Fencer
Pro Bowl Player
- Joined
- Oct 2, 2006
- Messages
- 14,293
- Reaction score
- 3,986
The problem with going for 2, unless you really have to, is you're only doubling the points from a near certain 1 but you're doing it on what is, statistically, a less than 50-50 shot. So mathematically it just doesn't work. Google tells me there's a 47.9% success rate. Think of the money Vegas could make offering double or nothing when the house wins 52.1 percent of the time.
Going for it on 4th down, of course, is wholly different, partly because the success rate on 4th and 2 or less is good and there's so many other variables (likelihood of scoring if you make it, likelihood of allowing point if you fail/punt).
But going for two is nothing more than an easy statistical exercise - with the one exception being if you have a play that, based on film study, you are confident will be successful.
It's not that simple. Certain eventual point scores are more likely than others (some are even impossible). So expected number of points is not the whole discussion.
Also, there's the value of information. If you attempt the conversion and either try or fail, both you and the other team can make subsequent decisions based on the result. The best example of this argues against a 2-point conversion, as follows:
If you're down 7 and score a TD late, it's a reasonable choice to go for 2 and let the game's outcome ride on that single play. However, if you break a tie by scoring a TD late, it would be crazy to go for 2. If you fail and the other team scores a TD, they go for 1. If you succeed and the other team scores a TD, the go for 2. (And if they don't score a TD, it doesn't matter what you did for a conversion, so we can leave those contingencies out of the analysis.)