Game Theory Analysis of Superbowl 49 Ending
The Seattle Seahawks suffered one of the most memorable Superbowl losses of all time when the Seahawks, just a yard away from victory, threw a last minute interception to the Patriots. The Seahawks were down 4 points with 26 seconds left. They had 3 more attempts to move the ball one more yard. The Seahawks had an all-star running back, arguably the best in the league. Why did Pete Carroll, the Seahawks coach, choose the pass? Was it an awful choice? Justin Wolfers and Robert Lane Green have their analysis of why the choice was a bad one outlined in the article “Three’s a Charm” in The Economist.
First, lets outline why Carroll chose to pass. It’s likely that Carroll was relying on a game theoretic model in which he had established a ratio in which pass to run was ideal. The Seahawks had a much stronger running game, but if they ran all the time then their running game would be uselessly predictable. The Seahawks used a mixed strategy in order to develop an unpredictable reputation over the course of the season. If we look at the statistics from the regular season, the Seahawks chose to pass 46.2% of the time and run 53.8% of the time. However, during the post season, this percentage transitioned to pass 44% of the time and run 56% of the time(ESPN statistics). If we look at the Superbowl statistics, The Seahawks chose to pass 42% of the time and run 58% during the Superbowl (ESPN statistics). Wolfers and Green argue that the value of high-percentage plays increases and reputation decreases as the number of plays during the season decreases. These statistics substantiate this idea.
They claim that the second down choice of Carroll was the last chance that the Seahawks had to establish unpredictability in the minds of their opponents. “Something like this must have been going through Pete Carroll’s mind. After the game, he inelegantly said that he was intentionally “wasting” a play. Had he said he didn’t expect the play to succeed but that he was sending a signal, hoping his opponents would look out for a pass also on third down (the second of the three plays in question)” (Wolfers). They also claim that Carroll went wrong in his hope to send a last minute signal of randomness to the Patriots. There were 3 downs remaining. Marshawn Lynch, the running back, had not had a fumble the whole season, while Russell Wilson had thrown 6 interceptions during the season. Lynch averaged over 5 yards per carry during the season. In the short term, over 2 or 3 games and given the scenario at hand, run would have been a clear dominant strategy because the game was ending.
The important takeaway is that while over the long term a mixed strategy is the best strategy for not getting taken advantage of. But in the short term, when your reputation is firmly established as random, your dominant strategy is to go with your higher payoff strategy.
Superbowl Stats: http://espn.go.com/nfl/boxscore?gameId=400749027
Season Stats: http://www.seahawks.com/team/statistics
Economist Article : http://www.economist.com/blogs/gametheory/2015/02/applying-game-theory