I didn't watch the game but I know the result (saw the re-cap). Hehehe. Heeheehe. HAHAHAHAHA!!! Burn baby burn! F you Giants! What's that smell? It ain't victory, pal!
This is very dependent on the probabilities - if you lower the probability of making the 2 points the advantage drops quickly (of course if the probability of making the kick is less than 100% it goes up). This also assumes independence between the two events and constant probability of all events. I think both assumptions are questionable. The only way this matters is if the Giants score TDs on two straight drives, while the Falcons do not score at all. That clearly implies that the Giants offense is playing well, while the Falcons offense is not; any momentum is on the Giants side, giving them more of an advantage in overtime than they would otherwise (I wouldn't have gone for two either time, but see below). I also think that they would have a better chance of succeeding on the two-point try at the end of the game when they're down 20-19 than earlier in the game when they're down 20-12 because the defense at that point would be much more worn down and demoralized, having just given up back-to-back TD drives. It is of course very difficult to quantify things like that; the data are out there, but not easily available. Having said all of that, I believe that my initial strongly negative reaction was incorrect. The people at fivethirtyeight.com have looked at this (https://fivethirtyeight.com/features/when-to-go-for-2-for-real/), and their calculations imply that a team down 8 points with 10 minutes left or less should go for two, in the sense that there is a bigger increase in the probability of winning the game going from 8 points down to 6 points down than there is a decrease in going from 7 points down to 8 points down. The actual gains are far smaller than the numbers you gave would suggest, however, with this game being an example - it didn't end up mattering because the Falcons scored anyway (which is why none of this applies if there's more than 10 minutes left in the game). The fivethirtyeight article is interesting because they find that a bigger gain in going for two is when you're down 4 points, even if it's fairly late in the game, which I doubt anyone would expect. Before we have too much sympathy for Shurmur getting criticized for this decision, however, we should still remember that even if he was right (or certainly not wrong) in this decision, he deserves to be completely trashed for his decision to call quarterback sneaks twice down 11 points with no timeouts and less than a minute left in the game (and he confirmed that it was his call, not Eli's, despite what the announcers said at the time). That shows a shocking lack of understanding of the importance of time at the end of a game. His justification was that not calling a sneak because it would be disastrous if it failed is a "defeatist" attitude shows an amazing amount of arrogance, particularly for a team with a lousy offense in general but several quality weapons as receivers. I really don't see any evidence to suggest that he isn't completely over his head.
