Give up punting and rational maximizing

I see this almost every weekend: a team is in a bad but not completely hopeless position, they’ve got the ball, as a practical matter they have to score to keep hope alive, and — they punt. It’s a give up play: one which unambiguously decreases the chances of winning. Three examples from the past two days, one from college and two from the NFL:

Iowa-Maryland

Maryland is down by 16 (two scores), fourth and eight from their own five, three minutes left, and no time outs.

Jets-Oakland

Jets are down 14, fourth and seven from their own 21, 3:19 to go, not sure of time out situation.

Green Bay-Denver

Green Bay is down 19, 6:50 to go, fourth and seven from their own 32, one time out left.

Now in each situation the team with the ball is a big long shot to come back, but stranger things have happened. What’s completely obvious — so obvious that you don’t need David Romer to run the numbers — is that voluntarily turning the ball over (why punts aren’t characterized as turnovers is an interesting question) at this point in the game decreases your chances of winning from slim to a lot closer to none.

So why do coaches do this?

One explanation is that in each case the coach has performed a quick Machiavellian calculation that his own self-interest is best served by attempting to minimize the margin of defeat rather than continuing to try to win.

I don’t really buy this, or at least I don’t think it’s as much of a factor as the reflexive and indeed almost instinctual impulse that drives all sorts of coaching decisions. That impulse is to avoid an immediate bad result (turning the ball over in scoring position for the other team), even when doing so increases the odds of a much more costly bad result (losing the game).

All of which is to say that a lot of things can be explained more plausibly by assuming the decision makers in question are idiots rather than rational maximizers of their utility.