Cycle Analytics for Traders: Advanced Technical Trading Concepts

John Ehlers is a widely respected expert on signal processing, so any book on this subject written by him should be welcomed. And I do indeed welcome this text. I have a strong background and interest in this subject, and I found “Cycle Analytics for Traders” to be a useful blend of theoretical foundation and practical implementations (mostly in TradeStation’s EasyLanguage).

On the other hand, the “for Traders” part of the title implies a certain target audience, and I can’t help but feel that most members of this particular audience will have some serious difficulties when they approach the book. The author makes an organizational choice that is less than ideal: he begins the book by diving directly into the Z-transform approach to filter design, a quite advanced mathematical concept normally taught well into a college-level course on signal processing. Indeed, Page 2 of the book hits the reader with a series of equations fearsome enough to make 99.9 percent of ‘traders’ roll their eyes and slam the book shut in dismay. And it only gets worse from there for an extended stretch. One must wade through graduate-level mathematics all the way to Page 40 before the first small EasyLanguage code sample appears. For those of us who’ve already grappled with this material in college, this is wonderful. He does an excellent job of condensing much of an entire semester of filtering theory into a few dozen tersely worded pages. But for his implied target audience, this approach is worse than worthless… it is so intimidating that few would have the fortitude to press on. Had I been given the opportunity to edit this book, I would have banished this material to an appendix, and instead begun with a more gentle introduction to signal processing applications relevant to traders.

The situation does eventually improve when Dr. Ehlers mostly abandons formidable mathematics and instead focuses on an intuitive approach to his chosen topics, accompanied by practical EasyLangauge code. In this regard, I might still criticize the fact that he goes to the opposite extreme and offers little or no theoretical justification for his techniques, nor does he offer examples of actual trading situations in which his techniques produce profits. Terse, rigorous summaries of the sort that filled the first 40 pages of the book would be good to have in an appendix that some of us would read and others could safely ignore. Also, because his title says “for Traders,” some examples of real-life trading payoffs would be better than vague promises that the techniques might be useful. Nonetheless, he has done an excellent job of choosing fascinating and potentially useful techniques. For mathematically unsophisticated readers who are not interested in anything beyond vague intuition and code that they can use to experiment with their own trading systems, most of this book is accessible.

Toward the end of Chapter Three the author does introduce one interesting topic, a low-pass filter of his own invention that he calls a SuperSmoother. Unfortunately, he provides neither rigorous theoretical justification for it, nor does he provide EasyLanguage code to implement it. The reader is left to trust the author’s idea and code it from a few minimally explained equations. Then, in Chapter Four, the author introduces another sort of low-pass filter which he calls a Decycler. EasyLanguage code for this filter is provided, but guidance for choosing between the two types of filter is minimal, a couple short sentences. Chapter Five introduces band-pass filters, though again with extremely complex mathematics. This topic does include EasyLanguage code as well as several illustrations using real market data, a welcome addition to the book. I will not comment further on the first five chapters, as with few exceptions they will be incomprehensible, and hence irrelevant, to the vast majority of readers.

The book begins to become interesting in Chapter Six, when concepts such as the fractal dimension and Hurst Exponent are discussed. EasyLanguage code and some market results are shown. But again, the explanations are for the most part terse and highly technical, with the author assuming that the reader has a strong mathematical background. Most traders will have to be content with picking up a few vague intuitive concepts.

Chapter Seven introduces yet another interesting author invention, the roofing filter. EasyLanguage code is provided, along with several example illustrations with real market data. Unfortunately for most readers, the explanations are highly technical, as in this typical example: “Since the roofing filter consists of the serial connection of a one-pole high-pass filter, a SuperSmoother, and another one-pole high-pass filter, it is clear the better formulation of the roofing filter would be the serial connection of a two-pole high-pass filter and the SuperSmoother.” Yeah. Obviously.

I’m always a sucker for pretty pictures involving data analysis, and Chapters Eight through Ten have these in spades. The author provides (very long and fussy) EasyLanguage code for producing colored plots of autocorrelation, periodograms, spectrograms and comb filter displays of market data. Lovely.

Chapter Eleven is arguably (and the author would likely argue) the first chapter in the book that is truly relevant to traders. It deals with concepts familiar to the trading community and presents techniques that would seem to have obvious utility. In this chapter he applies some of the concepts from earlier parts of the book to (hopefully) improve on several of the traditional workhorse indicators: RSI, Stochastics, and the Commodity Channel Index. As usual in this book, much of the supporting mathematics is terse and advanced, but at least he provides solid intuition and much EasyLanguage code that the reader can copy or download from his website. All that’s missing here is actual trading results comparing the original versions and his improved versions of the indicators. But of course readers should be able to run these tests themselves.

Chapters Twelve and Thirteen present other inventions of the author, a Sinewave indicator that he claims can have predictive power when the market is in a cyclic mode, and a Convolution indicator that he claims can identify market tops and bottoms. As with the prior chapter, he provides no rigorous real-life justification for these claims, but the ideas behind these indicators are undeniably intriguing.

Chapter Fourteen presents still another indicator involving a Hilbert Transform. Prior comments apply.

Chapter Fifteen discusses methods for transforming indicators to make them more usable. These methods are clearly important, if perhaps a bit superfluous for indicators that are well designed from the start.

My love for pretty pictures is triggered again in Chapter Sixteen, which discusses Swami charts, a nifty way to visualize indicators in the context of their history. His examples using RSI and Stochastics are excellent and will definitely appeal to most traders.

Chapter Seventeen concludes the book with a brief overview of some basic facts-of-life for the serious trader. This chapter mentions the need for out-of-sample testing, the danger of integrating stop-loss orders in a strategy, the utility of a simple bootstrap test for estimating the variation in possible net outcomes, and so forth. I got the feeling that this chapter was a last-minute add-on, something totally unrelated to the rest of the book, included only as a token obeisance to prudence. The material in this chapter is extremely important, but treated so lightly that it is all but useless.

In summary, there are two groups of readers who will find great value in this book. Readers with a strong mathematical and signal-processing background, as well as an interest in market trading, will find it to be a delightful summary of applied digital filter theory, with all of the most important equations laid out in a compact presentation, and with trading-oriented applications throughout. And readers who have little mathematical background but who are fluent in EasyLanguage or a similar trading platform language, and who are willing to experiment with techniques without knowing exactly what they are doing or why they are doing it, will also find this book fascinating. But this book is definitely not for the faint of heart. The mathematical presentations are fearsome, and will surely intimidate casual readers. Also, the book has a total lack of rigorous empirical justification for its techniques. An appendix containing statistically supported trading results with real market data would have gone a long way toward reassuring readers that the enormous effort of slogging through this text is worthwhile. Only serious readers need apply.

There are thousands of books covering the use of several technical analysis indicators. However, what is seldom discussed is why and how such indicators may achieve a particular trading objective. Moreover, many were developed when computing power was extremely limited, which is not the case today.

The latest book by John Ehlers addresses these glaring gaps in the field of technical analysis, drawing upon his extensive experience first in the world of digital signal processing – in many ways comparable to analyzing and processing capital markets data – and later as a trader and leading technical analyst.

For instance, many people use simple or exponential moving averages to smooth financial data and get a better sense of market direction. Mr Ehlers provides the theoretical background for their use and a host of innovative indicators which achieve the same objective – smoothness – but with much less cost – the lag relative to the actual data being smoothed. This in turn provides much more accurate and reliable trading signals.

Moreover, various conventional trending and oscillating indicators are contaminated by prior signals and noise, giving an inaccurate picture of market phase and direction. As a result, many authors have provided “fixes” by tweaking their trading rules. In contrast, Mr Ehlers identifies the root cause of this problem, which in itself is worth the price of the book several-fold, and proposes new indicators that accurately reflect current market conditions.

Digital signal processing is a complex topic, particularly when applied to technical analysis. But Mr Ehlers does a fine job in navigating the reader through the theory and its practical application. The code for every indicator, fruit of many hours of research and analysis, has been provided for everyone to see and use as they see fit.

I consider this book to be the seminal work of Mr Ehlers, in my opinion one of the best technical analysts alive today.

I’ve read all of Ehlers’ books. This one is by far my favorite. All the code runs well and is quite useful.

This is an enlightening book. And the code examples work. If you want to try some indicators that remove much of the ‘noise’ of the market, then this book is one you’ll want to have. A bit ‘mathy’ in places, but you don’t have to slog through it to make great use of the programming examples.

Having been trained in Electrical Engineering and Signals and Systems Theory, John does a great job borrowing and building filters which allow us to create non-lagging indicators to help market analysis. The only complaint I have is that usually he does a little better job on developing the theory. Non-the-less, these are mathematically accurate indicators which perform as advertised and are a great asset in the hands of someone who knows how to use them.

This is a good introduction to applying signal processing methods to stock trading with many clever techniques. However, the reader should _already_ be familiar with basic signal processing. A senior-level undergraduate EE course in signal processing will meet this prerequisite.

Great book….great free code. Warning: this book assumes you are an advanced trader and/or trading systems developer. Also…some of the indicators from Ehlers previous works are not included here as Ehlers only includes the indicators that matter – in his judgment. The book takes some time to get through…but is quite insightful for the technically adept reader.

In this book John Ehlers updates and significantly extends his prior research and insights on cycle analysis and stock market analysis using Digital Signal Processing (DSP) tools and concepts. Ehlers has spent many years doing very original research on cycle indicators for the market, notably the development of MESA (Maximum Entropy Spectral Analysis) to measure market cycles. Ehlers translates very complex material into understandable and usable indicators which are well illustrated in the book. Despite the sophistication of the tools and techniques, Ehler’s writing is extremely clear, yet concise, as he builds successively useful analysis tools.

About me: My (decades-ago) undergraduate degree (plus Masters coursework) is Electrical Engineering — where I was introduced to filter theory and z-transforms/difference equations. I like this edition much better than Dr. Ehlers’ older “Rocket Science for Traders” and “Cybernetic Analysis for Stocks and Futures”. However, it took me an introductory reading plus several follow-up readings/studyings before I felt I personally owned the concepts mentioned on pages x and xi (more important since I don’t have TradeStation). For my purposes, there’s chapters that could be in an appendix (since they don’t seem to make a major contribution to the author’s most-recommended versions of the software (e.g., Chapters 9, 10, 14 and perhaps 13). I think the flow of the book could be a lot better. In summary, if you’re attracted to Dr. Ehlers work, then I recommend this book. I also found getting the white papers or original article publications to be useful since there’s no color graphs (for heat maps, etc.) in the book. If you have TradeStation, then you’re in an better place than I am — you can just download the code using the access card that comes with the book.