This book explains how a scientific approach can be used to determine if a technical analysis (TA) rule is worth using or not. If you are looking for a recipe book, this is not it. There is no specific TA rule or approach which is promoted here.

The author first point is that TA – as is generally practiced today – is subjective** **and lacks discipline. Yet it is widely used, because it successfully exploits the fallacies of the human mind (for example: overconfidence, hindsight bias, the predilection to find patterns where none exist). For example, technical analysts always seem to have a ‘good story’ about what is happening on a chart, and the mind tends to fall for a good story. Another example is the Eliott Wave Principle, one of TA most enduring theories (and a good story too) because of its ability to minutely curve-fit historical data.

The author advocates a more scientific, objective approach to TA, grounded in statistics. A clear and well-illustrated introductory statistics section is included. More than mathematics, it is rigor and logic which is required from the reader. Eventually the author presents a statistical significance test for trading rules found via data mining and the results of testing 6400 simple rules (the result is that no rule is found to have statistically significant returns).

A great contribution of this book is its layman explanation of the **data mining** problem. An insidious problem with multiple forms, data mining explains why it is easy to devise a strategy that was **a winner in the past, but will fail miserably going forward**. There are many software platforms available to backtest strategies and optimize them. It has become easier than ever to fall in the data mining trap. However, designing and testing a strategy – and in particular backtesting – requires great rigor and discipline. If you are inclined to believe this, then this book is a must-read.