I think that it is a duty of a professional trader, both institutional or private, to buy all the books that the publishing industry issues every year. I agree with you that in the most part of them you will find nothing new and they will not help you in becoming a better trader. But as always there are exceptions. One of these exceptions is “Seasonal Stock Market Trends” by Jay Kaeppel, Wiley, 2009.
First and foremost it is not “another book about seasonal” in the sense it fills the gap many other books on seasonality has as far as statistical significance is concerned. Seasonal patters are often introduced with a sample size which is too little to have any statistical significance. What is statistical significance ? Let’s take the sniper example: if you give one shot to a sniper and he will hit the target you will never know if he is really a killer or simply a lucky man. If you shoots 100 bullets and you put all them in the center of the target you surely will be a damn good sniper.
The same with a pattern: if the stock market goes up 15 times out of 17 when you grand mother cooks a cake this does not mean that the relationship in between your grandmother’s cake and the stock market is a real one. The reverse would be true in the case you would be allowed to check the relationship over a 200 cakes’ sample. Many seasonality students, like Jerry Toepke, by Moore Research Center, one of the most respected website about seasonality, openly admits that in many cases you cannot study seasonality over a big samples size because seasonality is changing decade after decade so the same the economic relationship that rely on it.
Jay Kaeppel avoids to be trapped in this doldrums because he deals about stock markets and the Dow Jones data he uses go back in time 70 or 100 years. The hope is that economic relationship among variables that influence the stock market would be much more lasting and enduring than those on the commodities markets. In any case many of the seasonal pattern Kaeppel investigates are not new but new and intriguing is the way he checks all them with a sound statistical testing. Kaeppel examines the January effect, that is the effect by which if the first month of the year the stock market is bullish the full following year is bullish too. And he modify the original approach in a new ways so that signals are historically right the 90% of the times.
Then Kaeppel studies the holidays effect, that is why to stock market tends to perform on a better than average basis during trading days that lead up to major stock market holidays. Another topic of the book are the monthly trends, that is the author tries to break down the trading month as much as possible into favorable and unfavorable trading days. Readers will be amazed to learn that the best days to enter the stock markets are the second one and the last ones, the resulting equity lines are unbelievable ! I did not like very much the yearly trends, that is the behavior of a year depending of the proximity with a decade or the next one: when I do not understand the logic of something or simply I think there is no logic even interesting results from a test look suspicious.
Eventually Kaeppel research also the “sell in May and go away” effect detecting some pattern that readers will find really astonishing.
In any case, this is a book you really need to buy.