Maybe I don't understand the relative importance of the coaches in the outcome of the game, but if you fire the defensive coordinator and replace him with the secondary coach and the secondary begins rapidly deteriorating (I believe they were allowing about a 50% completion rate before and 75% after the change), then doesn't that imply the secondary coach is actually good and desperately needed back in the secondary? And not necessarily that he's a horrible coach?

I think theguyotc actually did get several things wrong. First of all, numbers of penalties shouldn't be modeled with a linear regression since they're count data, so you should use a Poisson regression. If you do this correctly it goes from an increase of 0.46 penalties/game to an increase of 0.05 penalties/game. Furthermore, neither of these are significant, indicating there's no reason to believe any increase. The second problem is obviously sample size, so if you do something as simple as removing week 9 from the equation you actually get a decrease in penalties per game, though it's still not at all significant. And really, after the travesty of officiating that happened in that game, don't you think we should drop it? :) -Max