Top critical review
9 people found this helpful
Statistical over simplicity
on July 1, 2012
As a statistician, and someone who also loves sports, I have totally enjoyed the "new" satistical approaches to sports stats, from tversky to moneyball to learning you should never punt. Sports arguments are often a lot more fun these days (e.g., 90 percent of the ESPN analysis on NBA draft night is how well the players scored the ball in college, when that appears statistically to be only a minor indicator of professional success.
The problem I have with much of the literature aimed at the general public though, is that is over simplifies the problems, and all to often takes away the argument by assumption. John Maynard Keynes taught us that the big problem with statistics is not the methods, it's having no way to validate the numbers we put in.
So here we have an assumed method of picking the best offensive and defensive teams in history, no discussion of why most of those teams did not win a championship, no discussion of alternative methods. We get probabilities of winning streaks, but only a couple paragraphs on problems with those stats. (player injuries as the only example.). What about the fact that NBA teams almost always lose the second game of back to back road games? What about teams tanking at the end of the season to improve draft position?
I appreciated large parts of this book, but also found myself deeply frustrated with it at points. There are better books out there for people who want to get started on modern sports statistics.