CFBTN and Vegas disagree by more than field on two games this weekend. The model still likes Pitt to win, but by about 2, not a touchdown. The logic here, it seems, is that USF has nothing to play for while Pitt does (i.e., bowl eligibility). It's a smart model, but it struggles with contextualized human incentives.
It's funny to think that the world decided West Virginia was good because they scored 70 against Baylor. Scoring 70 on Baylor is as hard as scoring drugs at TCU (scoring points against TCU is another matter). At the time I discounted the victory because they only beat Baylor by 7. WVU might score 70 if they played again, but I'm pretty sure Baylor would win. WVU is 4-7 ATS this season. They'll finish the regular season 4-8. Kansas is terrible, but they lost to Northern Illinois, Oklahoma State and Texas by a combined 17 points, not the 20 WVU is giving.