FREE hit counter and Internet traffic statistics from freestats.com

Monday, September 29, 2003

Moneyball

While thinking about baseball today I wanted to post a short review of the most interesting baseball book I read this summer called Moneyball: The Art of Winning an Unfair Game by Michael Lewis.

This is the story of the Oakland A's and their general manager Billy Beane. The question the book tries to answer is why the A's, who are one of the lowest revenue and payroll teams in baseball, have done so well from 1998 to 2002 twice winning 100 games and making the playoffs three years in a row and now making the playoffs again in 2003 winning 96 games.

The basic answer Lewis gives is that the A's don't use traditional baseball "insider" knowledge to evaluate the worth of players but rather rely on statistics (specifically "sabermetrics" which is the term coined for non traditional baseball statistics and developed by writers such as Bill James and Craig Wright). In fact, the assistant general manager for the A's is an outsider Harvard graduate and stathead. More generally Lewis casts the A's success as the triumph of science over superstition and the rigorous recording of observations over perception. One of the interesting things in the book is how long it took for any major league team to use this new knowledge that started to be developed in the early 1980s.

A couple of the strategies the A's employ are:

1) Value on base percentage over any other offensive statistic, especially batting average, since it records the fundamental goal of offense which is to avoid making outs. Sabermetricians have long known this and it is always interesting to watch major league managers put some of their worst on base men high in the order because of some secondary characteristic such as speed, the ability to bunt, or even power.

2) Draft almost all college players since the likelihood of high school players reaching the majors is much smaller (traditional scouts overvalue high school players because of the physical "tools" they see)

3) Value deception over velocity in a pitcher. Other teams overvalue how hard a pitcher can throw even though the value to deceive (via a strange release, changing speeds, and ball movement) are ultimately more important in getting major league hitters out.

4) Build closers and then trade them or let them go the free agent route to obtain draft picks. The basic premise is that the idea of the closer is vastly overrated and so it is easy to "create" a pitcher with the kinds of statistics ("saves" particularly) that other teams value but which aren't very meaningful.

5) Don't try to manufacture runs using stolen bases, hit and run, and sacrifices. These are largely self-defeating strategies.

Overall the book is fascinating for a baseball fan although the number of four letter words used was disappointing and in my view was way more than was needed to convey any sense of "reality".

Another criticism I had is that Lewis treats the financial inequalities in baseball as trivial and those who talk about them as unenlightened. Even though the A's use this "new knowledge" to gain an advantage now, eventually it will spread (now the Blue Jays and Red Sox are employing the same tactics) and things will even out. Teams like the Royals and A's will then still be at a great disadvantage since teams like the Yankees have payolls five times greater. Yes, the A's and other teams can win today by being smart, but being rich allows you to make a lot of mistakes and still come out at or near the top. Of course, one of the other overriding points is that the Beane methodology really only came into full swing with the draft of 2002. So it will take several years to see if their drafting tactics pay off.

This book is provoking a lot of discussion throughout major league baseball as I hear announcers and managers talking about it in interviews. In the spirit of full disclosure I should mention that I am a member of SABR (the society for american baseball research), the organization that serves as a home for sabermetrics.