A reader asked: “I am curious how reliable some of the factors referenced in the Stock Trader’s Almanac are, but I see no reference to it on your site. Could you review the book and/or the primary strategies in the book? I would be curious to have your perspective on how rigorous its analysis is.”
Stock Trader’s Almanac covers a lot of ground. The description indicates that it includes a lot of reference data and that there is very likely overlap in anomaly coverage with CXOadvisory.com, especially in categories such as “Calendar Effects,” “Political Indicators” and Technical Trading.”
You can use these three categories (and key word searches) to identify similar analyses and thereby get a second opinion on specific anomalies.
Some general risks of data-intensive asset return analyses are:
- Data snooping (mining) bias, both direct and indirect, which tends to incorporate luck into inference.
- Assumption of normality of asset return distributions, which tends to overvalue statistical metrics (like mean, standard deviation and significance test outputs).
- Look-ahead bias, which assumes knowledge not available in real time.
Realistic accounting for trading frictions and use of capital are other risks. For further discussion, see “The Demon’s Drudgery”.
