Book Review: Super Crunchers

This is a review on the book "Super Crunchers" by Ian Ayres. "Super Crunches" is available in NLB at 519.5 AYR

Super Chrunchers describes how regression and randomization has been applied in reality. For people who are studying or have studied statistics or econometrics, the book will show you how these statistical stuff are applied in reality. Overall, I find that this is a fascinating book. However, if you do not understand regression/randomization or if you hate numbers, you may not like this book.

Some interesting points:
1) Companies are using algorithms or super crunching to exploit the long tail. In other words, super crunching may help companies to sell the same product to different customers closer to the customer's reserve (maximum price that the customer is willing to pay).

2) On the other hand, super crunching also helps customer to discern price differences better. Examples are Priceline etc.

3) Randomization (Real-life small scale experiments using randomization principle) may shape government policies as government can test-run certain policies in reality and observe whether the intended policies are beneficial or not baased on the experimental outcomes.

4) Equations (aka regressions, randomization etc) is more likely to outperform a group of experts in predictions. This is especially so if the underlying database is relatively free of errors.

5) Standard deviation can be thought in the following way. That is, 95% of the outcomes roughly fall in between 2 standard deviations of the mean. (assuming that the event belongs to a normal distribution.)

The book also discuss how Larry Summers reasoned out very roughly why hard sciences tend to have more male academics than female academics. (The reasoning is logical and very smart but it is also very sketchy.)

The author, Ian Ayres, has also some interesting prediction tools (weblinks) on this website. (http://islandia.law.yale.edu/ayers/predictionTools.htm)