April 6, 2012
Posted by Jay Livingston
Did the bounty system work?
Even people with no interest in sports have heard about the strategy of Saints defensive coordinator Gregg Williams. He offered his players a bounty for injuring opposing players – $1000 if a player was carried off the field, $1500 if the player didn’t return to the game that day.
On the audio released yesterday (listen here), you can hear Williams giving pep-talk instructions to to the defense just before the playoff game against the 49ers. He specifies the injuries he would particularly welcome – a blow to the chin for quarterback Alex Smith, a concussion for receiver Kyle Williams, and as for receiver Michael Crabtree, “He becomes human when we fuckin’ take out that outside ACL.”
Much of the reaction to this story is shock and horror – some of it real, no doubt, by people unaware of football’s backstage, and some of it affected. Among the players, there is anger and genuine surprise. (“One word WOW,” tweeted 49rs safety Reginald Smith.) Others were more sanguine, saying in effect, that football is a violent game where people get injured. Jets linebacker Bart Scott said that getting rid of the bounties wouldn’t change that.
But so far I have seen no data on whether the bounties worked. Did the Saints injure more players than did other NFL teams? Surely those numbers are available.
The only evidence I’ve heard is that the Saints had the highest number of roughing-the-passer penalties in the NFL. That’s probably because they blitz more. Blitzing is a high-risk strategy, and there’s some question as to whether it’s effective. In theory, blitzes should increase the defense’s chances of injuring the quarterback. But the Saints were below the NFL median in sacks.
None of that speaks directly to the question of injuries. The bounty system is a recent and distasteful example of “incentivizing” (a recent and distasteful coinage among economists). Has no sports economist or Freakonmist even counted up the injuries let alone run econometric statistics to see if these incentives worked?
Planet Money addressed the question, although not in great depth. I think there are probably too many confounding factors. It's not like they instituted the bounty alone, it seems like they were also pushing a more aggressive defense, and since the two happened around the same time it seems like it would be impossible to tell.
ReplyDeleteIt doesn't make a huge amount of sense to me. I don't know how much Saints linebackers get paid annually, but you'd think it would be enough that a possible $1k/week is a relative drop in the bucket.
I must have been on some planet other than Money when they ran this, but it says what I said -- more blitzes. If you combine the three stats -- more blitzes, not so many sacks, more roughing calls -- it looks like they were trying to hit the QB even if it was after he'd thrown the ball.
ReplyDeleteAgreed that the money was of symbolic value to the players. but symbolic incentives are still incentives. I don't know what Williams's salary was. Maybe it seemed a little bit like real money to him.
http://endogeneityblog.com/2012/03/30/how-successful-was-the-saints-bounty-program-part-1-get-the-right-data/
ReplyDeleteThanks. Your post on your blog is the first data of any sort I've seen on this question. If the data are as available as you say, that' surprising.
ReplyDelete