I received an interesting phone call from my friend Phil last week. I hadn’t heard much from Phil, but he called me in the middle of a ferry ride to San Francisco to tell me that he was reading a book called Moneyball by Michael Lewis, and that he was sure that I’d enjoy it. I’m a hundred or so pages into it now, and I must agree, I do like it.

The book details the inner workings of the Oakland A’s: the baseball team that has fueled my fledgling enthusiasm for the game and which has posted remarkable win/loss records in the last few years, despite spending only a small fraction of what larger market teams spend on payroll. Moneyball attempts to show how this is not an aberration: that Beane and the A’s are actively pursuing a rational, scientific method for selecting undervalued players that other teams have passed over.

I first became exposed to similar ideas when I studied parimutuel betting systems in my late teens and early twenties. I realized through study that the goal of trying to make money at horse racing wasn’t really so much about picking the horse that would win, but rather selecting horses that paid well when they did win. The idea was to identify inefficiencies in the race: horses which had a greater chance of winning than the money odds would indicate. If you could identify these horses often enough (enough to overcome the cut by the track), then you could conceivably make money at horseracing. Being rather young and undisciplined, I didn’t pursue this idea too much further, but the limited reading I have done since then has indicated that such market inefficiencies are hard to find, as others have tread this path fairly often since then.

The lesson that I should have taken away from these early explorations is that if you can shed the biases and preconceptions that surround an activity, you can conceivably exploit the biases and preconceptions of others and make money. Not a bad lesson.

Lewis’ book points out that the vast majority of major league baseball teams have yet to pursue the goal of winning with this same degree of dispassion or rationality. This is perhaps not surprising as baseball organizations are full of former minor league and even major league players who act as scouts and advisors. They often value players by purely meaningless features (such as weight) and often ignore features which statistically are shown to be valuable. For instance, it is fashionable for teams to draft promising high school players and pay them a great deal of money, yet only one player in four that is drafted out of high school will ever play in the majors. If you wait to draft such players in college, you double the chances that they will see the majors, and you have a lot more data to study to decide whether the player is a good risk or not.

The other bias that leads to misconceptions is the use of batting average to evaluate players. The single most important indicator of eventual big league success isn’t batting average: it is on base percentage. If you think about it, it might not be surprising that this is so: after all, the only thing that really matters in baseball is three outs. If you don’t make three outs, you can still score. The probability of getting an out is just one minus your on base percentage. You minimize the number of outs by maximizing your on base percentage. It seems simple, but the league continues to present awards based upon on batting average.

I hesitate to recommend a book this enthusiastically after only reading a hundred pages, but I must thank Phil for turning me onto it. It tickles my appreciation for the meta-game of baseball: the spending of money to formulate a competitive baseball team. As I sit in the stands watching Barry Zito arc curveballs into home plate, a different portion of my brain will be alive with the possibility of understanding the game of baseball at this deeper level. Two thumbs up.