Glad to see this topic get resuscitated, because I had been thinking of starting a similar one. My interest is less in debating the specific choices than it is in exploring the psychology behind the decision-making process. Consider three scenarios:
1. The one in the OP: Should you go for two after scoring in the 4th quarter to make it an 8-point game
2. Should the Pats have gone for it on 4th and 2 in the famous 2009 Colts game?
3. Should the Packers have gone for two after last week's Hail Mary, rather than sending the game to OT?
In all three cases, the numbers argue pretty strongly in one direction, and in all three cases, what they argue is different from what the conventional wisdom would say you should do.
What I find really fascinating -- and Bill Barnwell touched on this when he appeared on 538.com's
"Hot Takedown" podcast -- is that whenever people argue that the stats should be ignored in the specific case under discussion, it's always to argue in favor of the CW. You never hear anyone say, "Well, the numbers say we should punt here, but based on these conditions it makes sense to go for it."
I think what happens is that people start out with the assumption that
of course you should kick the XP/punt/etc., and then, when challenged, come up with rationalizations to support it. Read this entire thread from the beginning, and it's like the living embodiment of the famous Gandhi quote "First they ignore you, then they laugh at you, then they fight you, then you win." The initial reactions to the OP were incredulous ("The owner would fire you on the spot!") Then people started acknowledging the numbers and saying that maybe it was a close call. Then they started accepting that it might make sense in a vaccuum, but listing exceptions where you should ignore the numbers. Maybe in another year or two everyone will have come around to the default assumption that you should make the higher-percentage play. (And then, five years after that, NFL coaches may even start doing it.)