[MCS]

There are times it may make sense to go for two even when it seems unnecessary. For example, I wrote this a while ago about being down by 14 late, scoring a TD, and going for 2 anyway:

"Assume you score two TDs, you make the 2-pt conversion 40% of the time, the PAT 100%, and the opponent doesn't score at all. Then the following are possible:

(1) You make your first two-point attempt. In this case you win because you kick the PAT on your second TD and win by 1.
(2) You miss your first and make your second. In this case you tie.
(3) You miss your first and second. You lose.

(2) is identical to kicking two PATs. (1) happens 40% of the time. (3) happens 36% of the time. So you come out ahead!

The problems is that coaches get blasted for unconventional gambles that don't work, but get only a bit of credit for gambles that do. You will be fired if you miss two 2-pt attempts in a playoff game and lose by 2.

...

The breakeven percentage for this is (3-sqrt(5))/2 which is around 38.1%. "

So even under conservative numbers, this works out well. In real life, it makes even more sense to follow this strategy, because 2-point conversions have more than a 40% chance and PATs have less than a 100% chance.