Customer comments on this selection.
 Definitely worth reading, but some significant flaws...
I had the opportunity to hear the author discuss his book at a Market Technician event at the Bloomberg building in NYC just over 2 years ago.
br /Even though I have some disputes regarding his methodologies, reading his work was time well spent.
br /
br /The first half of the book was heavy on philosophy and psychology, driving home the idea that the "scientific method" (whatever that is), is our only guard against self delusion.
br /
br /As I am familiar with both philosophy of science, as well as psychology, it seemed to me as if he was stating the obvious, but I suspect for most traders, it is something that needs to be repeated.
br /
br /The next half of the book elaborates on how he feels a trading system should be tested. This part of the book has some very good ideas, as well as some unfortunate errors.
br /
br /The very good parts involve his discussion of the permutation test. The Classical hypothesis testing we are all taught in intro stats, has problems when multiple comparisons are made, as researchers in genetics and biology are finding out.
br /
br /In most cases, some correction (ie. the Bonferroni) is made for "false positives" ie. significant results that are due to chance because so many tests were made. These corrections have a drawback in that they make the problem of false negatives--ie. accepting the null when it is false, more likely. In other words, corrections reduce the power of the test. The permutation test the author describes does not have that problem.
br /
br /Yet, this is where the author makes some significant errors (forgive the pun). He seems to believe that significance testing alone tells him all he needs to know: ie. if a rule fails to attain "statistical significance" it is worthless. That is completely false.
br /
br /The "significance" test, or P value, gives the probability of getting results equal to or greater than the observed result ASSUMING THAT THE NULL IS TRUE P(data|null).
br /
br /What we are interested in is: what is the probability that some rule has a positive (negative) return, ASSSUMING OUR DATA IS TRUE P(alternate|data). These are two different questions, and it is the latter one that we are interested in.
br /
br /In order to determine how informative a statistical experiment is, we need to define a few things: 1. the EFFECT SIZE (abbreviated as d)--the size of the difference between the null and the alternate hypothesis.
br /
br /We also need the POWER (1-B), where B is the probability of a false negative--accepting the null when it is false. For a complete analysis, would also need a prior probability--ie. a subjective probability that some hypothesis is worth investigating.
br /
br /Once we have some idea how the alternate hypothesis might be distributed, we can then compare it to the null, and see how much these distributions overlap. if the null and the alternate hypothsis have distributions that share a large area, the experiment does not provide much information, either way, to decide. Other, subjective factors, will lead us to choose one hypothesis over another.
br /
br /More formally, it is logarithm of [(data|null) / (data|alternate)]. This is called the likelihood ratio, and measures the evidence of the experiment for or against a particular hypothesis.
br /
br /The author is completely unaware of the drawbacks of his testing procedures, nor is he aware of the need for prior (subjective) probabilities to determine the utility of his results.
br /
br /In short, he seems completely unaware of the Bayesian approach to this problem.
br /
br /In spite of these drawbacks, the book is worthwhile. Some of the indicators that aren't significant, look interesting when you closely examine their confidence intervals.
br /
br /Some of the rules (available on the website), have a confidence interval where the area on one side of zero is 4-7 times larger than the area on the other side.
br /
br /Take the last rule on his website:
br /[...]
br /
br /The mean daily return of this rule is -0.102501. The confidence interval is -0.23433 to 0.03027. Even though it isn't "significant" most of the area of the confidence interval is negative, up to 7.6 times greater than the area above zero. Suffice it to say, the probability that the rule has a negative return is very high, and might be a worthy candidate to act on from a contrarian POV.
br /
br /The next half of the book elaborates on how he feels a trading system should be tested. This part of the book has some very good ideas, as well as some unfortunate errors.
br /
br /The very good parts involve his discussion of the permutation test. The Classical hypothesis testing we are all taught in intro stats, has problems when multiple comparisons are made, as researchers in genetics and biology are finding out.
br /
br /In most cases, some correction (ie. the Bonferroni) is made for "false positives" ie. significant results that are due to chance because so many tests were made. These corrections have a drawback in that they make the problem of false negatives--ie. accepting the null when it is false, more likely. In other words, corrections reduce the power of the test. The permutation test the author describes does not have that problem.
br /
br /Yet, this is where the author makes some significant errors (forgive the pun). He seems to believe that significance testing alone tells him all he needs to know: ie. if a rule fails to attain "statistical significance" it is worthless. That is false.
br /
br /The "significance" test, or P value, gives the probability of getting results equal to or greater than the observed result ASSUMING THAT THE NULL IS TRUE.
br /
br /In order to determine how informative a statistical experiment is, we need to define a few things: 1. the EFFECT SIZE (abbreviated as d)--the size of the difference between the null and the alternate hypothesis.
br /
br /We also need the POWER (1-B), where B is the probability of a false negative--accepting the null when it is false. We would also need a
br /prior probability--ie. a subjective probability that some hypothesis is worth investigating.
br /
br /Once we have some idea how the alternate hypothesis might be distributed, we can then compare it to the null, and see how much they overlap. if the null and the alternate hypothsis have distributions that overlap (this area can be calculated), the statistical experiment has low power, and does not provide much information to decide what is preferable to believe.
br /
br /The author is completely unaware of the drawbacks of his testing procedures, nor is he aware of the need for prior (subjective) probabilities to determine the utility of his results. In short, he seems completely unaware of the Bayesian approach to this problem.
br /
br /In spite of these drawbacks, the book is worthwhile. Some of the indicators that aren't significant, look interesting when you closely examine their confidence intervals.
br /
br /Some of the rules (available on the website), have a confidence interval where the area above zero is 4-7 times larger than the area below.
br /
br /Take the last rule on his website:
br /[...]
br /
br /The mean daily return of this rule is -0.102501. The confidence interval is -0.23433 to 0.03027. Even though it isn't "significant" most of the area of the confidence interval is negative, up to 7.6 times greater than the area above zero. Suffice it to say, the probability that the rule has a negative return is very high, and might be a worthy candidate to act on from a contrarian POV.
 Must read for serious traders
After reading some of the reviews of this book on Amazon, I was quite skeptical about the relevance and value of this topic in terms of actual practical trading applications. After all, as a trader who's read hundreds of market books, very few of them have actually contained real educational content with practical alpha generating ideas. Usually, it's the same old hogwash: "when you see this pattern you do this (buy or sell) because these 3 charts I've selectively chosen show you that it works" or "a low P/E ratio means the stock will go up, whereas a high P/E ratio means the stock is likely to go down". For serious traders familiar with the random and complex adaptive nature of financial markets and cognitive processes leading to the adoption of flawed trading methodologies, those types of books are seldom satisfactory. Moreover, most market books have either been written by academics with very little actual trading experience or traders who do not possess the intellectual rigor to write something of value for serious market practitioners. David Aronson convincingly punctures this traditional mold by incorporating disciplines such as philosophy, psychology, technical analysis, probability, statistics and others into one coherent analytical framework designed to do one thing: generate alpha by teaching you how to fish rather than handing you a fish. Never out of the hundreds of books that I've read have I encountered a book that has been able to do what the author has done: put many of these interrelated disciplines and ideas together into one book in a detailed, thorough and rigorous manner. Truly, David Aronson has done all of us traders who try to add more rigor to our methodology and analysis a gargantuan favor. For that, we should all be grateful.
br /
br /However, an important caveat is in order. As a previous reviewer has pointed out, this is not a book for those looking for a free lunch. It is presenting the reader with a conceptual framework, a thought process, a rigorous way of analyzing market action that will change the way you approach the markets and seek profit. And that very approach will always adapt and change as do the markets. After all, biology does not have the monopoly on evolution. Therefore, those disappointed by the lack of tradable patterns or ideas are simply missing the point: those patterns and ideas lose value as soon as they are widely disseminated rendering them ineffective and even dangerous since they act as psychological anchors. Not only does the author introduce and apply the scientific method to TA, he walks the reader through the logic of discovering actual predictive price patterns, relates those patterns to human nature through cognition, destroys the premise on which the EMH is based and introduces traders to an empirical method worth the time and effort of learning to avoid being "fooled by randomness". Read it, study it and learn from it. Then repeat the process. Not doing so will provide the counterparties to your trades with the edge. And eventually, your capital.
br /Good study and good trading.
br /
br /
TA Turned Upside Down - Must Read For All Serious Investors
Professor Aronson has produced the next generation of Technical Analysis guide that debunks much of what everyone has known but has never wanted to admit. Written in a concise and clear manner without Greek notation and mathematical proofs it is the working technicians guide that will bring solid scientific analysis to a "sorcery" based following. The book starts with philosophical and psychological explanations that lead to a solid non-mathematical framework of scientific reason for technical analysis. For those having basic statistical training this book will be easily mastered and it's insight immediate. The beauty of this volume is you do not need a doctorate to get the underlying message that what you see is often not reality. As a practitioner of everything random this book puts in perspective the integration of the new technologies and techniques that match the widespread use of technical analysis by everyone who investigates before they invest. As the old maxim has so often been quoted says, "one picture is worth a thousand words", the stock chart now has new meaning for anyone who will read this book. Should be on every investors bookshelf and must read for the new generation of technicians fresh out of graduate school or beginning trading for a living.
TA as a scientific endeavor, it's about time!!
EBTA is a great book for any technical analysis (TA) based trading system developer. This book attempts to put technical analysis on a more rigorous scientific footing that will allow it to flourish as an empirical science.
br /
br /As I see it, this book makes two main points:
br /
br /1) Traditional TA techniques are often defined so vaguely that they are properly classified as pseudoscience. Pseudoscience is characterized by claims that are so vague, they cannot be falsified. When a TA claim cannot be falsified, it cannot be subjected to scientific scrutiny, and so it can be perpetuated even when it holds no merit in predicting market movements.
br /
br /Therefore, any TA techniques that one uses should be programmable (in a computer language) so they can be backtested and either falsified or tentatively supported.
br /
br /2) Even if a backtested trading rule/strategy performs well in a backtest, this is merely a necessary and not a sufficient condition for believing this rule/strat will perform well in the future. Market movements are governed, at best, by a combination of randomness and predictable market structure. If one is not careful, backtested rules/strats may merely represent curve-fitting to the random elements of market movement that will not repeat themselves in the future. If/when one trades based on such overfitted strategies, one is almost sure to lose money.
br /
br /To combat this, one must submit backtesting results to rigorous statistical tests in order to reject results that are likely to be examples of such overfitting. Along these lines, the author gives a very clear introduction to basic statistical concepts/procedures, describes several out-of-sample testing methods, and explains 2 other advanced methods (White's Reality Check and the Monte Carlo Permutation method), that can all be used to analyze backtesting results. Again, applying these methods to backtesting results can help the researcher reject those results that are likely to be examples of overfitting due to data-mining efforts.
br /
br /------------------------------------------------------------------------
br /
br /The author goes into great detail about how and why the procedures he presents should be used, and he also includes a backtesting case study clearly illustrating how to apply the methods he has introduced earlier in the text.
br /
br /The clarity and explicitness of the entire book is a refreshing change from most books on TA. The author provides enough detail for anyone interested to implement any and all the procedures he describes (including White's Reality Check and the Monte Carlo Permutation method) for themselves.
br /
br /If you are looking for a book that will provide you with a specific trading strategy or an in-depth description of relevant data-mining techniques, look elsewhere, but don't ignore this book even in those cases.
br /
br /For, no matter what particular systems/strategies you create/discover, this book will show you how to properly analyze your results. If your results pass the tests this book suggests you use in such analyses, you will be much more likely to make money going forward as opposed to being fooled by randomness.
An island of clarity in a sea of pseudo-science.
A well done treatise on the application of the scientific method to the field of TA. Highly recommended unless you would rather believe in fairy-tales ;)
|
