Wednesday, November 25, 2009

This research is pretty foul

This story from ESPN summarizes research on basketball fouling. The key points are: home bias, compensatory foul calling and more compensatory foul calling on national TV. None of these seem particularly surprising.

Missing from the ESPN summary (but hopefully not the paper): standard errors!

Missing from the ESPN summary (and maybe the paper too): behavioral responses and social welfare calculations!

Consider this:
"The bigger the difference in fouls between the two teams playing, the more likely it was that the next call would come against the team with fewer fouls."
This is interpreted in the ESPN summary as resulting solely from compensatory behavior by referees, but it could also result from compensatory behavior by players. After all, the players are, throughout the game, trying to learn about the preferences of the referees. A big foul disparity in favor of one team might be interpreted as the referees being tougher on one team, with the result that that team responds by (optimally) fouling less, as it faces a higher cost of fouling.

And consider this:
We'd like them to have no memory and strictly call what's going on on the court," Anderson said.
Anderson is one of the study authors, and I think he is wrong. If fans get more utility from close, exiting games, then compensatory foul calling may be socially optimal.

That's two chapter ideas for Freakonomics III in one day. Maybe I'll get a cut of the royalties?

Hat tip: Greg Nicholson

No comments: