Statistical Analysis: Projecting NFL QB Failure Using College Stats

[Dante84]

I think a metric that should be included would be quality of opponents. He slightly accounts for that with a broad stroke of "BCS or Non-BCS." They should look at the teams they played against and calculate those opponents win/loss percentage over the course of the respective QB's starts.

Also, you could take that one step further and take into account the quality of the defenses they played against over the course of their starts.

This way it would better categorize those shitty ass QB's from USC and the rest of the lame ass PAC 10/12.

EDIT: Additional thoughts -

ALSO

You should take into account the quality of the QB's own Defense, Special Teams, Running game, and WR dropped balls. This way you know what kind of support they had.

If they have a ton of support, it might ding them. If they performed well in spite of shitty support (Geno) it would be a credit to them.