One thing I don’t get on this is are they doing the win probability by the outcome or the decision?
seems like it has to be the outcome, because I can’t imagine the decision to punt on 4th and 11 from the 50 increases your win probability by 3.5 percent. So that being the case, of course this thing is going to approve of your decisions when they work. Point is there is a downside when it doesn’t work.
it’s like the analysis of the 2 point play: oh you increase your win percentage from being down 10 to 9… of course! What about when you don’t get it and now you’re down 11? And the article says you convert a 2 pointer at 45 percent so each conversion attempt is worth less than a point, while 2 extra points is an almost guaranteed 10.
so this brings up another concern of mine on all this, which is this isn’t really statistical analysis, this is a football coach trying to apply statistical analysis, which I’m not sure he’s qualified to do with the precision necessary to grab these fractions of percentages.
