Monday, March 10, 2014

Moneyball

Michael Lewis's book Moneyball tells the story of how the Oakland A's outperformed most other baseball teams year after year with one of the lowest payrolls in the game. It's a lively narrative with characters like Billy Beane, a under-performing ball player turned over-performing general manager, and Bill James, a nightwatchman with a quirky interest in baseball stats who rejected conventional wisdom and built a new discipline called Sabermetrics.

Conventional wisdom says that a great baseball player has a 0.300 batting average, knocks in 100 RBI's and hits for 30 home runs. But according to James the most important number on the field is three: that's how many outs you can get until the inning is over. And until the inning is over, you can score. So the most important thing when you come to the plate is to not cause an out. From that perspective, batting average has a glaring hole: it doesn't count walks. And why not? Because early baseball statisticians, like Henry Chadwick back in the 1850's, regarded the walk as a pitcher's mistake, not a batter's skill. But he was wrong. That mistake has led thousands of baseball experts to miscalculate a batter's value, and that miscalculation creates an opportunity.

James and Beane and others re-examine all manner of strategies and statistics: sac-bunts and stolen basebases (aka small ball) are statistically not worth it, RBI's are a poor measure of performance compared to slugging percentage, good fielding isn't best measured by errors but by all fielding chances, and that maybe fielding hardly needs to be measured because good batting far outweighs bad fielding. All of this in an effort to find under-valued players and build a low-budget team that could win.

It's the story of how the information age, with the ability to gather and grind all kinds of stats, can overturn time-honored beliefs based on appearances and feelings, not on cold hard reality. And Lewis points to how this is much bigger than baseball. After reviewing James's work he says, "There was but one question he left unasked, and it vibrated between his lines: if gross miscalculations of a person's value could occur on a baseball field, before a live audience of thirty thousand, and a television audience of millions more, what did that say about the measurement of performance in other lines of work? If professional baseball players could be over- or under-valued, who couldn't?" (page 72)

Well, software engineers for sure. In my 20+ years of churning out code, I've seen countless projects go poof, tens of thousands of lines of code that will never ship, and all kinds of people trying to be productive - or managers hoping they are productive - but without any way of knowing how to measure it. We are working with deterministic computational machines, yet the best we can do is estimate and hope that what we're attempting is worth it. It's a sobering exposure of our ignorance.

But it creates opportunity. If something appears to be guesswork, you can step back, find some metrics, and possibly get a new and better view. If you can see where everyone else is miscalculating, you can jump in and make something better. Cold hard reality has a wonderful quality to it: it's real. You can count on it. Once you see it rightly, you can do something wonderful with it. So the only thing that turns out to be truly cold and hard is just our ignorance and stubbornness.

PS. No, I've not seen the movie. People tell me it's great. Brad Pitt and all that. Maybe someday.