[MCS]

I still think you underestimate the potentially very significant negative impact to coaches if variance decides it doesn't like them and they lose a couple of games this way. At least until owners/fans/writers understand this concept better and accept it. In baseball, EXECUTIVES are completely willing to ignore all sorts of data re: who is a good player and what helps a team win, and the baseball data are much stronger than the football example.

There is a large enough penalty for losing in an "uncommon" way that it is not fully offset by the increased win probability. The USC play against Notre Dame is a good example. If they lose on that, Pete Carroll gets crucified (well, or he would if they hadn't won 800 consecutive games, though he's kinda earned a pass).