Not to derail, but another thing that won praise from the emerging football analytics school of thought was a decision to go for two by the Eagles. Down 14 but still in the 3Q (I think), the Eagles scored a TD. Rather than going with the conventional PAT kick, they went for two (which carries an assumed probability for success of 50%) Let's say the score was 21-7 (14 point deficit). The trailing team scores a TD and goes for two. * If successful, they now trail only by 6 points. If their defense holds the opponent (or if the scoring is a back and forth seesaw battle matching them point for point), they now only need to score a TD and kick the much easier to convert PAT kick to win the game. You are playing for the win rather than the tie by shifting the more risky attempt (2 point conversion) to earlier in the game while you have more time on the clock to score more points. * If unsuccessful, they still trail by one score (8 points) and would have another 50/50 shot to tie the game at the end. Of course, going for 2 early in the game and ultimately losing by 2 points because neither attempt was successful is the type of stuff that gets coaches fired if they don't have a progressive-minded owner/GM backing them.