More evidence of the value of two-point attempts
Yesterday, I published a piece detailing why NFL teams could really help themselves by going for two far more than they already do–perhaps almost all the time. If two-point tries yield more expected points than extra points (which I believe to be the case, particularly for a team with a potent offense and poor kicker, like Dallas), then teams should only kick an extra point in very specific game situations, such as a tie late in the game.
I did a little research and found Advanced NFL Stats published a similar article on why teams might want to go for two almost all the time. In it, they mention the conversion rate for two-point tries from 2000-2009 was 47.9 percent. Not stellar, but only because of poor play selection. Teams passed the ball on 74.2 percent of attempts, despite an overall conversion rate of only 43.4 percent on passes. On run plays, the rate jumps to 61.7 percent.
ANS goes on to suggest (as I did yesterday) that, even after an increase in rushing attempts, the Nash equilibirum for the two-point conversion percentage would equalize at a number over 50 percent. They write:
In 2009, the success rate for extra point kicks was 98.3%, and so far in 2010 it’s 98.8%. So for 2-point conversions to be the higher expected-value play, it would only need to be successful about 49.5% of the time. A strategy mix that’s heavy on running would almost certainly exceed that rate.
Later, ANS suggests a two-point try should be attempted whenever the game is still one of point-maximization. Isn’t the game always about maximizing points? Well, no. Early in the contest, teams attempt to secure as many point as possible. Later, however, game situations dictate strategy. It would be foolish to attempt a two-point conversion in a tie game with 20 seconds remaining, for example. In that case, even if a two-point try yields more expected points than an extra point, the game is no longer one of point-maximization.
ANS suggests games are generally point-maximization contest until the fourth quarter, but I disagree. While I don’t think there’s a set cutoff point as to when two-point tries should be automatic, I know I wouldn’t be looking to maximize points during a tie game late in the third quarter. At that time, the chance of the game yielding no more scores far outweighs the expected points gained by going for two. Thus, after scoring a touchdown to tie the game late in the third quarter (or even in the beginning of the third), I would attempt an extra point to take the lead.
The point-maximization cutoff point is less important, though, than the realization that two-point tries are almost certain to yield more expected points than extra points. Thus, at least in the first half of games, the stats seem to indicate that two-point attempts should be a no-brainer. And contrary to popular opinion, kicking an extra point is the “risky” play, only to be used in specific game situations.