Saturday, September 24, 2005

Moneyball

I'm not the biggest baseball fan, but this book intrigued me. Moneyball is the story of how thee 2002 Oakland A's were competitive, despite having a very low payroll.

The answer was through statistical analysis - the team's management found certain traits were valued incorrectly by the market. For example, batters who hit lots of home runs typically command a premium. However, what is more important is the guy who reliably singles or draws walks. Similarly, pitchers were often rated by their earned run average, but what is more important is the ratio of strikes to walks. A statistical analysis showed the ERA is basically utterly random - it is theorized a pitcher doesn't actually control this at all, it is more what the fielders are doing.

Another interesting point was evaluating position players on their number of errors. The error is unique in sports, since it is a play scored by what some else (the official scorer) thinks should have happened. The A's management at the time felt that a guy who had an error at least did one thing right: he was where the ball was. Whereas, the guy who never had any errors might be like that because he reacted too slow or couldn't run fast enough.

Even more interesting were formulas they developed to predict how many runs were needed to win a certain percentage of games, and in turn figure out how to form a team to get the right number of runs.

I'm a so-so baseball fan, but I found this book a very interesting mix of sports, statistics, and economic analysis focused on evaluating talent and picking up bargain players. Now that the A's secret is out, I wonder if the market has changed out from under them? Or, how have they done since 2002?

No comments: