Stats in football (in the wild)
Cassell to Moss. TOUCHDOWN. With only seconds left the Patriots had completed a drive started deep in their territory to pull within one point of the Jets. They kicked the extra point, went to overtime, and lost. (After having the Jets at 3 and 17.) If the Patriots had gone for two after the touchdown, they could have won the game right there. So should Belicheck have gone for two?
Endgame Technologies has developed a simulator for football games called ZEUS. According to these simulations Belicheck should have gone for two at the end of regulation instead of kicking the extra point and sending the game to overtime.
I’ve been interested in decision making in football for a long time, especially the decision to go for two points after a touch down instead of kicking the extra point. The article “Refining the Point(s)-after-touchdown decision” by Harold Sacrowitz is an excellent article on the subject. In his results he develops a table for when to go for two or kick the extra point in order to maximize a teams chances of winning.
More recently, an article “Do firms maximize? Evidence from professional football” by David Romer, investigates NFL teams decision about going for it on fourth down. He argues that NFL teams are kicking (punting and going to field goals) too often and they would increase their probabilty of winning by going for it on fourth down more often.