Monday, February 18, 2008

Super crunchers need intuition


I finished the new book by Ian Ayres, Super Crunchers: Why Thinking-by-Numbers is the New Way to be Smart. It got the recommendation from Steve Levitt the author of Freakonomics that it could change the way you think. I should like this book as someone who has been a quantitative analyst for his entire career. I have always felt comfortable with the discipline of systematic investing, yet my feelings were mixed at the end. There is the conclusion that looking at the numbers will provide better insight. Looking for long-term relationships will be able to eliminate the problems of emotions with investing. There will be less risk of being swayed by emotions or the whims of crowds.

Ayres does convincing job of making you feel like this should be the only alternative to decision making. In fact, he was thinking of calling this book, the end of intuition. He provides great examples of how quantitative investing does a better job of diagnosis than most physicians. Equations do better than many experts. Randomized trials are an excellent way of providing insight on alternative theories.

Still, in the investment area, there may be room for intuition. The room for error with investment decisions is large. Many regression equations trying to provide explanation of returns find that they can explain less that 15% of the total variation in returns. The results of many regressions are a function of underlying assumptions. The variables in many regressions are unstable. Intuition is needed to provide the right quantitative framework and power to know when to change models. My conclusion is that super crunchers need intuition in order to properly set the problem and interpret the results. There is still no substitution for experience in interpreting what models are telling the analyst. There no chance for randomized trials in the normal sense with investing. Quantitative investing without intuition is as dangerous as using the seat of your pants. There has to be a balance between the intuition that comes from experience and the power of focusing on the numbers.

1 comment:

Mark Rzepczynski said...

It is not an issue of being for or against quantitative approaches. The fact is that as you do more work with numbers you understand their limitations.