Michael Giberson
At Stumbling and Mumbling Chris Dillow ruminates on “Norms, agency, and competition,” which is just some fancy econo-speak for a post about why football coaches prefer conventional strategies that reduce the chance of their team winning. Dillow notes David Romer’s work (via James Kwak) on American football, which shows coaches punt too often on fourth down, and research by Christian Grund and Oliver Gurtler on “proper football” (Dillow’s term), which shows coaches in the Bundesliga too often add an attacking player when down a goal even though it reduces their chance of winning.
For more on Romer’s work, see this article in Contingencies (a magazine of the American Academy of Actuaries). Related, “Never Punting,” a guest post at Football Outsiders.
Dillow notes the two responses are both the conventional approach, but observes that in American football the convention favors an overly conservative choice while in proper football (I’ll accept Dillow’s terminology) the convention favors a too risky approach.
For Dillow, “The question is: why? Is it because coaches are ignorant of the statistics and so follow herd mentality?” He doesn’t offer an answer, just a good question. It could be that the outside analysts are wrong, and coaches right, but then there should be a problem in the analysis to point to. Is there such a problem? (The Contingencies article notes that Texas Tech Red Raider coach Mike Leach is among coaches with a reputation for aggressive action on fourth down, but adds that even he doesn’t go for it as often as the statistics recommend.)
Interesting topics as the American football season gets underway, and most proper football leagues just underway.
The reason is simple: coaches who do what the statistics say, lose their jobs faster than coaches who play “by the book.” Their teams may win at a marginally higher rate. But when they do lose, more of the criticism is put on the coach, and it’s much more vicious. Ask “The Always Innovative” Sam Wyche, who actually got the Cincinnatti Bengals to a Super Bowl. His 12-man huddle was so effective the NFL banned it, his no-huddle offense was widely copied. But when something didn’t work, suddenly he was “Wicky Wacky Wyche” and a complete idiot.
That sounds right, but then it just pushes the question up to owners, who fire deviant coaches (even though, presumably, they have better than average records overall).
Owners may run a simpler heuristic: Winning? Keep coach; losing but like coach? Keep one more season; losing and don’t like coach? Get a new one. (Slightly more formally: hard for owners to monitor coaches individual contribution to team output, so owner substitutes by watching the easy to monitor win-loss record which is in principle correlated with coaching quality.)