then High of Sec - High of Sec 1 Endif + If Low of Sec > Low of Sec 1 else Low of Sec 1 unit ago - Low 1 unit ago unit ago Else 0 1 unit ago unit ago Else 0 unit ago Then 0 of Sec endif, TimePeriod ) Shortly after being introduced to the Elliott wave concept in the early 1970s, I sought out experts to educate me regarding this ap proach. Unfortun ately, the only practitione rs of whom I was aware were Joe Collins from St. Louis and Jack Frost from Canada. I con tacted both , an d they, in turn , referred me to a number of investors who had experimented with Elliot t wave analysis and with Fibonacci numbers. Two individuals, both from Florida and both physicians, were recommended to me as possibl e resource s. The experience an d information they provided—more precisely, the lack thereof—were instrumental in the creation of my own approach to wave analysis. By recounting two unrelated incidents, I might better commu nicate the effort in futility that I expended. I invited one doctor to Wisconsin to deliver to my business associates a speech about his inter preta tion of wave researc h. When I had arrived at the airp ort gate to greet him and all the passengers had deplaned and he was nowhere to be found, I called his office to determine whether he hadhad missed thethe plane. His she nurse/receptionist me said that he he made plane; had seen him reassured depart. She would find me. I continued to wait, and eventually an individual approached and asked if I was "Tom" and I said yes. He said he recognized me because I had been walking in Fibonacci angles. I 102 Wave Analysis Wave Analysis 10 3 realized that I had my hands full and that my expec tations were for the worst. He did not let me down—the meeting was totally unproductive and a complete dis aster. Rather than give up entirely, I proceeded to in terview the other wave specialist over the telephone before extending an invitation to visit. It proved to be a wise decision. My concerns were justified. I learned tha t the doctor was an Elliott wave and a Fibonacci en thusiast and that his life's routine had been guided by both. For example, he informed me that he had been married th ree times, h ad five children, made it a prac tice to work for eight consecutive days without a break, and then would take 13-day vacations. I had heard enough. I was determined to conduct my own research and draw my own conclusions regarding the "experts" the identical chart to analyze separately, there would be ten totally different interpretations. Given all the derivation applications, such as retracem ents, if the cor e is lacking substance, how can its by-products be relied on? To address these seem ingly legitimate concerns and, at the same time, capi talize on the wave principle, I developed my own approach to wave identification. It is distinguished from the conventional version in the sense that it is definitive, concise, and logical. Elements of the Fibonacci number series are in terwoven throughout the Elliott wave principle. The thread stretc hes from the numbe r of waves to the extent of retracements and of price projections. Un fortunately, prior to my development of a totally me subject matter. I was convinced that something of value existed but I had to uncover it myself. From the construction of the Egyptian pyramids to the conver sion of kilometers to miles, the pervasiveness of Fi bonacci numbers and relationships was obvious, but it was a struggle to decode and to define the role of Fibonacci and of wave behavior in the markets. My conclusions regarding Fibonacci retracements were discussed in Chapter 2. In this chapter, I discuss my wave techniques. chanical wave method, no one had produced the fabric necessary for an objective approach to wave analysis. Simply, my approach—D-Wave analysis— incorporates patterns that are defined by a series of highs and lows. The number of days selected for each series is Fibonacci-derived. Although the specific Fi bonacci numbers used may vary, the same require ments apply. In each instance, however, sufficient data must be accumulated to construct a workable template. Specifically, I identify a 13-day high close— a close greater than all previous 13 days' high s. Next, I locate the first close, subsequent to the 13-day high close, that is an 8-day lowest close—a close less than all previous 8 days' closes. Once those points a re iden tified, the first wave is complete. The second wave is not begun until a 21-day highest close is recorded—a close greater than all previous 21 days" closes. The second wave is considere d complete once a 13-day low est close is formed—a close less than all previous 13 days' closes. Finally, th e third wave begins once a 34day high is realized—a high greater than all previous 34 days' highs. The third wave is considered complete when a 21 -day low is made—a l ow less than all previ ous 21 days' lows {see Figure 4.1). As with most other illustrations presented throughout this book, the ex amples and references demonstrating various con cepts are replete with daily charts. This in no way implies that their application is limited to just that The elevation of wave analysis from total obscu rity in the early 1970s to today's widespread accept ance is impressive to say the least. Approximately 20 years ago, I literally exhausted myself attempting to integrate the facets of both wave identif ication and ap plicati on. Retrospecti vely, nothi ng worked better, but to apply it prospectively was next to impossible. There existed no hard and fast rules to accurately define the complet ion, let alone the inception of the wave patte rn as it was unfolding. It was as if one were trying to catch smoke—it was visible but elusive. Some proclaimed practitioners have applied the theory successfully, but, wil whenl lack asked, their explana tions rega rding specifics consistency an d will be riddled with exceptions. What has always disturb ed me is, given the widespread usage of the "theory," why are the interpretations so diverse? Jokingly, I have often remarked that were one to give ten Elliott 104 Wave Analysis Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 4.1 Identified on the chart is a map o f the D-Wave concept. At point A, a 21-day low is identified—a low less than all previous 21 days' lows. Up side movement labeled 1 actually was identified 4 days before the peak— once price closed greater than all previous 13 days' highs. D-Wave 1 was officially completed once a subsequent close was recorded tha t was less than all previous 8 closes. The next wave up began on the first day price closed greater than all previous 21 days' closes. This upleg was completed once price closed less than all previous 13 days' closes. The final wave up began once a 34-day high was recorded—a high greater than all previous 34 days' highs. This wave was completed once a subsequent 21-day low was made. Time is not a critical factor other than in establishing the number of days to complete waves. Price movement is important, however. time frame. These principles apply equally to all peri ods, whether hourly or monthly. As is apparent by examining the sequence of numbers defining the three waves, the number of days requir ed to establish the lows is .618 of the num ber of days necessary to qualify the highs. A critical Figur e 4.2 Point A identifies a 21-day low close—a close less than all previ ous 21 days' closes—and thereby initializes D-Wave 1. Once a 13-day high close is recorded and subsequently an 8-day low close is made, D-Wave 1 is completed. Note point A, which is the 21 -da y low close. consideration is the point of inception. It is essential that at least a 21-day low close be recorded before the first D-Wave count commences (see Figure 4.2). Once the D-Wave technique has been mastered, the retracement method described in Chapter 2, including the identification of the critical price and the use of the TD Retracement Qualifiers, can be applied similarly. The Fibonacci numbers selected to identify the waves do not have to conform with those selected above. To establish a longer-term perspective, the se ries can begin with a greater number of days tha n 13. For example, 21, 34, 55, and so on, days can be used, but once the initial number is determined, all subse quent ones follow in sequence and each wave must 106 Wave Analysis Source: Logical Inform ation Machines, Inc. (LIM), Chicago. IL. Source: Logical Information Machines, Inc. (L1M) , Chicago, IL. Figur e 4.3 D-Wave 1 was properly initialized and the move to point 5 con formed well. Had this been an hourly chart instead of a daily one, the same requirements and relationships would have been adhered to. Figure 4.4 By multiplying the first price leg from A-B by 1.618, 2.6 18, 3.618, and 4.618, the subsequent price objectives identified by points C, D, and E are projected. experience a .618 number of days' low to record wave completion (see Figure 4.3). Once the D-Wave is understood and installed, it is possible to estimate price objectives by multiplying the first wave by various Fibonacci numbers, such as 1.618, 2.618, and 3.618 (see Figures 4.4 and 4.5). I have always believed that the holding period for trades can be categorized as short-, intermediate-, and long-term. I am certain that this terminology is acceptable to most traders. I am confident, however, that these same traders do not consider their defini tions for these words to be identical to mine. Whereas they may relate these words to specific time intervals, I choose to consider the m in ter ms of price movement. In the past, volatility was not as pronounced as in re cent years. Consequently, it was not unusual for a price move of 10 percent to consume a period of at least 1 to 2 months, a move of 10 to 20 percent to re quire 2 to 6 months, and a move of 20 to 30 percent to unfold over 6 months. As a result of large pools of funds deciding to buy or to sell at the same time be cause they have similar trend-following techniques or share simultaneous reception of information, prices move in vacuums quickly. What took weeks to accom plish in the past, conceivably could occur now in minutes. This is one of the major reasons why cycles are so unpredictable. In fact, this is why I feel that concentrating on time rather than price movement is 10 8 Wave Analysis Accumulation/ Distribution Source: Logical Information Machines , Inc. (LIM), Chicago, IL. Figure 4.5 Observe how, by multip lying D-Wave 1 (points A-B) by 1.618, 2.618, and 3.618, price objectives C, D, and E are identified. outdated. D-Wave analysis acknowledges the impor tance of price movement, and I believe this recogni tion is one of the most important elements of this theory. This has been a superficial description of D-Wave analysis, but it sends a distinct message: If the ana lyst subscribes to the belief that market advances and declines unfold in waves, it is not difficult to translate these movements into patterns that are objectively identifiable. Such a procedure facilitates the process of price projections and retracements. In addition, it ensures consistency and uniformity in wave identifi cation and selection. My introduction to conventiona l technical analysi s and to its preoc cupation with subjective and artistic interpretations left me totally frustrated. Consequently, I rebelled and reverted to the other ex treme: I searched for techniques that were totally objective and me chanical. At the same time I was poring over charts and divining mechanica l cha rt techniques, I employ ed my economi cs and math e matics background to create supply-demand models capable of identifying buy and sell opportunities. My journey to accomplish this goal is described in this chapter. Although I initially experimented a nd developed these v arious techniques for the equity markets, my research confirmed that, with a fe w minor adjustments, these same me thods could be appl ied to the futures markets. Through the use of various price volume studies a nd formulas, I ultimately created the produ ct that satisfied my needs. I had learned in Economics 101 tha t an increase in demand oc curring at the same time as static or diminishing supply translated into an advan ce in price. Conversely, an inc rease in supply tha t was coincident with co nstant or reduced demand caused price to decline. > 110 Accumulation/Distribution Accumulation/Distribution 111 With those principles in mind, I researched all the tec hniques I could find that dealt with price move ment and volume. This mission included the basic onbalance volume approaches used by various market analysts. Specifically, in these instances, price activ ity was compared to numerous volume-wei ghted calcu lations. The b asis for this type of analy sis was simple: volume is considered the fuel required to move prices up an d down. By successful ly identif ying whethe r the big buyers are accumulating or distributing their posi tions, a trade r can benefit by them. Big-block stock ac tivity is generally associated with large, sophisticated, and informed investor s. To capitalize on and to partic i pate in both their research and expectations, it is im portant that a market model be sensitive to shifts in position was becoming more profitable. I asked the name of the stock he was involved in and he said Dis count Corp. My reaction was uncontrolled laughter. This individual personified the true market analyst— he knew absolutely nothing about the fundamentals of the company in which he was invested, let alone what type of business it was in. The company he be lieved to be a retail operation was actually a govern ment securities broker. This trader was a devoted disciple of the market and, with the exception of this one momentary lapse, he did not allow fundamentals to interfere with the signals generated by the systems he used. This is, admittedly, an extreme example; nevertheless, it describes the extent to which some traders ignore fundamentals and concentrate on their supply and demand caused by their campaigns. I have always described technicians as parasites because, in the true st sense, they are not infor med nor do they care about the fundamental factors contribut ing to investment decisions. Their only goal is to iden tify and to ride the trend. An episode that occurred many years ago demonstrates the tru th of this state ment and best typifies the personality and attitude of a pure trader. I introduced a close friend of mine to one of the marke t timing systems 1 ha d develop ed. He was so fascinated with its ability to mechanically identify and forecast price trends that he invested his own money in a trade based on a signal generated by this system. I was unaware that he had done so. By chanc e, d uring one of our pho ne conversations, he was interrupted by the release of information over the newswire regarding retail sales. Apparently, the news was totally unexpec ted. His response was, "There go es my profit." I asked what he was talking about. He in formed me that he was so fascinated by a technique that I had shared with him that he took a position in a stock that had generated a signal. He was now con cerned that this fundamental news was going to affect his position adversely. I remarked that he was not a fundamentalist and he should not concern himself with the news anyway. He agreed and at the same time made the observation that whereas all the other retail stocks were reacting negative ly to the n ews, h is timing models. In their trading lives, there exists no color gray, only black and white. Whereas fundamentals dictate the movement of prices over an extended period of time, short-r un price wrinkles a re best identifi ed by employi ng mark et tim ing devices and techniques. Occasionally, short-term price movements successfully camouflage the under lying price trend established by the large operators, but, because volume typically precedes price move ment, the prevailing trend can be identified by price change and volume analysis. One simple technique developed to detect the ba sic trend merely involves the summation of daily vol ume; for example, if the closing price for the day is up, a positive value for that day is added to the cumulative total. Conversely, if the closing price for the day is down, a negative value for that day is added to the cu mulative total. The index created is compared to the actual daily price movement, and divergences are identified to forecast price flow (see Figure 5.1). Another, more complicated method analyzes each transaction (tick) and continuously recalculates the index multiplying theRather price than change number of by shares (contracts). scru by the tiniz ing each trade, however, other technicians isolate only the large-block (more than 10,000-share) trans actions, assuming they are initiated by informed, so phisticated investors whose degree of aggressiveness 112 Accumulation/Distribution Accumulation/Distribution 113 Source: Logical Informat ion Machin es, Inc. (LIM), Chicago, IL. Source: Logical Infor mation Machines , Inc. (LIM), Chicago, IL. Figure 5.1 The concept behind this method is s imple—volume precedes price movement. If the price close is up versus the previous day, accumulation is taking place. If the price close is down versus the previous day, distribution is dominant for that day. Practitioners of this method believe that, camou flaged behind seemingly random price movements, a cumulative index of ac cumulation and distribution will alert traders to the true market picture. Figure 5.2 By accounti ng for not only volume but also for the extent of the price movement (price concessions) from one close to the next, an index of accumulation-distribution can be enhanced. is revealed by their price concessions. Other volume anal ysts rely on price chan ge and volume, but merely compute their calculations daily at the conclusion of sensitive. A close friend and fellow market timer, Larry Williams, was working on precisely the same project. He convinced me that the proper reference point was the current day's open, rather than the previous day's closing price. Although all price serv ices—whether presented daily in the newspaper or in trading. They multiply the price change by the vol ume on that particular day to create an index {see Figure 5.2). All these techniques were helpful, but none ful filled my needs. I wanted something more exact and other media sources, or obtained from quote ma chines—report price change from the previous day's close, this practice does not represent a true picture of price accu mulat ion or distributio n. It is not diff i cult to conclude that what occurred yesterday is 114 Accumulation/Distribution histo ry. Be cau se of new s events, the open could spike up or down and, consequently, a particular day's close could be higher than yesterday's close. As a re sult, what ostensibly appears to be accumulation or distribution based on the relationship of the current close versus the previous day's close would in actual ity be precisely the opposite when compared to the same day's opening price (see Figure 5.3). Further more, to acco unt for those exceptional cases whe n the open price is significantly different from the previous Source: Logical Informat ion Machines, Inc. (LIM], Chicago, IL. Figure 5.3 What occurre d yesterday is h istory. A more meaningful relation ship than closing price movement is the relationship between a price close and that same day 's opening price. If the close is greater than the open, accumula tion has taken place; conversely, if the close is less than the open, distribution has taken place. Accumulation/Distribution 115 day's close, an adjustment formula can be Inserted to account for the price gap and to deemphasize the movement from the current day's open and close. It was apparent that the current day's price range was a major component in measuring accumulation and distribution. By comparing the movement from the current day's close and open with that of the high to the low and by incorporating the factor of volume, a significant basis for a legitimate supply-demand model can be constructed. What I finally arrived at, however, was something considerably more complex and sensitive to shifts in supply and demand than this relatively crude approach. Specifically, although the relationship between price activity and the index was a good indicator of the direction of price, it was virtually impossible to compare the relati ve attractive ness of various secu rities because some issues were much more active (in volume) than others. Let me de scribe and explain in more detail how I was able to rec oncile the issues of both comparing and ranking securities. Just because I present my conclusions regarding the relationship between price change and volume, I do not me an to imply that my approach is necessarily the best; rather, it is the one I created and rely on after having researched and examined countless others. Its features are its logic, simplicity, and versatility, as well as its integration of numerous analytical ap proaches. Once mastered, it is designed to afford the user the ability to evaluate on a relative basis a large number of securities. Ideally, he will be able also to draw inferences regarding the cause of price ad vances or declines—was the rally the beginning of a sustain ed advance or jus t short covering? Rather than recite numerous other benefits derived from the techniques described herein, I will present them and highlight their advantages and disadvantages. The techniques that follow are related, and the composite approach presented evolved over a period of many years. The creation of the nucleus was the most critical element in the process. Often, I chal lenged the logic and the foundation of my basic as sumptions. Given what was available in the public 116 Accumulation/Distribution domain, however, I was convinced that the basis was sufficient ly sound to sup port all my derivat ive studie s. As I discussed above, the critical item in distinguish ing between accumulation (demand) and distribution (supply) is the reference point selected. Many years after I had concluded that the open was the proper pivot price for almost all measures of accumulation/ distribution, I had this supposition confirmed by one of the major operators in the stock market. More specifically, everyo ne is aware of the st ature and dom inance assigned th e spec ialists on th e New York Stock Exchange. I h ad the ple asure of establi shing a special kinship with one of the most respected of these indi viduals. I shared with him my theories and formulas regarding accumulation and distribution and the im portan ce of the open price. He was amazed to learn of both my discoveries and my approach to analysis. He indicated to me that I had accomplished mathemati cally precisely what he had acquired intuitively over many years on the floo r. To transla te it into a workable code enabling an investor to monitor a large number of securities simultaneously was an effort he had previ ously believed unattainable. His endorsement rein forced my commitment to this method and further research. Although the formulas are elementary, in order to fully appreciate their power and potential I recom mend you experiment thoroughly with every facet of the concepts of accumulation/distribution presented (see Figure 5.4). The basis of all measures includes this formula: Source: Logical Information Machines, Inc. (LIM), Chicago, IL . Figure 5.4 By using the price cha nge from close to open, as well as the vol ume, another index can be constructed. In other words, this calculation depicts the relation ship of the clos e on a parti cul ar day to tha t day's open. price range for tha t day (high to lo w), and t hen multi plying this ratio by that day's volume. In and of itself, this index value, when run cumulatively and com pared with t he underlying price activity, is a good in dicator of future price movement. Before discussing the technique for standardizing If it is positive, it can be argued that price accumula tion has taken place; if it is negative, price distribu tion has occurred. The intensity of the accumulation or distribution can be determined by relating move ment fro m the open to the close, comparing it with the securities, it is important to present an adjustment to this formula to compensate for significant opening price gaps of eight percent or more (see Figure 5.5). In these rare instances, even if price retraces back to the open, th e exceptional inf lation or deflati on in the open Close-open High-low Volume for th at part icu lar day 118 Accumulation/Distribution Apr May Source: Logical Information Machines, Inc. (LIM), Chicago. IL. Source: Logical Inform ation Machines , Inc. (LIM), Chicago, IL . Figure 5. 5(a) By dividing the price movem ent from close to open by the price movement traversed for the entire day, the degree of aggressiveness is determined. For example, if price were to open on its low and close on its high on a particular day, itwould suggest a more intensiv e buying campaign than if price were to open or close at the daily price midrange. Figure 5.5(b) One chart [5-5(a)] illustrates with vo lume included and the other [5-5(b)] with no volume. price must be considered by introducing a formula that supersedes the srcinal one. To calculate the buying (accumulation) pressure with an open or close of eight percent or greater than the previous day's close, the then multiplied by this value and added to the cumu lative index. Conversely, in order to calculate the sell ing (distribution) pressure with an open eight percent or more less than the previous day's close, the differ ence between yesterday's close and today's low is added to the difference between today's high and to difference today's high and today's yesterday's is added tobetween the difference between closeclose and today's low. From this sum, the difference between to day's high and today's close is subtracted. In turn, this value is divided by the difference between today's high and yesterday's close. The volume for the day is day's close. From this sum, the difference between to day's close and today's low is subtracted. In turn, this value is divided by the difference between yesterday's close and today's low. The volume for the day is then multiplied by this value and added to the cumulative index (see Figure 5.6). 120 Accumulation/Distribution Accumulation/Distribution 121 2. Determine the difference between yesterday's close and today's low (if less than zero, then ig nore), and add the difference between today's high and today's close (selling pressure). 3. Add together the buying and selling pressure, and divide this value into the buying pressure figure if today is an up close; divide this value into the selling pressure if today is a down close. 4. Multiply this number by the total volume for the day and add to the cumulative index (see Figure 5.7a). Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 5.6 When an open occurs that is eightpercent or moregreater than or less than the previous day's close, an adjustment must be made to accen tuate this atypical price behavior. Additionally, this formula can be used in lieu of opening prices when they are unavailable. Should the open price be unobtainable for any reason, a version of the gap compensation formula would serve as a credible substitute. A few revisions are required, however: 1. and Calculate the difference between today 's high yesterday's close (if less than zero, then ig nore) and the difference between today's close and today's low, to arrive at a measure of the buying pressure (buying pressure). What you have learned so far are variations of what you may have already seen in the public domain. What I am about to share with you now is proprietary and essential to the ta sk of comparing various securi ties to determine relative attractiveness. The concept is easy and straightforward, but it is important that you follow each step to ensure complete understand ing and total mastery. After the formula for calculating accumulation/ distribution has been selected—I recommend the one using the open reference with the adjustment for open gaps of eight percent or more— a sch edule of variou s time intervals is to be selected. I recommend an ar ray of Fibonacci numbers beginning with five days and extending to 13, 21, 34, 55, 89, 144, 233, and 377 days. Each day, a value representing buying or selling pressure appears. Add together all the positive (buy ing pressure) numbers over the prescribed period of days, and then add together all the negative (selling pressure) num bers over that same period. Then divide the su m of all the buying pressu re values by the abso lute value of the sum of all the buyin g pressure values plus the selling pressure values. This number is total a ratio all defining the buying pressure divided by the activity (buying pressure plus selling pressure) and can be converted to a percentage by merely multiply ing times 100 percent. 122 Accumulation/Distribution IBM Volu Source: Logical Informat ion Machines, Inc. (LIM), Chicago. IL. Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Figure 5.7(a) The adjustment for exceptional (eigh t percent or more) open gaps is introduced into the basic formula. Figure 5.7(a) includes volume. Figure 5.7(b) This chart does not include volume. In order to get the flavor of wh at is tak ing place withi n the dynamics of the market, t his procedure can be applied to other time periods in the same m an ner. These percentages measur e the de mand over vari ous time intervals, so it is possible to relate one security to another to determine which is being more most reliable in identifying attractive investment op portunities. The rate of change is calculated easily: divide the current day's percentage by the percentage X days ago. I typically wor k with Fibonacci numbers . Once a particular number is selected, I calculate the rate of change by dividing the value that day by the value at least four Fibonacci levels lower. For example, aggressively accumulated or distributed. A chart of each security is even more helpful in displaying the movement of thi s oscillator (see Figu re 5.8). Even more important, however, is an indicator tha t displays the rate of chang e of the perce ntage s. In fact, my experience has proven this value to be the if I employ an 89-day series, to calculate the rate of change I would compare today's value with tha t of 13 days ago—i n the Fibonacci series, 13 increase s next to 21, 34, 55, then 89. As you can see, 13 is positioned four degrees beneath 89. If one were to use a 144-day series, then a comparison between the current day's 124 Accumulation/Distribution Accumulation/Distribution 125 Source: Logical Information Machines , Inc. (LIM), Chicago. IL. Source: Logical Informat ion Machines, Inc. (UM |, Chicago, IL. Figure 5.8 It has always been difficult to relate the attractivenes s of one se curity versus another. By calculating the percentage of buying pressure di vided by the total pressure (both buying and selling), comparisons can be made. This measure is a serious breakthrough in trading analytics. Figure 5.9 By determining the rate of change of buying pressure divided by the total pressure (both buying and selling), the degree of aggressiveness among various securities, as well as for individual securities, can be measured. value and that 21 days ago would be used; if one were remain static for all securities compared. For example, if an 89-day value is used for one stock and the rate of to use a 233-day series, then a comparison between that day's value and the value 34 days earlier would be selected. Keep in mind that these are merely sugges tions. You may have greater success at using other number series or comparing other periods of rates of chan ge. Once sele cted, however, the period should change is based on the value 13 days prio r, then the same time periods should be used when evaluating the relative appeal of other stock candidates (see Figure 5.9). Once a value is selecte d, t he index for most stocks will move within the same band. Experimentation with that index will yield the parameters associated with 126 Accumulation/Distribution Accumulation/Distribution 127 stocks have no restrictions as to the upside or down side movement they can traverse daily, futures have prescribed daily price limits because of the extreme leverage involved in those markets. When price moves limit, trading is virtually suspended. Although trans actions can be effected at those price extremes, de pending on the pool size looking to buy or sell, the volume the marke t is capable of producing may be sig nificantly less than if there were no price limits. To ac count for this pent-up supply or demand, I recommend combining all consecutive days extending from the first day a limit move occurs until the last day of the series. The open price of the first day and the closing price of the last day, as well as the range and the vol Source: Logical Informati on Machi nes, Inc. (LIM). Chicago, IL. Figure 5.10 No volume and a different time period (34 days versus 34 days 5 days ago) are presented on this chart. price tops and bottoms. Generally, the rate of change of this index will reverse prior to the actual price turns. Together with other techniques, entry and exit prices can be identifi ed and the relative attraction of a situation can be compared with other opportunities. The above technique describes an accumulation/ distribution model for stocks. The same approach can be applied to futures, with one exception. Whereas ume for This the entire period, are treated as ifapthey proach were one day. method incorporates the basic described and compensates for the shortcoming asso ciated with limit moves. Some success is derived by ex cluding volume totally and running the formula described in Figure 5.9 with various other time peri ods both long and short term (see Figure 5.10). As you can see, th e model just described ha s ap plication to both equities and futures. Variations of this model have equal application to these same mar kets with similar results. As with every other tech nique presented in this book, the best results are realized in combination with other proven approaches, thu s enhancing the prospe cts of trading success. Co n sequently, my recommendation is not only to experi ment and to introduce the accumulation/distribution model into your trading tool kit but also to utilize other methods that confirm your trading rules. Moving Averages For many years now, one of the most popular trend-following meth ods has been the moving average. The simplicity of its construction and the e ase of its interpretation contribute to its widespread usage and acceptance. Unfortunately, this tool's success is derived from a particular market's ability to trend. My research indicates that markets generally move in trading ranges and trend much less fre quently. Historical observations suggest that, approximately 75 to 80 percent of the time, price of a partic ula r security tends to move in a trading range. On the other hand, 20 to 25 percent of the time, price trends are either up or down. Furthermore, additional re search shows that price accelerates in a downtrend generally about two to two and a half times faster than price in an uptrend. This phenomenon can be easily accounted for by the fact that whereas investors typically accumulate a position over a period of time, their recognition of a pri ce decline is immediate a nd th eir tendency is to liquidate the entire position at one time. The most common and b asic c alculation of the moving aver age is arithmatic and involves only adding together the closing prices of a security over a prescribed period of time, dividing this sum by the total number of entries, and then plotting that period's value on a chart coincident with that interval's price range. Unfortunately, ) 130 Moving Averages this is the common practice but it is not necessarily the ideal or the correct one. Most market timers ig nore or fail to recognize many questions that arise. Specifically: • Why is each time period equally weighted when in fact the most recent price activity is more im portant? • Why is the final calculation plotted on a chart immediately beneath the last price entry when an average is calculated? • Why are only the closing prices averaged and other critical price points such as open, high, and low overlooked? • Why are some periods of time more popular than others? • Why are moving averages so widely accepted and used when in most instances they are_applied to trading range mark ets which causes the u ser to get continuously whipsawed? My experience suggests that the results achieved by using traditional moving averages are no better than those realized by using most other conventional trend-following approaches. Moving averages are eas ily calculated and understood and can be found on most quote service graphics displays and in almost all graphics software packages. Don't confuse this uni versal availability with utility and trading success, however. I have found that, in this industry, there is no correlation between acceptance/usage and per formance results—in fact, precisely the opposite is true in most instances. Given some exceptions, despite extensive re search, I have uncovered only a few circumstances in which moving averages can be applied and respectable results can be expected. Specifically, by definition, moving averages identify turning points in trend well Moving Average s 13 1 after they have occurred. As I stated earlier, markets operate within a tradin g range most of the time. Oc casionally, however, prices do break out of this pat tern. My adaptations to moving average analysis have proven to be worthwhile in those breakouts be cause the risk of being whipsawed is diminished con siderably. Basically, the various moving average techniques I recommend all cope differently with the issue of trad ing range whipsaws. One version projects moving av erages into the future; another averages highs, lows, and closes for a period of time to create a fictitious av erage price to compare with the moving average; and still another employs a moving average only when price breaks out of a trading range. A description of each is presented here. Conventional trading analysis provides for the moving average entries to be aligned with the trading days such th at the las t movin g average entry coincides with the last price entry. There is nothing sacred about this particular relationship, and this practice has always bothered me. I experimented with center ing the moving average and realized some improve ment in performance results. Instead of calculating the moving average and positioning the values beneath the most recent entry, as most traders would, I experimented with the technique of having the pro jec ted averag e co incide w ith the cu rre nt d ay' s pri ce. In a sense, you might say that 38 percent of the moving average band appears prior to the current price entry and 62 percent is projected into the future. In other words, 62 perce nt of the movi ng average has been p ro jec ted into th e fut ure . I found th at th is shift re tai ne d the pattern of the moving average and at the same time reduced the likelihood of whipsaws inherent in trading ranges. Another approach for calculating a series of mov ing averages to avoid the problem ranges and ofattempts whipsaws by making certain of thattrading the short-term moving average exceeds upside the longterm moving average in the case of a buy signal, and that the short-term moving average exceeds downside / 132 Moving Average s Moving Averages 13 3 the long-term moving average in the case of a sell signal. At the same time, both moving averages must exceed either a fictitious price peak or price trough that is an average of the most recent two days' highs and two days' lows. I have bee n the most comfortable using moving averages of 5 and 21 days' duration. To demonstrate just how this method works, calculate the value of each moving average by summing the opens, highs, lows, and closes over each respective time period. Next, project both averages into the fu ture: in the case of the 5-day average, project 3 days into the future; in the case o f the 21-day moving av erage, project 13 da ys into the future. If the 5-day projected value is more than the 21-day projected value, the n look to buy; if the 5-day projected value is applied. In every instance , however, the key ingredient of any approach is its ability to rem ain d ormant while price moves sideways. Once price breaks out of the trading range, the technique should be sufficiently sensitive to detect any movement that would precede a trend reversal . For many years, I observed a central tendency for price activity to move within a band defined by a mov ing average that was identified by multiplying each day's price low by 110 percent and each day's price high by 90 percent. This band can be smoothed by multiplying an average of the previous 3 days' lows and highs and by increasing the band factors to 115 percent and 85 percent. When price exceeded this moving average band, overbought and oversold read less than the 21-day projected value, look to sell. Performance results can be enhanced further by making certain that both averages exceed the hypo thetical two-day price high (buy) or low (sell). Further improvements are realized if both moving averages are declining or advancing together. Finally, by making certain that the 5-day average is greater than the 21day average for a buy signal and less than the 21-day average for a sell signal, you should realize improved results. To avoid a multitude of signals while locked in a tradin g range market, I created a movin g average sys tem that only became active when price recorded ei ther a 13-day-high low or a 13-day-low high. Let me explain this concept further. If price advances and it records a low greater than all previous 12 lows, then a 3-day moving average of lows is installed and fol lowed for a period of 4 trading days to identify a place to sell. Conversely, if price declines and it records a high less than all previous 12 highs, then a 3-day moving average of highs is installed and followed for a period of 4 trading days to identify a place to buy. ings were generated. The percentages presented can be adapted to specific markets. One technique I developed many years ago I call the TD Moving Average techni que. It is designed to ini tiate buy a nd sell signals on t he first day both of two moving ave rages—a long term a nd a short term—turn up or down simultaneously for the first time. Gener ally, the short-term average responds first and subse quently the long term confirms—that is the day action is to be taken. In other words, the first instance they both move up or down versus the previous day's TD Moving Average readings is the trigger day. Typically, the two moving average periods I use are 13 and 55 days, but the latter period has been adjusted to as many as 65 days. I believe another approach h as merit but, because of both software and data constraints, I have been un able to test it. This method involves identifying and av eraging the median (middle) price recorded each day for a parti cular period of the day. I p lan to experiment with variations of this technique now that I possess the software required; I am awaiting the necessary The is active a period only 4 days moving after theaverage higher low or the for lower high is of recorded. As you can see, the application of the moving aver ages is depende nt on the fulfillment of various prereq uisites. Other variations of this approach can also be data.My moving average techniques are unconven tional. They have been designed to counteract the nemesis of all moving average approaches—trading range and sideways markets. I believe that these 134 Moving Averages methods circumvent the obstacles confronting the average trader. Together with the other trading ideas presented throughout the book, these approaches can be implemented to giv e the savvy tr ader a mark et edge. Sequential At the time I entered the investment business, it was commonplace to attempt to identify price tops and bottoms by using cycles. The length of these cycl es was determined by calculating either the n um ber of days from one price low to a succeeding price low or the number of days from a particular price low to a subsequent price peak (see Figures 7.1 and 7.2). This method of market timing is subjective and, because the periods are not static, it does not lend itself t o statistical analy sis and testing. In fact, the interpretation of cycles is sufficiently vague that, often, where a low or a high might be expected to be found, a condition called "inversion" occurs in stead: prices actually do the opposite of that which was anticipated. I was disturbed by this lack of predictability. Consequently, I exper imented with the application of the Fibonacci time series to cycles and obtained somewhat better results but nothing extraordinary. I have always been skeptical of the practice of relying on cycles to identify price tops and bottoms. It is difficult for me to accept the fact that an a rbitrarily derived number of days possesses repetitive properties. Quite to the contrary, my research studies suggest that the price action of some trading days is actually meaningless. Consequently, I researched exhaustively to create a technique that would employ a mechanized timing device to identify price highs 136 Sequential™ Sequential" 137 CIS Bond (weekly) Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Figure 7.1 The approximate period of time from a price low to a subse quent price low is 39 weeks. Each low for this cycle is marked with an X. Source: Logical Information Machines, Inc. (LIM), Chicago, IL . Figure 7.2 The approximat e number of weeks from a price low to a subse quent high close is 10 weeks. The periods are marked X and Y. and lows as they occurred. I experimented endlessly to determine what price relationships typically ap peared prior to and coincident with market turning points. Let me describe, in basic terms and by using simple examples, the process I followed. Once demand exceeds supply, price advances. Whether this advance is due to short covering, pur chase recommendations, positive news, or any other factors is immaterial. At a certain point in time, figu typicall y, prices peak not because of sophisticated and knowledgeable sellers' accurate identification of the price high, but rather because the "last buyer" has bought. Conversely, as supply exceeds demand, price declines. Whether this selloff is attributable to bad news, sell recommendations, short selling, or any other factor is immaterial. Eventually, however, all the potential sellers will have sold. My research has proven that price bottoms are made once the last seller ratively speaking, all the potential buyers will have bought. Unless there is a catalyst to entice a new crop of buyers , th e mar ke t is vulnera ble to decline for two reas ons: (1) an exhaus tion of buyi ng or ( 2) an incr ease in the pace of selling. My experience confirms that, has sold and, by default, price moves when aggressive buying prematurely occursup. at In anfact, interim low, typically it is associated with a short covering rally. Consequently, after the buying frenzy dimin ishes, a price vacuum is created and ultimately price 138 Sequential™ Sequential" declines even faster and further until a new equi librium level between supply and demand is estab lished (see Figure 7.3) . At that point, if the d ynamics are right and the selling pressure has been exhausted, price has an opportunity to reverse upside. As price advances, it approaches its ultimate peak; conversely, as price declines, it moves closer to its eventual low. These comments are made not to in sult your int ellige nce, but rat her to emphasize th e ob vious: selling into strength and buyin g into weakness are practices often overlooked and believed unattain able by traders. As a result, most tra ders a re tren d fol lowers who subscribe to the notion that a particular trend will continue in force. They b elieve tha t attem pts DJIA (weekly) Apr 1973 Jim Aug Sep Source: Logical Information Machines, Inc. (LIM), Chicago, Nov Dec Jan 1974 IL. Figure 7.3 Observe the 10-week advance (A-B) and the abrupt, severe 5week decline to new lows (C). 139 to buy into a declining market are akin to catching falling daggers. To overcome this fear, some traders utilize cycles, but as I mentioned earlier, cycles are too arbitrary and subjective. Consequently, I created a technique that incorporates in vague terms the con cept of cycles without the handicap of time rigidity. Whereas cycles traders employ a prescribed time series, I rely on a dynamic set of variables that ad justs to marke t action. In other words, I wait fo r the mark et to speak to me as the price movement unfolds. Actions speak louder than words, and what better source of market direction is there than the market itself? All information known, including all the hopes and fears of trade rs, is tran slated into one impo rtant item— price. Should unexpected developments regarding market fundamentals arise and the supply-demand equation be shifted as a result, the price movement should reflect this change. Just as the market's price personality constantly undergoes change, so too any system that attempts to identify price tops and bot toms must adapt to these character swings in money flow and measure them precisely. This major defi ciency of conventional cycles analysis makes it inferior to the versatile approach I will share with you. My research confirms the fact that, prior to a price top or bottom, the market announces its inten tions regarding price direction loud and clear to any trader willing to listen. Specifically, the market fore warns the trader whether it is predisposed toward a price top or toward a price bottom. In other words, the market's environment or inclination to top or to bot tom is first defined by this setup phase. Because this entire approach is mechanical, as are the additional filters required to actually generate the buy and sell signals, I developed a checklist to simplify this process (see below). The feature of this approach is its design to buy into weakness and to sell into strength. Once all the prerequisite s are satisfied in the order required, a signal occurs. Hence, the name Sequential™ was given to the system. The procedure followed to establish a Sequential buy or sell signal is straightforward and uncomplicated. 140 Sequential™ Setup 141 In fact, the simplicity associated with its implementa tion concerned me at the time I developed the tech nique. I was puzzled that no one else had previously discovered and integrated the same time series and price relationships. Continuously, I checked and rechecked my studies to make certain that I had not overlooked some key element. Keep in mind that the period of development and testing was not by any mean s recent—this all t ook pla ce in the 1970s. It was conceived and researched prior to the era of comput ers. Since that time, this technique has had universal application to various markets, including stocks, fu ture s, and indexes. As a result of additional researc h, I have developed enhancements to the srcinal Sequen tial™, but the core of the method still exists and thrives. How many other defini tive market timing a p proaches designed to anticipate price tops and bot toms have endured and have withstood a similar test of time? Not many, if any at all. Setup To generate a sequential buy signal, the market envi ronment must first be predisposed to rally. My re search determined that a prerequisite to a buy is a particular relationship among closing prices over a period of nin e consecutive days. Specifica lly, once a period of at least nine consecutive trading closes less than the close four trading days earlier is recorded, then the buy setup is complete. For exam ple, if the close of trading on Friday is less than the close of trading on that same week's Monday (assum ing trading occurred on Thursday, Wednesday, and Tuesday), one day of a possibl e set of nine is defined. Had the close of Friday been equal to or greater than the close on Monday, however, this day would not have qualified as one day of a nine-day buy setup. Figures 7.4 and 7.5 describe the requirements for a valid buy setup. The first day of the nine-day buy setup must be preceded by a close on the trading day Source: Logical Information Machines, Inc. (LIM), Figure 7.4 Chicago, IL. Requirements f or a valid buy setup. immediately before it that is greater than or equal to the close four trading days earlier (see Figure 7.6). As is apparent from these examples, it is not uncommon to witness a sho rt-term bottom or even a price reversal upon completion of the nine-day setup. Unless the market is in a free fall or in a short-term correction within an uptrend, this short-term price hiccup is ju st a repr ieve in th e do wn tre nd an d th e dec line should resume. In order to generate a sequential sell signal, the marke t environm ent must be predispose d to a decline. The prerequisite for a sell setup is exactly the opposite of the prerequisite for a buy signal. A buy setup re quires a series of nine consecutive days' closes less than the close four trading days earlier; a sell setup 142 Setup Sequential" Source: Logical Information Machin es, Inc. (L1M ), Chicago, IL. Figure 7.5 The first close less than the close f our trading days prior to it is marked with an X. All subsequent days are marked numerically. It is not un common for price to reverse or move sideways once the setup of nine days is Source: Logical Information Machines, In 143 c. (LIM), Chicago, IL. Figure 7. 6 The close on day B is greater than or equal to the close on day A. The next day's close is less than the close four days earlier and is counted as day one of the setup. Note how price bottomed out on the ninth day of the setup. recorded. requires a series of nine consecutive trading days' closes greater than the close four trading days earlier. For example, if the cl ose of trad ing on Frida y is greater th an the close of tra din g on that same week's Monday (assuming tradi ng occurred on Thursday, Wednes day, and Tuesda y), on e day of a possible set of nin e ha s been defined. Had the close on Friday been less than or equal to the close of Monday, this day would not have qualified as one day of a nine-da y sell setup. The requirements for a valid sell setup are described in Figur es 7.7 an d 7.8. The first day of the ni ne-da y sell setup must be preceded by a close on the trading day immediately before it that is less than the close four trading days earlier (see Figure 7.9). Similar to the buy setup, a derivative benefit of the nine-day sell setup is the identification of a short-term high once the setup is formed. Unless the market is in a blow-off phase or the overall trend is defined as down, this pullback should be temporary, however, and the ad vance should resume. As you can see, the setup is very simple to estab lish. Either it requires nine consecutive trading days' closes less than the trading days' close four days be fore each for a buy signal, or it require s nine cons ecu tive trading days' closes greater than the trading days' close four days before e ach for a sell signal. It is impor tant that: 144 Sequential 1 Setup Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Figure 7.7 Nine consecutive days in w hich the close is greater than the close four trading days earlier are counted on this chart—in other words, a sell setup has been completed. Notice the price "hiccup" or "stutter" as 1 de scribe it upon completion of this series—although it is not always the case, frequently the price movement stalls or reverses upon setup completion. Each of the three days between the current trading day and the trading day's close four days ago is a trading day; 2. The tra di ng day's close of the day prior to day one of a buy setu p is greater th an th e close of the day of fo the ur days earlier, andone th e trad ing tra day'sdingclose day prior to day of a sell setu p is less tha n the close of the tr adi ng four days earlier; 145 Source: LogicaJ Informa tion Machines, Inc. (LIM), Chicago, IL. Figure 7.8 A series of nine consecutive closes greater than the close four trading days earlier defines the sell setup. Note that price established shortterm peaks upon completion of the setup period—the ninth day. 3. If the close of a tra ding day Is equal to the close of the trading day four days before it, the setup series is interrupted and must begin anew; 4. The series of consecutive closes may exceed nine but the required period for a valid setup is satisfied once the requirement of nine consec utive closes is met. There is a natural rhythm defined by the setup series of nine consecut ive closes greater tha n or less than the 146 Sequential" Setup 147 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 7.9 On the day prio r to the first day of the sell setup, the close (A) was less than the close four days earlier (B). close four days earlier. Generally, the market will ex perience a reversal or a stabilization in price at that time. In fact, as mentioned earlier, in some instances, price will record a significant tur n at ju st tha t point. These observations are universal and apply to all mar kets and to all time intervals. I made an observation several years ago regard ing the comparison between the most recent price setup and the most recent priceprice setup in the other price direction. As the current setup is being formed, I compare (1) the extreme price peak or low— depending on whether the movement is up or down— recorded from the first day of the most recent setup Source: Logical Informa tion Machines, Inc. (LIM), Chicago, IL. Figure 7.10 Note how the nine consecutive up closes (setup) in early-mid May exceeded the nine consecutive down closes (se tup) recorded in la te April. through its completion with (2) that of the most re cent "inactive" setup through its completion. An in active setup is defined as one in which the series of consecutive closes versus closes four days earlier has numbered at least nine but has been interrupted and one in which the trend contradicts the current setup (see Figures 7.10 and 7.11). By definition, this must be the case because the current setup occurs in the other direction. Although, technically speaking, the setup is comprised of a series of nine consecutive closes greater than the close four trading days earlier for a sell (and less than for a buy), the comparison of the two setups does not necessarily require the 148 Setup Sequential™ 149 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 7.1 1 In both instan ces (Figure 7.10 as well), you can see a series of setups. In both cases, note short-term top (bottom) generally seen after nine setup—price "hiccup" or "st utter" if you will. Figure 7.12 Price interse ction did not occu r on day 8, but it did on day 9 once the price high that day exceeded the low three day's earlier—day 6 in this example. completion of the e ntire curr ent series of closes be cause the peak or trough can exceed the trough or the peak—depending on whether the current setup is up or down—of th e inactive se tu p befo re the se tu p is completed. In fact, it may not be formed at all. This particular technique has enabled me to define the trend of vario us mark et s on nume rous occasions, and it is a valuab le derivative benefit of a Sequential" setup. entry. Conversely, to avoid the problem of early entry in an upside blow-off, an indication that the price brakes have been applied is a necessity. Once the setup has been properly qualified, the next phase of Sequential—the countd own—begins. The setup quali fication process, called "intersection," is very easy to understand. Simply stated, intersection requires that the price range of either the eighth or the ni nth day of the setup overlap the price activity of any setup day A vital element is required to validate the Sequen tial setup. Its absence underscores the fact that the market is in a runaway phase. For example, if price is declining in a waterfall manner, it is important that a retardation of the decline occur to prevent premature 1 three or more earlier. other for a buy setupdays takes place In once thewords, high ofintersection either day 8 or day 9 of the setup is greater than or equal to the low three , four, five, six, or seven days earl ier (see Fig ures 7.12 and 7.13). On the other hand, intersection 150 Sequential" May 1986 Setup Jim Ju l Aug Sep Source: Logica l Infor mation Machines, Inc. (LIM). Chicago, IL. Figure 7.13 Both setups (Figure 7.12 as well) fail to record intersection on day 8 of the setup but they do so on day 9. 151 Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Figure 7.14 In both regards—sell setup and buy setup—intersection did not occur on day 8 or day 9 of the setup, but rather on day 13 and day 11, respectively. for a sell setup occurs once the low of either day 8 or day 9 of the se tup is less th an or equal to the high three, four, five, six, or seven days earlier (see Fig day is a continuation of the setup or not. All that is required is that the price intersect the price low three or more days earlier in the case of a buy setup or in ures 7.14 andinstan 7.15). Intersection can also place in one other ce: if intersect iontake does not occur on day 8 or on day 9 of th e setu p, t he n it may oc cur on any subsequent day, rega rdless of whet her that tersect the price high three or more days earlier in the case of a sell setup. However, in both instances, the countdown phase is postponed until intersection is satisfied (see Figures 7.16 and 7.17). 152 Sequential" Countdown Jan 1993 Source: Logical Inform ationMachines, Inc. (LIM). Chicago, IL. Figure 7. 15 Whereas intersection did occur on the sell setup on both day 8 and day 9, for the buy setup it was delayed until day 10. Feb Mar Apr 153 May Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 7.1 6 The intersection for the first buy setup was postponed until day 10 and countdown does not begin until intersection and the setup are both complete. Countdown There are two instan ces in which the setup can be canceled. They are very simple and straightforward. The most common is a phenomenon called "recycling," which is described in the section on countdown below. The other is related to the closing prices recorded at any time between th e completion of the set up and the generat ion of a signal. More precisely, should a subse  quent closing price exceed either the highest intraday high—in the case of a buy setup—or the lowest intra day low —in the case of a sell setu p—th e setu p phase is canceled and it must be reinitialized (see Figures 7.18 and 7.19). Once setup has been satisfied, the countdown process begins. Countdown describes the relationship of the close to either the high or the low price two trading days earlier, depending on whether a sell or a buy setup is active (see Figures 7.20 and 7.21). With re spect to a pending buy signal, the close must be less than the low two days earlier; with respect to a pend ing sell signal, the close must be greater tha n the high two trading days earlier. Once a total of 13 closes less than the low two trading days earlier in the case of a buy, or once a total of 13 closes greater than the high 154 Sequential™ Countdown Source: Logical Informa tion Machines, Inc. (LIM), Chicago, IL. Figure 7.17 The intersection for the sell setup occu rred on day 9 of the sell setup. The intersection for the buy setup took place on day 11 of the buy setup. Source: Logica l Information Machines, Inc. 155 (LIM), Chicago, IL Figure 7.18 Note how close at point A on chart exceeded the highest hig h of the buy setup in late March counted on the chart. 156 Sequential™ Countdown 157 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 7. 19 In both instanc es, Figure 7.18, as well as this one, closing price A exceeds the highest intraday high of the buy setup. Source: Logical Informati on Machine s, Inc. (LIM). Chicago, IL. Figure 7.20 The "X" marked days identify the 13 countdown days recorded upon completion of the sell setup. 158 Sequential™ Source: Logical Inform ation Machines, Inc. (LIM), Chicago, IL. Source: Logical Information Machine s, Inc. (LIM), Chicago. IL. Figure 7.21 The "X" marked days identify the 13 countdown days initiated upon completion of the buy setup. Figure 7.22 Approximately 25 to 30 days lapsed fr om the initiatio n of the sell setup to the thirteenth day of sell countdown. two da ys earl ier in th e case of a sell are recorded, a signal is generated. These 13 closes need not occur consecutively; they will occur only rarely, if at all. Once intersection has taken place to validate comple tion of the set up—a nd beginn ing no earlier th an day 9 of the set up—th e count down begins . By definition, the countdown period cannot be completed any sooner than 12 days after setup, and that assumes that day 9 qualifies. Typically, however, one might expect 15 to 30 days to lapse between the setup and formation to start over. The srcinal setup is negated at any time subsequent to setup, and prior to a signal, a setup in the opposing direction occurs (see Figures 7.24 and 7.25). The second situation does not require additional time to form a new setup, but it does recycle (start over) the countdown phase. In this case, a sub sequent setup is formed simultaneously as the count down process is taking place. This new setup replaces the srcinal setup and is in concert with the srcinal setup, not contradictory (see Figures 7.26 and 7.27). the completion of countdown (see Figures 7.22 and 7.23). This occurs often and is a function of the market's reevaluating the supply and demand equation and reestablishing the path and the time parameters to the ultimate top or bottom. In both instances, the srcinal setup and the countdown are repealed; in the Two situations that could arise after setup would cancel countdown. The first situation invalidates the srcinal setup and requires the process of setup 16 0 Sequential™ Countdown 161 Source: Logical Information Machines , Inc. (LIM), Chicago, IL. Figure 7.2 3 Approximately 36 days lapsed from the initiation of the buy setup to the thirteenth and final day of buy countdown. Source: Logical Informat ion Machines, Inc. (LIM), Chicago, IL. Figure 7.24 The ninth day of the sell setup coincided wit h the exact price high. Subsequently, a buy setup was formed that did not co nfirm intersection until setup day 13. 162 Sequential™ Countdown Source: Logical Information Machine s. Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 7.25 The sell setup negated the buy setu p that had been forme earlier. Figure 7.26 The first sell setup was superseded by a second sell setup . 163 164 Entry Sequential™ 165 Source: Logical Information Machines , Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM). Chicago, IL. Figure 7.27 The second buy setup negated the first and became active. Figure 7.28 Mote the sell signal was generated at the high close of the sec ondary price peak. second situation, a new setup has been formed through the recycling process. Now that the setup and the countdown phases have been discussed, three other important aspects of Sequential remain outstanding: (1) the entry, (2) the exit, and (3) the stop loss techniques. Entry Three metho ds are recommended fo r Sequential entry. The first approach enters the market on the close of the day in which countdown is completed (see Figures 7.28 through 7.31). This is the riskiest entry because the setup can be recycled and the srcinal signal will evaporat e in the process. A new signal cannot be gen erated until the countdown process has been re played. Although the potential exists that the trade may produce a loss, it is the only entry of the three that offers the opportunity to buy or to sell at the ab solute closing price low or price high. The second method ensures that price does not recycle and consequently does not forfeit the active signal. However, it requires a price "flip"—a close greater than th e close four days earlier in the case of a buy, or a close less than the close four days earlier in the case of a sell (see Figures 7.32 through 7.35). By awaiting the fli p, ins uranc e is bought that a se tup will not be recycled. 166 Sequential™ Entry Source: Logical Informati on Machin es, Inc. (LIM), Chicago, IL. Figure 7. 29 example. The precise low day was identified as the signal day in this source: JLOgicai inform ation Machin es, Inc. (LIM), Chicago, IL. Figure 7.3 0 Once again the exact hig h day was the sell day—point A. 167 168 Sequential™ Aug 19B6 Entry Sep Oct Nov Dec 169 Jan 1987 Source: Logical Informat ion Machines . Inc. (LIM), Chicago, IL. Figure 7.31 By executing the trades on the close of the thirteenth day of countdown, ideal entries were selected. Source; Logical Informati on Machines, Inc. (LIM), Chicago, IL. Figure 7.32 Note the "flip" day occurred after the price peak and thir teenth day of countdown, but it confirmed the sell. 17 0 Sequential" Entry 171 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 7.3 3 See how "flip" day occurred subsequent to actual price high and thirteenth day. Source: Logical Infor mation Machine s, Inc. (LIM), Chicago, IL. Figure 7.34 Once again confirmation close occurred subsequent to the thirteenth day. 172 Sequential™ Exit 173 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Source: Logical Information Machines. Inc. (LIM), Chicago. IL. Figure 7.3 5 By awaiting the initiation of trade s until the "flip"—clos e greater (or less than) the close four days earlier—ideal entries were selected. The entry close is circled and marked A. Figure 7.36 See the first close (circled), subsequent to the thirteenth (sell) day, less than the low two days earlier. The final entry technique is to await a two-day range "flip" once the thirteenth day is identified. In other words, once the countdown is completed, buy the first time a subsequent close greater than the high two days earlier occurs or, conversely, sell the first time a subsequent close less than the low two days earlier occurs (see Figures 7.36 through 7.39). This entry perfects the "flip" entry and generally serves as a compromise between entry one and entry two. Exit There are two ways in which to exit a trade other than being stop ped out of the tr ade wit h a loss. The first method is to liquidate the position once the current setup is completed and price fails to exceed the fur thest price level recorded by the most recent inactive setup (see Figures 7.40 and 7.41). This exit assumes that because the trend has not reversed as defined by the point of termination of the active setup failing to exceed the furthest price of the most recent inactive setup, there is a likelihood of an impending reversal and thus the trade should be liquidated. The other exit also compares the two setups, but in this instance if any price recorded during the cur rent active setup exceeds the furthest price of the in active setup, then the position is held until a reverse signal is generated. 17 4 Sequential" Exit 175 Source: Logical Informati on Machines, Inc. (LIM), Chicago, IL. Nov 1969 Dec Source: Logical Information Machines, Inc. (LIM), Chicago, IL Jan 1990 , Figure 7.37 In this instance the two-day "flip" confirmation trans lates into a much less favorable sell entry than was the sale on the peak day—the thir teenth count day. Figure 7.3 8 The entry for the two-day "flip" is close to the signal day in this example. 176 Stop Loss Sequential" Source: Logical Infor mation Machine s, Inc. (LIM), Chicago, IL. Source: Logical Information Machi nes, Inc. (LIM), Chicago, IL. Figure 7.39 Rather than executing the buy on the thirteenth day's close, by awaiting the two-day "flip" confirmation, a less advantageous entry price was recorded. Figure 7.40 As you can see on the chart, the nine up setup identif ied nu merically did not exceed the down setup established beginning 11 days be fore the October price low. price range of the highest range day throughout the entire pe riod of setup an d countdown fo r a sell signal. In the c ase of a buy signal, the t rue range for the low  est range day is calculated by subtracting the low that day from the high that day or the close the previous day, whichever is greater. The stop loss level is defined as that value subtracted from the low that day (see Figure 7.42). The reverse technique is applied when calculating the stop loss for a sell signal. The true Stop Loss The final element to consider is the stop loss. Unfortu nately, not all trades are successful. To protect against the cha nce of failure, a stop loss should be installe My research suggest s that two techniques accomplish this money management goal. Both consider the price range of the lowest range day throughout the entire period of setup and countdown for a buy signal, or the 177 d. range for the highest range day is calculated by sub tracting from the high that day either the low of that day or the close one day earlier, whichever is greater. The stop loss level is defined as that value added to the high that day (see Figure 7.43). In both cases, a 178 Sequential™ Stop Loss 179 Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 7.41 The exits are generat ed if nine consecutive closes greater than the close four days earlier are recorded, a profit is realized, and price does not exceed the setup in the other direction which got you into the trade in the first place. Source: Logical Information Machines, Inc. (LIM), Chicago. IL. Figure 7.42 Subsequent to the Buy signal, nine consecutive up closes were recorded, and a profitable exit was made. subsequent close must exceed the calculated values in order to activate a stop loss. The second stop loss technique is more conserva tive. In the c ase of a stop loss for both a buy signal a nd a sell signal, the same day selection technique is used. Instea d of usi ng the tru e range, however, a stop loss for a buy signal is calculated by subtracting, from the Both stop losses are based on the assumption that the market expressed a certain degree of pessimism on that extreme price day and to exceed it on a closing basis would constitute a deviation from its price char acter a nd consequently jeopardize the signal. I was convinced that the sequential method worked well when applied to daily data. Occasionally, I would study charts of a shorter time duration than low the difference between the closebyand the to low; stop loss for a sell signal is calculated adding the a high the difference between the high and the close. Just as in the previously discussed stop loss, a subse quent close must exceed the stop loss parameters. daily. I observed that the sequential technique worked well. Although I don't recommend its use on any time series other than daily, Figure 7.44 demonstrates the successful application of Sequential™ to a minute chart of the Japanese yen. 180 Summary Sequential™ 181 Source: CQG, Glenwood Springs, CO. Figure 7.44 Note the precis ion of the 9-13 sequence on identifying the low est price on the minute chart. Source: Logical Informa tion Machines, Inc. (LIM), Chicago, IL. Figure 7.43 To establish a stop loss, subtract the true range (Current day's high or close one day ago, whichever is greater - Current day's low or close one day ago, whichever is less). For a buy signal, sub tract this value from the lowest day of the setup/countdown period. For a stop loss after a sale, the reverse is done—i n this e xample, the difference between point A—close one day ago—and the high—point B—is added to point B. If price closes above this value, the stop is activated. For a buy stop, everything is reversed. Summary I have shared with you a technique that appears su perficially to be valuable; however, nothing is infal lible. Because I was the creator, it is difficult for me to admit the chance of failure. Consequently, it took Paul Tudor Jones and Peter Borish to challenge the method. Their creativity and mindset as floor traders prepared them to accept losses. They queried me re garding the subsequent activity of signals gone awry. Initiall y, I was in sulted by their unwillingness bot h to focus on the signals and to accept the fact that they would be successful. The price activity of the failures was most instructive. The ones that didn't work were really bad. In a sense, by concen trating on them alone, a trader could enjoy success. 182 Sequential™ You have been introduced to a powerful tool de signed to identify potential turning points in the mar kets. My experience proves that the technique has universal application to all markets, both foreign and domestic. Its features are its mechanized nature, its long-term perspecti ve, a nd its ability to fi ght the pre vailing trend. Because of a trader's passion to always be active in the markets, Sequential™ may be consid ered boring and unappealing. My experience has proven, however, that the most successful traders are at least cognizant of the big picture and take advan tage of it. Once again you may be asking yourself why I have given up th is valua ble' brain child o f mine for adoption. I do not view its publication in that context, however. I see it as an oppo rtunit y for you to do some babysitting for me while I pursue other market av enues of interest. In addition, in recent years I have introduced two notewo rthy enhancements to this ba sic approach which I have not discussed. I am confi dent that as you acquire proficiency with this technique, you will also be equipped to make similar improvements. Gaps A market psychologist will confirm the fact that emotions such as fear and greed play a significant part in determining price swings in the market. In fact, many market letter writers earn a living by measuring these sentiments and making recommendations based on their assessment of the collective opinions of the trading masses. There is a belief that in order to be successful in the mar ket, a trader must buy when everyone is selling and sell when everyone is buying. In general, this is a valid observation because the consensus is generally wrong. Simply stated, logic dictates that as price moves higher the number of potential buyers is depleted until there remains, figuratively speaking, only the "last buyer" left to buy; consequently, by default, price declines. Conversely, as price declines, the number of prospective sellers diminishes until there re mains only the "last selle r" left to sell; consequently, by de fault, price advances. Just consider the roles of the specialists and the floor traders. Both provide liquidity to the markets by selling strength and by weakness. At the same time, living they are al ways battling thebuying trend and they make a comfortable doing so. Although the most pronounced dislocations typically occur at the opening, these traders are afforded an advantage because they are responsible for setting the opening price. Opening price moves 184 Gaps Gaps 185 that exceed the previous day's close and fail to be filled by the close of tr adi ng are cal led eit her price gaps or price laps. Gaps occur when a particular day's price high or low fails to intersect the previous day's price high or low. Laps occur when the price high or low intersects the previous day's high or low but not the previous day's close (see Figures 8.1 and 8.2). Although much discussion in the past has been devoted to the topic of gaps, most of the r esea rch ha s been lacking. For the sake of ease of presentation, I will label both gaps and laps as gaps. My research Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 8.2 Points A, B, and C are downside laps: the price highs intersect the previous day's lows, but do not intersect the previous day's closes. Point D is a downside gap: the price high does not intersect the previous day's low. regarding gaps is unconventional and sheds light on a different perspective in which to view them. An old market adage teaches that all price gaps are filled. I'm cer ta in tha t a seller of Chrysler stock in the 1930s will dispute this claim, as will a seller of the Dow Jones Industrial Average in early 1975 (see Figure 8.3). To date, neither gap has been filled. These a re prime exa mples of how mark et folklore ha s Source: Logical Information Machines. Inc. (LIM). Chicago, IL. Figure 8.1 Points A, B, and C are upside gaps: the price lows do not overlap with the previous day's highs. Points E and F are considered upside laps: the price lows overlap the previous day's highs, but do not overlap the previous day's closes. been an d ha been those promoted as doctrine . All gaps accepted are not filled, ands even that are will some times leave the trader poorer by the time they are. In my studies of gaps, I have made some worthwhile ob servations. Once again I have elected to draw my own 186 Gaps Gaps 187 of time. Taken tog ether with t he followi ng conditions more substance and significance can be assigned to a gap. As was mentioned earlier, emotions are generally a major contributor to a price gap. When I looked to identify instances when this influence is either elimi nated or muted, I uncovered four: 1. When news is announced over a weekend or, ideally, over a long holiday weekend; 2. When news is either decide dly negative or pos  itive and a price gap occurs unexpectedly in the opposite direction; 3. When both a period of more than 11 days has elapsed since the gap occurred and when the close of day 8, day 9, or day 10 after the gap is the extreme close since the gap day; Source: Logical Information Machines, Inc. (LIM), Chicago. IL. Figure8.3 The lap of January 2, 1975 (A), andthegapof January 27, 1975 (B), were never filled, contradicting the notion that all gaps (laps) are filled. conclusions and ignore those misconceptions widely held as fact. Most price gaps are filled within a few days of their occurrence. Should you accept this as fact, however, I assure you that just when you initiate a trade , it wil l not work. To reduce t he likelihood of this event occurring, I observed those instances when price gaps were not filled for an extended period of time. What I found interesting was the fact that if gaps are precipitated by minor news events, they are usually filled qui ckly, even the same day. If the gap is associated with a major, unexpected announcement or with no news whatsoever, it is more likely to be le gitimate and remain unfilled for an extended period 4. When volume on the opening is light and con tinues afterward, suggesting a shortage of ei ther supply in an upside move or demand in a downside move. Over an extended period of time such as a week end, a tr ader is able to evaluate events more rationally than overnight. Removed from the office or news re leases, he is able to detach himself, temper his emo tions, and make cooler, more calculated decisions. For those reason s, Monday— or, in th e case of long week ends, Tuesday—gaps assume a special significance (see Figures 8.4 and 8.5). Furthermore, committee meetings are generally held on Mondays and at that time all the facts are dissected thoroughly, thus providing a sense of premeditation, deliberation, and forethought, rather of than to a decision. I am particularly aware gapemotion, price activity on Mondays and often review weekly charts; Monday gaps—on the first trading day of the week—are easily identif ied. By incorporating the other gap observations mentioned, 188 Gaps Gaps 189 Source: Logical Information Machines. Inc. (LIM), Chicago. IL. Source: Logical Infor mation Machi nes. Inc. (LIM), Chicago, IL. Figure 8.4 By using a weekly chart, Monday gaps (laps) are identified eas  ily. They occur at points A, B, C, D, E, and F. such as volume and time considerations, authenti city can be further confirmed and impact can be anticipated. If a part icul ar event ha s been antici pated for a pe riod of time an d a number of fals e s tar ts rega rding the release of this information have occurred prior to its actual release, then there exists a good chance that the news has already been discounted. On those rare occasions, the opening price may actually gap in the other direction and the gap remains unfilled (see Fig ure 8.6). Such an event would be totally unexpected and would carry significant implications—as well as opportunity—for the alert trader. Figure 8.5 By using a weekly chart, Mon day gaps (laps) are identifie d eas ily. They occur at points (A, B, C, D, E, and F ). Because the expectation is that most gaps are soon fi lled, I examined those in stance s when a gap re mained unfilled for an extended period of time. My conclusion was, essentially, if a gap is not filled within the ensuing 11 trading days after it appears, price usually continues to move in the direction of the gap until its momentum is exhausted. This observation is valid with the following exception—the closing price of day 8, day 9, or day 10 after the gap mu st be the most extreme close day since the gap day (see Figures 8.7 and 8.8). Volume is an important factor when analyzing gaps. Heavy volume on a gap opening is generally news-inspired and short-lived. On the other hand, my 19 0 Gaps Gaps 191 Source: Logical Information Machines . Inc. (LIM), Chicago, IL. Source: Logical Informati on Machines, Inc. (LIM), Chicago, IL. Figure 8.6 Becau se of the Middle-East o il disruption and the threat of U.S. war involvement, expectations were for the stock market to decline. Unex pectedly, the DJIA rallied sharply beginning with a upside price gap (A). Figure 8.7 The close on d ay 8 subsequent to the gap was an extreme close (A)—greater than all previous 7 closes—thus qualifying the gap and sug gesting the trend would continue. The same signal in reverse was indicated by the downside lap, which was confirmed 10 days later by an extreme close (B). The same indication was given 8, 9, and 10 days after the downside gap at point C. research su ggests tha t light volume gaps are durab le and appear as a thief in the night—unexpectedly, without a ny warni ng whatsoever . Their significanc e is often dismissed because they come with no fanfare or hoopla. In these cases, relatively speaking, the volume is not exceptional and the price change on the opening is nominal. However, when these particular ingredi ents are packaged together, the impact they supply is descriptions and lame explanations offered by conven tional analysts regarding specific gaps, as well as the classification of varieties such as "breakaway," "mid-," and "exhaustion" gaps, are lacking excuses and often without merit. Taken in the context of the subject mat ter presented in Chapters 1 and 2, which discuss TD powerful trader should alert to their currence indeed. s and beAprepared to takebeadvant age oc of them. Gaps have been relegated to the trading doghouse. Although their presence is obvious, no one has at tempte d to justify an d explain their existence. The Line and retracements, as sume breakouts a significance not revealedhowever, before. gaps By being aware of the characteristics of gaps, a trader can be better prepared and equipped to handle their implica tions and turn them to his profitable advantage. 192 Gaps Chapter Daily Range Projections Feb 1992 Mar Apr May Jun Source: Logical Information Machines, Inc. (LIM), Chicago. IL. Figure 8.8 The downside and upside gaps and laps were confirmed by ex treme closes 8, 9, and 10 days later. In the early 1980s, I appe ared regularly on Financ ial News Network prior to the daily opening and announced the projected price ranges for various markets. The formula presented below is an enhanced version of the one I used to make the se projections . It is the prod uct of many hours of research, and its interpretation is important in defining short-term price movement. My research has shown that tomorrow's price range is influ enced by the relationship between the current day's price close versus the current day's price open. There are three possible rela tionships between the close today and the open today: 1. The close today is less than the open today; 2. The close today is greater than the open today; 3. The close today is equal to the open today. If relationship 1 exists, I use the following formula to project the range for the following day: 194 Daily Range Projections Daily Range Projections (High today + Low today + Close today + Low today)/ 2 = X Tomorrow's projected hi gh = X — Today's low Although the performance results achieved by this formula in the past have been respectable be cause it has established realistic parameters for the following day's price activity, I make no guarantee that this performance will continue. For purposes of illustration, Table 9.1 shows the range prediction s for Soybeans March 1994. Tomorrow's projected low = X - Today's high If relationship 2 exists, I revise the formula as follows: (High today + Low today + Close today + High today}/ 2 = X Tomorrow's projected h igh = X — Today's low Table 9.1 Soybeans March 1994 t Tomorrow's projected low = X - Today's high Open If relationship 3 occurs, I make the following ad ju st me nt s: (High today + Low today + Close today + Close today)/ 2 = X Tomorrow's projected high = X - Today's low Tomorrow's projected low = X - Today's high These values merely provide a benchmark for the ensuing day 's price activity. I recommend th at the fig  ures be use d as foll ows: If price opens withi n the pro jec ted pric e r an ge an d you ar e a d ay t rad er, an tic ipa te resistance above the projected high and expect sup port at the projected low. More importantly, should price open outside the projected range—above the projected high or below the projected low—the supplydemand balance has shifted significantly enough to imply that the short-term price trend will continue in the direction of the openin g breakou t. Two options ex ist for the short-t erm trad er if such a brea kout occu rs: 1. Ignore the projected ranges for the day; 2. Adjust the value for the projected low to just below the projected high in the case of an up side breakout; conversely, revise the value for 195 the projected high to just above the projected low in the case of a downside breakout. 1/26/94 1/27 1/28 1/31 2/1 2/2 704.0 694.5 683.75 686.5 *684.5 Actual Low High 704.5 696.0 687.25 690.75 684.75 693.5 683.0 681.5 686.0 675.0 Close 700.25 696.5 683.75 686.75 687.0 683.5 Projected High Low 700.5 685.75 690.0 691.25 *Open below projected low—revise projected high to level of srcinal projected low. 689.5 677.0 684.25 686.5 Rate of Change On numerous occasions, I've been approached by college students who wanted to know what courses would best prepare them for a career in the stock market or the futures market. Recalling my own educationa l exper ience—li beral arts, study abroad, gra duate school not to follow in of business, and law school—I invariably respond my footsteps. In the trading profession, what is most obvious is of ten most obviously wrong. I grant that fundamentals or the percep tion of them dictate the long-term trend in markets; however, over shorte r time periods, the recognition of these fundam ental develop  ments might be ignored or overlooked. As a result, the price of a security may either remain dormant or contradict reason and logic. Effective market timing techniques help alert the user to the appropriate times when the price may be disposed to respond favor ably or unfavorably. The identification of these precise points in time is accomplished by measuring both supply/demand and mar ket sentiment. My techniques concentrate on basic economics and mass psy chology. Consequently, I believe an unders tand ing of these two ar eas of knowledge is vital to success—the former for purpose of measuring supply and demand and the latter for evaluating the emotionalism of the market. ) 198 Rate of Change Rate of Change 199 Of the highly educated professors you know, how many have been successful traders? I woul d venture to guess: not very many. Their lack of success is no re flection on their intelligence; over the long run, their fundamental expectatio ns can make them successful investors. Short-term trading, however, is a full-time profes sion, and ma rke ts do not always operate ration ally. In fact, my experience suggests that an inverse correlation exists between education and short-term trading success. Much of the information taught in business school is of a fundamental nature and does not address the key emotions dictating short-term price movement—fear and greed. Markets are effi cient, and once information is released the discount process begins immediately defining the impact of for a weekly perspective, the chart is updated weekly; and for a daily perspective, it can be updated daily. In each instance, however, the close is compared with the close one year earlier. One major benefit is derived from this approach. If the market is in an oversold condition, buys based on another shorter-term system can be generated. Conversely, if the market is in an overbought condi tion, sells based on another shorter-term system can be generated. In any case, even as an indicator unto it self, this relationship measures the level of emotion associated with price mov es and visually displays how similar movements have evolved in the past. The de gree of advance or of decline can be evaluated, a nd the extent of movement within overbought/oversold zones this news. Inbecause many of instances, movements exaggerated dynamicsprice outside the arenaare of fundamental analysis, such as stop losses, systems' trading signals, margin calls, and so on. Conse quently, the immediate price movement may contra dict all logical expectations. Figuratively speaking, I have often said that price continues to advance until the last buyer has bought, and price continues to de cline until the last seller has sold, and that this se quence of events accounts for perceived illogical market responses. can be examined and compa red. In fact, price activity recorded a year ago dictates current price movement by defining extreme parameters associated with his torical turning points. Figures 10.1 through 10.15 are examples of charts used for this type of comparative analysis. Many years ago, I experimented with a price com parison that incorporated market timing and market sentiment. For the most part, I applied this approach to major market indexes and to futures markets, rather than individual stocks, because the likelihood of price declining to zero was a real possibility with a stock. Specifically, I divided the current price of the se curity by the price of the security one year earlier. Other time periods may work better, depending on the market, but I wanted to cover the same time periods for all markets. From the chart prepared for each market, I can design overbought/oversold bands that have histori cally defined areas of low-risk buy and sell levels. For a monthly perspective, the chart is updated monthly; 20 0 Rate of Change Rate of Chan ge Bar of Cocoa Rate of Change of Cocoa Bar of Coffee Rate of Change of Coffee Source: Logical Informati on Machines, Inc. (LIM), Chicago, IL. Source: Logical Information Machines, Inc. (LIM), Chicago, IL Figure 10.1 One-year rate of change, cocoa mont hly. Figure 10 .2 . Six-month rate of change, coffee monthly. 201 20 2 Rate of Change Rate of Change Bar of Copper Bar of Corn Rate of Change of Copper Rate of Change of Corn Source: Logical Information Machines, Inc. (LIM). Chicago, IL. Figure 10 .3 Six-month rate of change, copper mon thly. Source: Logical Inform ation Machines, Inc. (LIM), Chicago, IL. Figure 10.4 One-year rate of change, corn monthly. 20 3 204 Rate of Chang e Bar of Crude Oil Bar of DJIA 40.00 36.00 32.00 28.00 24.00 20.00 16.00 12.00 Rate of change of crude Oil _Rate- of Change of.DJIA Source: Logical Inform ation Machines, Inc. (LIM), Chicago, IL . Source: Logical Informati on Machines, Inc. (LIM), Chicago, IL. Figure 10.5 One-year rate of change, crude oil monthly. Figure 10. 6 One-year rate of change, DJ1A monthly. 206 Rate of Change Rate of Change Bar of DJTA Ba r of HWP Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Source : Logical Informa tion Machines, Inc. (LIM), Chicago, Figure 10.7 One-year rate of change, DJTA monthly. Figure 1 0.8 IL. Six-month rate of change, Hewlett-Packard monthly. 207 20 8 Rate of Change Rate of Change Rate of Change of MCD 20 9 Bar of MMM Bar of MCD 1 _Q Rate of Change of MM M Source: Logical Infor mation Machines , Inc. (LIM), Chicago, IL. Source: Logical Information Machines, I nc. (LIM), Chicago, IL . Figure 10. 9 Figure 10.10 One-year rate of change , Minnesota Mining Corporation monthly. Six-month rate of change, McDonald 's monthly. 21 0 Rate of Change Rate of Change Bar of Silver Rate of Change of MO Rate of Change of Silver Source; Logical Informat ion Machi nes, Inc. (UM), Chicago, IL . Source: Logical Information Machines, Inc. (LIM), Chicago, IL. Figure 10 .11 Figure 10 .12 Six-month rate of change, Philip Morris Corporation. Six-month rate of change, silver monthly. 21 1 21 2 Rate of Change Bar of Soybean Rate of Change Bar of SPX Rate, of Change, of SPX Rate of Change of Soybean 1956 1962 1968 1974 1980 Source: Logical Information Machines , Inc. (LIM), Chicago, IL. Figure 10.1 4 Source: Logical Information Machines. Inc. (LIM), Chicago, IL. Figure 10.13 One-year rate of change, soybean monthly. Six-month rate o f change, cash S& P monthly. 1986 21 3 21 4 Rate of Change Chap 119.0000 112.0000 105.0000 98.0000 91.0000 84.0000 17.0000 70.0000 63.0000 56.0000 Equities Rate of Change of US Source:Logical Information Machines, Inc.(LIM), Chicago, IL. Figure 10.15 Six-month rate of change, U.S. T-bonds monthly. Throughout this book, I have presented numerous methods and strategies designe d to decode the market puzzle. My experience sug gests that these techniques can be applied to all markets equally well. Whether your specific interest lies in the area of cash cur rency, commodity, fixed income, or stock trading, the applications and the results should be comparable. Because of the variations in market composition, however, composite indexes may vary from one market to another. In this chapter, I intend to highlight various tradin g strategies I have created and have applied to both individual stocks and the stock market. New Issues—Initial Public Offerings For many years, I have observed a similar trading pattern for most newly issuedofstocks. I have beenbecause particularly trad ing activity these new issues I was sensitive in chargetoofthe trading these stocks when I first entered the investment business over 23 years ago. Although nothing in the investment business is for cer tain, and most techniques, following a period during which they 216 Equities New Issues—Initial Public Offerings 21 7 work exceptionally well, undergo a period of erosion and disinterest, this method seems to have success fully withstood the rigors of various marke t environ ments and of time. Whether the climate for public offerings is hot or cold, this technique on balance seems to work, although the degree of success seems to be influenced somewhat by the overall market. Typically, once a new issue begins to trade, I look for the following charac teristics regardles s of the price at which the public offering is completed. Often, price will either advance or at least move sideways for a few days. The reason for this price movement is that the selling syndicate that initially offered the shares to the public fo r sale sup ports th e price of the stock for a period of time. If the srcinal offering is priced conser of stock the underwriter(s) will place with investors, the volume and dollar-weighted volume for each day since the offering, and so on. This is not to say that the price activity will not conform to the one I antici pated if these items are lacking; rather, I use these cri teria to market -time entrie s, as well as to reinf orce my expectations. Once I have infor mation regarding the n umber of shares offered for sale, I attempt to learn how much stock the underwriter has placed in the hands of buyers. Underwriters often attempt to place the stock in strong hands—in other words, with buyers who ar e prepared to hold the stock for a period of time rather than liquidate once trading begins. They accomplish this by den ying commissions to the broker if the stock vatively, the syndicate's price suppo rt activity is not as critical. However, if the und erwrite r attemp ts to raise the ma ximum amoun t for the selling company, it may stretch the market to the point where prospective buyers believe it to be too expensive; price then de clines. In any case, the syndicate usually is able to muster enough buyin g to suppo rt the offer ing price. In fact, I have heard of some instances in which the un derwriter has insured the absence of supply by failing to compensate its brokers with commissions when their clients "flipped" a new issue—bought the offering and immediately liquidated once tradi ng began, which forced the syndicate to buy it at the syndicate bid. Whether the stock's price remains above the pub lic market offering price or not, there is a tendency for price to retreat or move sideways for two to four weeks after the first two to three days of trading. Then, at about the time when most of the people srcinally in terested in the offering become distracted, a new, more subdued surge in buying generally appears. To finetune the arrival of this secondary interest, I often re view specific items regarding the offering. I make an is flipped (sold immediately) at a loss. Generally, the underw riter is obligat ed to supp ort the stock at th e of fering price level for a period of time, and may not want to increase its inventory. Consequently, I make the assumption that the stock held by the underwrit ers and their customers should not be a factor in the market for a period of time. Next, I calculate the bal ance of stock of fered by the other syndicate m embers. With that figure in mind, I observe the trad ing volume for the ensuing d ays. As a rule of thumb, once the syn dicate members' stock has been tur ned over two times, the upside move should resume. Other items, such a s the price of the stock and the exchange on which it will be listed, are incidental fac tors that should also be considered. Many institutions are precluded from buying a stock that does not ap pear on an approved list. Frequently, a stock cannot be included on the approved list if it is not priced above $10. Most reputable new issues are traded on NAS DAQ, with a few exceptions that qualify for the listed exchanges. Margin requirements associated with these exchanges are important considerations as well. effort obtain information the size of publicto offering, the name of regarding the underwriter(s), thethe size of the selling syndicate, the number of market makers in the stock if on National Association of Securities Dealers Automated Quotations (NASDAQ), the amount factors, such the lifting of the often "quietrevi pe riod"Other and the removal of as major restrictions, talize interest in the stock. Furthermore, a primary consideration is the subject of supply. As mentioned in Chapter 1, many investors who own a stock at a loss 21 8 Equities will sell once they break even. In the case of a recent new issue, there is no supply because there are no buyers with a loss. The various factors I review when I consider a new issue should not be confused with the basic tend ency inherent in stocks offered in the new issue aftermarket. A noticeable pattern to advance appears three to five weeks after the offer ing. I suggest th at, in order to be alerted to potential trading candidates, you subscribe to a chart service such as O'Neill's Daily Graphs, which monitors the daily price history of many of these stocks just after public trading be gins to follow this price behavior. Buy-Outs I have been in the inve stment busin ess for a long enough time to witness all the fads and market concepts that are imaginable. The era of corporate takeovers was a thrill fo r me. Fortunately, the supply -demand models I had created were installed and were being success fully used prior to the advent of this period. Initially, my work would alert me to those situations in which aggressive buying was taking place. Most technicians are parasites and require no fundamental justifica tion for their market activities, so I assumed that pos itive fundamental developments had precipitated this demand. I was soon to learn that a pattern had devel oped that correctly predicted pending buy-outs. At last count, between 1978 and 1982, more than 32 ac quisitions were correctly forecast. I even had the gall to notify corporate presidents and announce that their companies were being acquired. In fact, I was de scribed by one journalist as the "grim reaper." The techniques are described in Chapters 5 and 7, which discuss accumulation/distribution and Sequential© respectively. Both methods combined were sufficiently sensitive to identify these opportunities. In this chap ter, my goal is to share information I acquired from ex perience to further validate acquisition candidates. Having played basketball, I was never satisfied with the easy lay-up. When I was put in a lay-up Buy-Outs 219 situation, I often passed the ball or made an effort to score with a mor e difficult shot. The sam e strategy a p plied to inside information and potentia l buy-outs. Not only did I pref er to make the process a challenge bu t I also was confid ent that by the time pending rum ors of buy-outs were widespread and nothing was an nounced formally, the rumors were more than likely bogus. I tried to explain this to others, but they were unwilling to listen. I did much research to convince them with logic and with examples. In the early 1970s, there were occasional buy-out rumors. Most legitimate instances demonstrated a similar pattern. I noticed tha t a surge in price volu me was followed by a respite period of typically just over six months. I came to respect this pattern; under the tax laws at that time, long-term capital gain s required a holding period of six months. Furthermore, the gov ernmental agencies were not actively involved in pros ecuting traders for insider activity. Consequently, I was always aware of rumo rs from reliable sources and I advised others that, typically, the official announce ment was more tha n likely fo rthcoming much later be cause the insiders were probably still in the process of accumulating stock personally—even though I never capitalized on this fact myself. At the same time, the acquiring company was careful not to accumulate more than the maximum allowable percentage of the shares outstanding before the government required a formal acquisition announcement. After the tax law was changed to abolish the sixmonth holding period requirement, other factors that served to confirm rumors of a buy-out were tested and were applied successfully. My belief was that once the rumors were received by the lowest common trading denominator—the public—and still nothing was for mally released, more than likely they were nothing more than rumors. To prove this fact to others, I con ducted the the following exercise. First, I Next, researched confirmed total shares outstanding. I multiand plie d the number of shares ou tstandin g by 5 percent to arrive at a benchmark. In turn, I multiplied this figure by a f actor of 5. B ecause the Securities a nd Ex change Commission (SEC) requires any shareholder 220 Equities accumulating in excess of 5 percent of a company to divulge this information, I assumed that for every share a potential buyer accumulated, four shares were being bought by others. Were a rumor proven to be fact, then the critical 5 percent ownership would most likely be completed by the time 25 perce nt of the shares outstanding had been traded. This particular filter process served me well in convincing others to avoid disasters. Another important observation I have made throughout the ye ars relates to the price activity dis played by a stock which has been "put into acqui sition play" by the release of a statement from the acquiring company. Specifically, once the announce ment is made and it is a cash—as opposed to a stock— purc hase offer and the price of the share s immediately trade at or above the acquisition price, generally ei ther a higher price is bid by the suitor, additional buyers appear, or, in any event, the deal is easily ex pected to be consummated. New Listing on Exchange Once a stock is listed on an exchange or added to an in dex, the potential for additional interest is enhanced considerably. Index funds are required to include in their portfolio s all components in various indexes, and margin requirements are often more attractive once exchange listing is accomplished. For these reasons, the po tential audience is often increased significantly. It is not uncommon to witness a price advance even prior to listing, in anticipation of this tendency. Fur thermore, many committees of large investment com panies restrict investment to only listed stocks—and then only stocks priced in excess of $10. Because of the criteria use required for listing approval, these large in vestors the listing process and the active require ments to remain listed as additional safeguards to ensure t hat they are investing prudently. The reverse of this phenomenon occurs when a stock is delisted. Heavy liquidation of delisted stocks, together with th e New Highs-New Lows 22 1 prospects of the company itself failing, are legitimate concerns that are to be respected and expected. I noticed the same tendency back in the early 1970s, when exchange-listed stock calls were first introduced. It was almost a foregone conclusion that as soon as a call was listed, the underlying security would advance. This pattern was dominant for an ex tended period of time until the exchange-listed puts were introduced and prices for the underlying stock declined for a short period of time. Unfortunately, this tendency was short-lived. In any case, I remain vigi lant to observe the vagaries associated with the intro duction of any new product in order to identify any inclination for the pattern to repeat itself. New Highs-New Lows Newcomers to a race track can always be expected to bet on the long shots. Typically, those bets have odds over 50-to-1 and almost never win. The smart money— the sophisticated gamblers—conduct their research and bet their money prudently and realistically; any expectations of a long -shot winner are left to inexpe ri enced gamblers. A long shot does occasionally win, amid as many bells and whistles a s a big slot machine winner will hea r in Las Veg as, but th at outcome is the exception. The same concept applies to the stock mar ket. Invariably, inexperienced traders like to focus on yesterday's market winners. History has proven that it is the exception indeed for a strong stock or industry performer in one bull market to repeat its preeminence in a succeeding bull move. Generally, once it has be come a fal len angel, it takes a numb er of mark et cycles to recover and lead once again. Unsophisticated traders (and some experienced traders) ignore this fact and of ten become trapped in th ese losing propositions. Common sense dictates that, as price declines more and more, owners of a stock incur losses. For the stock to rally significantly to new highs, all t he supply created by prematu re buyers on th e way down must be overcome. How many times have you entered a trade Equities only to see price move immediately against you, and then said to yourself that once you break even you will liquidate? Either these buyers must hold their stock positions and not liquidate, or their supply must be absorbed before price can advance. If a stock is mak ing a series of new highs in price, there are no un happy buyers with losses. Thus, the expectation of liquidation once the trader breaks even is gone. Con ceptually, the argument of overhead supply does not exist. My experience of being a stock scavenger was short-l ived, once I viewed the prospects in the c ontext of overhead supply. My research proved that stocks making new highs during an overall flat market were candidates for purchase because they were able to defy the laws of gravity displayed by the market in dexes. In fact, generally, they were leaders in the m ar ket during any period of strength. Conversely, during a sideways market prior to decline, those stocks recording new lows were the leaders on the downside in any market selloff. Many years ago, I took my research regar ding 52week new highs-new low s and applied a technique t hat assigned a stock's relati ve position versus its 52-week high or low . For example, instead of ju st relying on th e list of new highs-new lows as they appeared in the newspaper, I wanted to know precisely how close a given stock, presently at neither a new high nor a new low, was to recording one. Often, the proximity of a stock to recording a new high or a new low is camou flaged. An index I created, the TD New High/Low In dex, provided me with a benchmark whereby I could confirm expectations of price breakouts either upside or downside. The index is constructed by dividing the 52-week price movement of a stock by 10, and then ranking the stock on that particular day. If, for exam ple, price records a close today within 10 p ercent of its 52-week high, then a rating of 10 is assigned to the stock. Conversely, if price close of less than percent of its 52-week high,records then a arating 1 is as90 signed to the stock. If the price closes 50 percent less tha n its 52-week high, then a ratin g of 5 is assigned to the stock. Next, I calculate a cu mulative value and plot Advance-Decline 223 this index beneath the price action of a market index to validate price moves and trends and to determine the durability and substance of a trend. This method of evaluating the relative price close versus the price range of the previous year and th en calcu lating a com posite index (TD New High/Low Index) to validate overal l mark et moves is a valuable contribution to the library of market indicators. Once again, a basic, widely accepted indicator—new high-new low—is en hance d to create a more complete market indicator. All it took was a little imagination and some creativity. The other indicators described below are designed to improve on those commonly used by most stock traders . I believe the enha nceme nts I have introduced, as well as the integration of the various approaches into a composite, yield benefits that greatly improve the potential of trading success. Turbo-Charged Indicators As is apparent throughou t the book, many of the ideas I have presented are improvements on the techniques employed by most traders. My personality is such th at I have never been content to accept what everyone else does. I want explanations and logic to substantiate what I do. In most instances, my research confirmed that the widely followed construction and the inter pretation of indicators had some validit y, but I wa nted to exploit them further and make them more effective and valuable with my adaptations. I was comfortable and confident knowing that no one else would be us ing anything similar unless I shared these indicators with them. Advance-Decline Most traders are familiar with advance-decline models. Conventional methods usually run a cumulative index Equities of net advanc es-declines plotted daily be low a particu lar market index. For over 20 years, I calculated a 5day average by summing the net advances-declines and then dividing by 5.1 also calculated a 13-day aver age by summing the net advances-declines and then dividing by 13. I established overbought/oversold boundaries for these averages and compared them to a series of other relationships (described below) to ar rive at ideal buy or sell opportunities. Generally, I would expect to see the 5-day figure exceed ±450 on the sa me day that the 13-day figure exc eeds ±25 0. Advance-Decline  I perform similar calculations of 5- and 13-day averag es for the Dow Jone s Indus tria l Average 30 com ponents. Generally, readings of ±14.0 and ±5.0 for the same-day readings for both the 5-day and the 13day averages coincide with turning points. Next, I divide the total advances by the total de clines for each day. I sum these values and average them for 5 days and for 13 days. The 5-day figure would have to exceed 1.95, and the 13-day calculation would have to exceed 1.70 on the same day. Con versely, the 5-day figure would have to be below .65 and the 13-day calculation below .95. In turn, these levels would be incorporated into the master model to identify ideal buy and sell entry points. The next comparison in my overbought/oversold matrix is the ratio of advancing issues/total issues trad ed. On a 5-day and a 13-day bas is, for an ideal oversold reading, I like to see the 5-day below 30 and the 13-day below 35. Conversely, for an ideal over bought reading, I like to see the 5-day above 50 and the 13-day above 45. Integrated into the model con structed by determining ideal overbought/oversold bands for the other indicators, this measure confirms high-risk and low-risk buy and sell zones. Most people are familiar with the trend index (TRIN) developed by Richard Arms and found on most quota tion term ina ls. This index divides the ratio of ad vancing to declining issues by the ratio of upside vol ume to downside volume. I recommend averaging the se daily value s over 5-day and 13 -day perio ds. If the 5- day value is above 1.35 and th e 13-day value is 225 above 1.20 , or the 5-day value is below .75 and the 13day value is below .85 on the same day, and the other relationships in the model confirm, low-risk entry lev els can be established. The last compo nent of the ma rke t model is a ratio of the cu rre nt day's Dow Jone s Indust rial s Close di vided by the close of the Dow Jones Industrials Close 55 tra din g days before. If the r atio is below .89 or above 1.13, usually turning points are identified. Used in conjunction with the other indicators de scribe d above, a package of indi cator s with a re spectable track record history is created. I created this overbought/oversold matrix over 20 yea rs ago. I entered the st atist ics myself daily. Today , computers simplify this task markedly. I present be low a sample of the sta tisti cs I accumu lat e daily. Date Total Issues Traded Advance (Adv) Decline (Dec) Net 5-Day 13-Day Advance Total 5-Day 13-Day Advance Decline 5-Day 13-Day Net DJIA 5-Day 13-Day TRIN 5-Day 13-Day DJIA High Low Close Momentum Close Toda y Close 89 D ays Ago S&P High Low Close Momentum Close Today Close 89 Days Ago Chapter Options When I first entered the investment business, the Chicago Board of Options Exchang e (CBOE) did not exist. Options wer e un derwrit ten by brokers and were traded over the counter. Since that time, how ever, much literature has been written regarding the topic of op tions. Unfortunately, most information concentrates on procedure and valuation studies. Although some attention is devoted to meth ods that evaluate market sentiment and that indicate market direc tion, the extent of this information is woefully incomplete. All the observations I have made srcinate from personal experiences ac quired as a result of numerous forays in the options markets and are not found in textbooks. As I have stated repeatedly throughout this book, "sweating out" personal trades seems to make a trader more alert to p otential pitfal ls; it also contribu tes to indelibly fixat ing on his memory various strategies and opportunities. I will de scribe my techniques and rules, and, I hope, impart some wisdom that will reduce the likelihood of your trading failure. These tech niques have than application to both equity and futures options.that af Rather recite numerous incidents and episodes fected my options trading life, I will share with you the lessons learned. They were acquired as a result of being taught by the "ultimate market teacher"—trading losses. Psychologists have said 228 Options that many traders possess an unconscious desire to lose in their investments. I am not one of those indi viduals. I have looked on trading losses, however, as the tuition cost required to be educated in successful market trading. This may sound trite, but I learned from experience. Had someone else offered this infor mation to me years ago, I would have applied it and would have avoided frustration and heavy market losses. Although I was hungr y and my ap petite for in formation was voracious, nothing existed to satisfy it. It has often been said that the only winner in the options game is the writer. Studies have shown that over 80 percent of option traders lose money. When the listed option markets opened, lack of sophistication characterized both the writers and the buyers of the options. The learning curve for the buyers, however, was longer than that for the writers, because little lit erature was devoted to their plight of trading failure. More than likely, this was a result of the fact that whereas the writers—sellers—were predominantly institutions and floor traders, the buyers were small investors who were naive and did not possess the infor mation and resources the writers did to eliminate these inadequacies. Emotions and expectations play an important role in options pricing. Strip away these human feel ings and the game of options trading becomes much simpler. Many models, developed to ascertain fair val ues, have been employed by writers for some time and, by using both computers and mechanized strategies, the emotional component has been effectively elimi nated and replaced with discipline. My goal was to de velop a suitable set of rules for the buyer. Through experience, I accomplished this goal and created a list that is readily accessible anytime I venture into this risky market. Foremost on my mind is the fact that I must control my emotions and must ignore the emo tionalism of all my market counterparts who are buy ing at the same time. My experience suggests that observation of the following rules offers a chance that a trade can turn into a profitable experience. Specifi cally, I adhere to the following: Options 229 1. Only purchase a call option when the overall market is down in price versus the previous day's close; 2. Only purch ase a call option when the underly ing indust ry group is down in price versus the previous day's close; 3. Only purchase a call option when the call is down in price versus the previous day's close. These rules have served me well and, with the excep tion of replacing the requirement of a down close ver sus the previous day's close with an up close versus the previous theas rules forOptions purchasing put is options haveday's beenclose, defined well. trading difficult enough without allowing your emotions to in terfere as well. Together with the simple mathematical comparisons and models presented below, this list should get you on the road to success. For years, traders have used a simple options ra tio to identif y sentimen t extremes th at coincided with price turning points in the underlying securities. Al though pra ctitioners have exacted some degree of suc cess in forecasting price reversals, the results are spotty. Simply, what they do is divide overall put vol ume by overall call volume. My research suggests that approach is deficient f or a numb er of reasons: 1. The assumption is made that for each expira tion date and price, there are puts and calls listed; 2. No dollar-weighted adjustment whatsoever is made to the volume statistic s; 3. tion No considerati is given to the interplay of op volume andonopen interest. These are critical items that must be addressed to properly assess market sentiment. 230 Options Options 231 Initially, when options were listed on the ex changes, they were limited to calls. Slowly, puts were introduced, but had a trader calculated just the basic put-call ratio, it would certainly have distorted and skewed the results in favor of call volume. It makes sense that not only is the number of op tions traded im porta nt but so also is the price of these calls and puts. Why should the impact of options priced at one-eighth of a dollar be the same as those priced at five dollars? Consequently, I devised my own ratio of calls to pu ts by multiplying the volume by the dollar value of each option. I called this ratio the TD Dollar-Weighted Option Ratio. The same band of over bought/oversold th at is applied to the conventional ap proach can be used. However, the results should be an options expert, and he explained that most option models are updated over the weekend. Typically, the time premium is recalculated at that time. Subsequ ent to this episode, I concluded that it was prudent to ini tiate a position on a Monday rather than late on a Fri day unless I anticipated that an impo rtant event would occur over the weekend and could significantly offset the time premium erosion. This is particularly appli cable to options expiring within a mon th's time. I have provided you with some of the methods I use to evaluate the sentiment associated with various options markets . Not only can yo u use th ese models to evaluate the relative attractiveness of option opportu nities, but you can translate this information into opinions regarding the underlying securities. I would more indicati ve of true sentiment. Another ingredient in th e option equation is often overlooked: the interplay between option volume and open interest. Every time an option is written, open in terest exp ands. By invokin g a method ology that incor porates option volume as a percentage of its open interest and in turn dollar-weighting these numbers, another important ratio is created that has predictive value. To get an appreciation for the relationship be tween volume and open interest, just examine what occurs when the volume in a particular futures mar ket—not options—on a particular day exceeds the open interest on that same day. In essence, what has transpired is that the ownership in that market has been turned over and, consequently, its personality and price characteristics are prone to change. On an other le vel, a similar situation occurs with options. My personal experience in trading options re vealed a situation that is not obvious until one is con fronted with its consequences. Specifically, just prior to the marke t's close on a Fri day, I purch ased some call options. No news announcements were released prior hope that you can apply these indicators, as well as the rule s for entry, to your advantage by becoming an offensive options t rader capable of seizing opport uni ties as they arise, rather than being defensively dis posed struggling to preserve capital. to the following Monday's reopening. In my fact, the un derlying security opened much higher. To dismay, the option's price opened lower than Friday's close and my purcha se price. To no avail, I attempted to resolve this perceived dislocation with logic. I consulted with Cha "Waldo" Patterns* In recent years, a fictional cartoon character named Waldo© has been popularized and promoted in books, posters, puzzles, and a game entitled "Where's Waldo?"®. His creator, Martin Handford, has made it a challenge to find Wa ldo in the midst of hundre ds of other cartoon figures. Camouflaged and hidden in the extreme recesses of these pictures is Waldo. His location is difficult to identify, but once shown or discovered it becomes obvious. This exercise re minded me of a similar process I have been involved in for years— identifying, on a chart, price patterns that are obscured and overwhelmed by the price activity surrounding them. Once I be came aware of what to look for, however, this process was simple. Consequently, I have labeled these chart relationships and patterns as Waldo patterns. Rather than get into a lengthy discussion re garding the ir genesis, I will mer ely highlight their existence an d un  derscore some of the observations I have made regarding their implications. Suffice it to say that I suggest you research these Waldo pat tern s to dete rmine whether they might play a role in your trading program. Whether you deal in equities, futures, or cash markets, these patterns should convey similar messages. *Copyright 1987 by Martin Handford, Little, Brown and Company. 23 4 "Waldo" Patterns "Waldo" Patte rns 23 5 When I started in the investment business, I was introduced to all the generally accepted market mod els, indicators, and techniques. It took approximately one year of hard work and heavy indoctrination by market technicians before I fully grasped the com monly used systems and approaches to market tim ing. What I learned always appeared good on paper, but to reproduce it and apply it was difficult. My solu tion was to create srcinal research regardless of the time and expense involved to accomplish it. What sur faced from this research project was the revelation that many of the interpretations assigned to widely followed and often quoted market patterns were just the opposite of what they should have bee n. This dis covery shattered the faith I had in conventional wis time has increased the selling pool. These short covering rallies are typically character ized by their steepness. Many traders like to relate current up side price action to a reference high recorded some time before. Richard Russell, of Dow Theory Forecasts, made a notable observa tion many years ago, which I follow to this day. At that time, many trade rs were fo cused on a Dow Jones Industrial Average peak recorded many months earlier. The market was advancing on exceptionally heavy vol ume. Contrary to popular belief, Russell ob served that (1) the price movement was running into heavy resistance because of the dom and forced me to conduct my own research and market timing investigation. I emerged from my selfeducation with the following principles, which I have adhered to ever since: large volume, and (2) more than likely, price would stall and reverse prior to any penetra tion of the previous highs. He was correct. I conducted my own research and concluded that an ideal situation arises when price ad vances and volume is light, suggesting a shortage prior to the price peak. Once the high is exceeded, I look for volume to inc reas e significantly, for two reasons. First, I expect to see short covering and stop loss buying to occur at and above the old price high. Sec ond, trend followers are likely to initiate posi tions precisely at that price high and above. Consequently, I prefer to see volume explode subsequent to a breakout above a previous high but not before. These same observations apply to price and volume when price de clines and the prospects for a previous low's being penetrated are being evaluated. 1. Most trade rs beli eve that increasi ng volume is an impo rtant companion of a genuine price advance. I concur in some instances on a price continuum but my research suggests that such a development is definitely not al ways the case. For instance, once price has formed a low, I prefer to see light volume be cause, typically, it suggests a shortage of supply. I defy traders to prove that they cor rectly purchased th e absolute price low . I as  sociate such claims with the man y celebrated fish tales. My experience shows that lows are made once the last seller has sold and, by default, price moves sideways to higher. Gen erally, when heavy upside volume occurs co incident with a new price low, it is of a short covering variety and consequently short lived. In fact, such a situation generally es tablishes a price vacuum in which price declines even faster once the decline resumes; this occurs because premature buying has depleted the buying reserve and at the same 2. Many traders believe that reversal bottoms and tops are important trading signals. I dis agree. to examine pri ce ch fu art s of mostI invite of the you actively tradedthe stocks and tures. Generally, price reversals—days in which a low exceeds the previous day's low to the downside and price closes higher, or days 23 6 "Waldo" Patterns in which price exceeds the previous day's high to the upside and price closes lower— are caused by short-term traders, and prices typically resume their trend once they are completed. Trading days in which price de clines and a down close versus the previous day is recorded, or trading days in which price advances and an up close versus the previous day is recorded, are more signifi cant and durable and are likely to occur at price bottoms and tops. There is one pattern that assumes the same significance as the one previously described. In that pattern, a reversal occurs and price closes greater than all previous four closes in the case of an up side reversal or closes less than all previous four closes in the case of a downside reversal. "Waldo" Patterns 23 7 between the previous day's close and low, conclusions regarding the following day's up  side potential can be determined. Specifi cally, if the difference is greater one day ago tha n two days ago, the prospects a re good f or a rally, provided that the following day's low does not violate the previous low. Conversely, once a short-term high is formed, by sub tracting the difference between the previous day's high and close, conclusions regarding the ensuing day's downside potential can be established. If the difference is greater one day ago than two days ago, the prospe cts for a decline are good, provided that the follow ing day's high does not violate the previous day's high. 4. Assume the lowest price recorded is 10 or more days ago and, prior to that low, all pre vious 10 days' lows were higher. Label this low day as the reference day. If the next 2 days are down closes and both record range lows beneath the reference day's close, a price low is likely being formed. To establish a price high, reverse the conditions. 7. A downtrend can be reversed once a price open or close is recorded that exceeds upside the price close four days before the most re cent TD Point Low. This must occur within four days of the TD Point Low. If two price gaps are recorded on two days up to and in cluding the first reversal close, the reversal is suspect. Conversely, an uptrend can be re versed once a price open or close is recorded that exceeds downside the price close four days before the most recent TD Point High. This must occur within four days of the TD Point High. If two price gaps are recorded on two days up to and including the first rever sal close, the reversal is suspect. 5. Generally , when price closes uncha nged ver sus the previous day's close, price continues to move higher if the previous day's close is up and to move lower if the previous day's 8. If the close 1 d ay ago is less tha n the close 5 days ago, and the close today is greater than all previous 7 days' hig hs but not all previous 11 days' highs, a short-term top is formed. 3. Typically, markets experience a consolida tion phase whenever a particular day's price range is more than two times the previous day's price range. This is more apparent if price has been trending for a period of time. close is down. 6. Once a short-term low is formed, by subtrac t ing the difference between that day's close and low and comparing it with the difference Conversely, daythe ago is more tha n the closeif5 the daysclose ago ,1and close today is less than all previous 7 days' lows but not all previous 11 days' lows, a short-term low is formed. 23 8 "Waldo" Patterns 9. Prepare fo r a potential change in market per sonality when volume levels on a particular day exceed—or surpass significantly on a percentage basis—historical relationships between volume-open interest levels. This could suggest the potential of an ensuing di rectional price change. 10. Most trend foll owers buy when price e xceeds upside all previous highs for a prescribed number of days, or sell when price exceeds downside all previous lows for a prescribed number of days. A common practice is to bu y once a high exceeds all previous 40 days' highs and to sell once a low exceeds all previ ous 40 days' lows. Many trend followers re main invested at all times. However, some prefer to tak e profits and p osition themselves neutral once price records a 20-day low or a 20-day high. Typically, trend following pro duces less tha n 35 percent winners; if portfo lio diversification is applied, the likelihood of participating in the trending markets is en hanced considerably. 11. A number of market analysts have observed seasona l price tendencies in the commodities markets. Numerous books have described a numbe r of these ob servations. What ha s been often overlooked, however, is the fact that there exists a propensity for stocks to exhibit a similar seasonal behavior. I pred ict more research will be conducted in th is are a. 12. Most trade rs apply the identical overbough t/ oversold decision rules to bull and bear mar kets. In an uptrending market, oversold readings occur only for short dominant. time and overbought readings are a more Conversely, in a do wntrending m arket , over bought readings appear for only a short time and oversold readings occur more often. The "Waldo" Patterns 23 9 overall trend of the market, or the market environment, is a critical factor to consider • when interpreting overbought/oversold os cillators. (See Chapter 3.) 13. Prior to the completion of my work regar ding trendlines (TO Points and Lines, see Chapter 1), I used a technique to perfect trendline breakout entries. Specifically, I would draw a trendline. If it was an up trendline, I would wait until downside penetration was ex hausted. Then I would use the low defined by the sell-off as my sell entry once a subse quent rally was completed. Conversely, I would trace a down trendline and await the completion of the upside penetration. Then I would identify the high defined by the ad vance as my buy ent ry once a subsequent de cline was completed. 14. Traders like to identify support and resist ance levels and often anticipate specific price activity to occur at these price levels. My ex perience suggests th at the concept of support and resistance does have application to stocks, but due to the high turnover in the futures markets, the principle has no appli cation except in intraday trading. 15. One factor overlooked by many traders when operating a market timing system is the sig nificance of specific days of the week, of the month, or of the year. I have conducted some research regarding this area of systems anal ysis and my findings suggest a more than ca sual influence. Conclusion No trading methodolgy is perfect, and the many methods presented in this book are no exceptions. The techniques I have shared with you have been developed and refined in the most intense and rigor ous laboratory conditions imaginable—the trading arena. As I have repeatedly emphasized throughout this book, necessity has truly been th e mother of inventi on in my trad ing life . Nothing instills the challenge of creativity more than to see personal trading profits erode, only to be replaced with a series of sizable losses. The goal of successful and profitable trading performance has been an elusive one for me to attain. In those rare instances when my tradin g juices an d adrenalin a re flowing most v igorous ly and my level of confidence is at its peak, I am most vulnerable. One particu lar example, comes to mind. On a summer day in 1989, ma ny of the mechanical systems I had designed and developed at Tudor gener ated signals concurrently. Peter Borish, my counterpart at the com pany, and I were exci ted about the prospects. What occurred tur ned out toasbeaone of my most humbling al experiences. be gan modestly profitable tradeprofession and evolved into a trip It into traders' hell—we recorded the largest one-day account drawdown in our trading history. Fortunately, Peter and I had installed money management contingencies that curtailed our losses. Once again, I 242 Conclusion emphasize the inevitability of encountering just such unexpected events in this busine ss. To reduce their im pact, I recommend lessons in money management sur vival. In add ition, I strongly suggest tha t you prepa re a psychological profile highlighting your strengths and weaknesses as a trader, t hus enabling you to more eas ily deal with trading disasters when they arise. I have attempted to provide you with the tools re quired to generate profitable trading signals. It is my hope that you will use my research and experience to design and to craft your own trading rules. Nothing would please me more than to read or to hear of a trader's success and accomplishments srcinating from some ideas presented in this book. I was not af forded this luxury and opportunity when I entered the business many years ago; I had to acquire most of my knowledge from personal trading experiences. This book is a product and a culmination of those many years of work in a profession in which I am con sumed. Larry Bird, th e former Boston Celtics' baske t ball superstar, now retired, remarked subsequent to signing his first post-college multimillion-dollar deal, "It's a good thing Celtics management didn't realize that I love this game so m uch tha t I woul d have played it for nothing." I reiterate precisely the sa me words re garding my profession. I am confident that should a similar passion for the investment business burn in your life, your trading future is destined to be both emotionally and financially fulfilling and rewarding. Index A c Accumulation: distribution, 109-127, 218 equation for, 116, 118-119 model, 124-127 Advance-declines, 221-225 Aggressiveness, degree of, 125 Alas ka Air Group, 150 Allied Chemical, 148 Americ an Int'l Group, 149 Amgen, 20 Arcs, retracement, 69-70 Arms, Richard, 224 Audience of book, 4 B Bally Manufacturing, 154 Benefits, 16-21 of buy and sell si gnals , 2 Best Buy Corp., 151 Borish, Peter, 181, 241 Bowmar calculator, 4 Breakout values: in TD Price Projectors, CAC 40, 152 Calculating price projections, 15 Calculation of rate of change, 27 Career in stock market, 197 Cash, S&P, 213 Caterpillar, 29 CBS, 189 Centering moving averages, 131 Chart analysis: approaches to, 2 benefits, 2 buy and sell signals, 2 evolution of, 2 market timing indicators, 2 Char t highs and l ows, 15 Chase Manhattan Bank, 178 Chicago Board of Options Exchange (CBOE), 227 Cocoa, 42, 75, 175, 200 Coffee, 12, 32, 33, 54, 144, British 24-25 Gilt, 30 British pound, 192 Buy and sell signals, 2 Buy-outs, 218-220 Buy/sell zones, 85-86 159,Joe, 201 101 Collins, Contradictory signals, 38-39 Contratrend rallies, 59 Copper, 71, 142,202 Index Corn, 157, 203 Cotton, 8, 158 Countdown, 153-154 Critical price, 61-64, 68, 70 Cross rates, 177 Crude Oil, 185, 204 Cycles, as market timing devices, 135-136 deficiencies of, 139 D Daily charts: reasons for, 9 Daily range projections, 193-195 Demand price: pivot points, 10-11, 12 Demarker Indicator, 92-93, 95-99 Deutsche Mark, 50, 63, 171, 177, 179 Digital Equipment, 64, 91 Distribution/Accumulation, 109-127 equation for, 116, 118-119 model for, 124- 127 Dow Jones Industrial Averag e (DJIA), 66-6 7, 88, 138, 161, 170, 186, 196, 205, 206 Downside targets, projecting, 35-36 Duration: degree of price moves, 78 key variable, 89 D-Wave analysis, 103-109 E Elliott, R. N., 60 Elliott wave principle, 101-104 Entry points: evaluating, 3 sequential, 164-172 Equities, 215-225 buy-outs, 218-220 new is sues, 215 -218 new lis ting, 220 -221 Exit points: evaluating, 3 sequential, 172-176 Exxon, 39 F False moves, 50 Fibonacci numbers, 60, 101-102, 121, 123 Financial News Network, 193 Financial Times Index, 43 Flip, 165-166, 172, 173, 174, 175, 176 Frost, Jack, 101 G Gap compensation formula, 120-121 Gaps, 14, 15, 16, 40, 183-192 muting, 187 volume, impo rtance of, 189-190 General Motors, 11 Germand Bund, 165 Gold, 47, 62, 137, 173 Golden mean, 60 H Hangseng Index, 174 Hecia Mng, 160 Hewlett Packard, 108, 207 Historical price activity, 7 Holding periods for trades, 106-108 Homestake Mining, 107 I IBM, 53, 114, 120, 125, 93, 112, 113, 117, 118, 119, 122, 123, 124, 126, 162, 191 Imbalances, effects of, 88-89 Indicators: market-timing, 2 Initial Public Offering (IPO), 215-218 Intensity of accumulation/ distribution, 116-117 International Flavors & Fragrances, 180 Intersection, 148-149 Intraday price breakouts, 48-59 J Minnesota Mining & Manufacturing, 44, 209 Moving averages, 129-133 questions about, 130 techniques, 131-133 Multiple trendlines, 8 N National Association of Securities Dealers Automated Quotations (NASDAQ), 216-217 New highs-new lows, 221-223 New Issues, 215-218 Japa nese yen, 17, 18, 46, 58, 146, 181 Jones, Paul, 181 New Listing, Nikkei, 6, 68,220-221 164 K o Kmart, 163 L Laps, 184-185 Limit moves: shortcomings of, 127 Lumber LB, 21 M Magnet price, 61, 64-65, 66, 71 Marion Melrell Dow, 168 Market environment, 90-92 Market sentiment, 229 Market timing: analysis, 48 approaches, 1, 85-87 benefits of,2 3 indicators, standardizing, 9 success, 3 McDonalds, 143, 208 Microsoft, 26, 70 O'Neil's daily graphs, 218 Options, 227-231 Orange juice, 19, 172 Overbought/oversold, 85-100, 238-239 Matrix, 224-225 Overbuying, 87-89 p Pacific Telesis, 176 Pattern of consistency, 8 Peak descending TD points, 11-12 Pepsi, 145 Philip Morris, 25, 153, 210 Pivot point: high, 15 low, 14 supply price, 10-11 Price activity, 10-11 Price advance: as demand exceeds supply, 6, 7 Price breakouts: validation, 48-59 Index Index Price exhaustion area, 93 Price decline: supply exceeds demand, 6 Price gaps, 1 5-17, 18, 19, 20,21 Price movement: based on supply and demand, 6 graphs, 6-7 symmetry, 23 Price objective, 21 Price points, for retracements, 61-62 Price projections, 22-24 calculations, 15 Reliability of demand lines, 17 Resistance/support levels, 64-65 Retracements, 59-84 arcs, 69-70 exceptions, 65-69 factors, 61 qualifiers, 71-77 ratios, 60-61 selection of price poi nts, 61 Risks, reducing, 41 RSI indicator, 88, 90, 9 1, 92 Rules for options trading, Pricecomparison, range activity 94 Price setup, 140-152 invalidation of, 158-159 Price trend: Figure 1.3, 8 graphing, 8 Proctor & Gamble, 74 Proportional divider, use of, 77 Public offerings, initial, 215-218 229 phase, 148 Runaway Russell, Richard, 235 Q Qualifiers: retracement, 71-77 R Range Expansion Index (REI), 92-95, 96, 97 Range projections, daily, 193-195 Rate of change, 197-214 calculation, 27 Reference points: intraday low, 27-28 Reference values: selection criteria, 78-84 REI macro, 98 S Safeguards, 1 Schlumberge r, 41 , 52 Sequential, 135-182, 218 countdown, 153-164 entry, 164-172 exit, 172-176 price setup, 140-152 Set-up, 140-152 Short-term static, eliminating, 94 Silver, 31, 76 Snapple, 13 Soybean oil, 28, 56 Soybeans, 22, 57, 73, 80, 92, 141, 155, 184, 195, 208 S&P 100, 23 S&P 500, 16, 37, 55, 69, 72, 82, 99, 105, 106, 156 Standardizing securities, 117-121 Stockastics, 86 Stop-loss, 176-180 Stop loss, use of, 41, 176-180 Sugar, 51 Summatio n of dally volume, 111 Supply-demand, 6 equilibrium equation, 12 models, 109 Supply price: pivot points, 10-11 Swiss Franc, 14, 36 Symmetr y of price movement, 23 Teledyne-Post, 77, 169 Texas Instruments, 45 TD Breakout Qualifiers, 48-58, 71 TD Demand Line, 11-12 TD Dollar Weighted Option Ratio, 230 TD Line breakout, 39-40 TD Lines, 6, 11-58, 67, 71, 74-75, 239 construction of, 12-13 higher magnitude, 42-47 rate of change, 27 TD Moving Average, 133 comparisons, 40 distinctions, 32-33 problems, 38-41 TD Price Points, 6-2 2, 15, 237, 239 demand points, 12 resetting, 12 selection of, 13-22 supply price, 10-12 TD Price Projectors, 22, 24-37, 40 Method 1, 24-25 Method 2, 25-34 34-3711-12 TDMethod Supply3,Line, TD Supply Points: descending, 13 identifying, 13 refinement, 13-15 Time, as a factor in overbuying, 89 Time series, Fibonacci, 60 Topix, 167 Trend Factors, 77-84 Trend index (TRIN), 224 Trend lines, 1-58 breakouts, 5, 39-40 defined, 5 Trend reversals, 59 Trend turning points, 10 Tribune, 166 True highs and lows, 15 Turning points identified, 100 u Uniformity, achieving, 53 United Airlines, 65 U.S. Treasury bonds, 10, 38, 49, 66, 79, 83, 90, 104, 136, 147, 188, 214 Uptrending market, 26 w Waldo patterns, 23 3-239 Wave analysis, 101-108 Williams, Larry, 113 X Xerox, 7 z Zones: buy/sell, 85 85 identifying, 247