Friday, October 05, 2007

"Why do economists work on sports?" 

Guesting at Marginal Revolution, Justin Wolfers asks the question and gives six possible answers. In a comment, the seventh and I think best is given:
The elephant in the room here is that many sports produce a vast quantity of analyzable data in a reasonably convenient form.
I'm not sure why it's an elephant in the room. It's not just the volume of data but how it is collected. I sat at Veterans Stadium in its last year, in the upper deck, with a good friend (we were watching Phils-Red Sox), and sitting on my other side was a guy scoring the game. Now I score all my games, but often using an accounting style format that I learned from an old Bill James Abstract many years ago. I'm used to seeing the diamond boxes, or people drawing in their diamonds, etc. But what this guy had was entirely different, with recording of pitch type and location and speed. Given where we were, he wasn't a scout. So I ask him, and he says he's paid by Stats, Inc. $50 plus the cost of a seat to score the game. Not every game is scored, he said, but probably 90% are. (I looked at their jobs page, and you can see they pay for data.) And before them, of course, is the Elias Sports Bureau, which does 'official stats" for every major team sport. When you know that all the data has been created consistently by a small set of guys (see this SABR interview with Steve Hirdt), you have an ability to evaluate how people respond to incentives in a controlled setting that's almost as good as an experiment.

This competition in statistics between Elias and STATS (and a couple other providers) produces more and better data all the time, with controls put in place by the nature of games.

I'd also add one more thing to the list of things Wolfers mentions for why we study sports: I do not think it's an accident that there is much more on team sports than individual sports. Team production, and its impact on labor, has fascinated us for half a century because it's peculiar (to borrow Walter Neale's description) and applicable in other areas, like fishing boats or some management settings. Baseball gives you some chance to disentangle team production because it's more sequential (at least for offense and pitching), but even there it's not perfect. So more to do! And it's being done.