College Football Week 5
Welcome to the Week 5 edition of “TSI vs the Field”, my weekly article where I take a look at the college football slate through the lens of my T Shoe Index, and compare and contrast where other respected predictive models align and disagree on the slate. This analysis helps pinpoint where there are betting opportunities on model alignment, and where it may be best to proceed with caution or stay away if the projections vary too much.
Note: SP+ Rankings are the work of CFB analyst Bill Connelly (ESPN), FPI is the Football Power Index from ESPN Analytics, and Sagarin is Jeff Sagarin, formerly of USA Today.
Week 5 College Football Model Alignment:
Model Consensus
Ole Miss (-1.5) vs LSU, O/U 54.5
In a weekend full of high-profile matchups with playoff implications, this game is definitely in the mix as one of the most important. LSU is still somewhat being carried from a perception standpoint by their Week 1 win at Clemson, but Clemson is now 1-3 and has fallen over 20 points in my TSI ratings since the preseason.
Ole Miss is seemingly flying under the radar as a playoff contender, despite being undefeated and really clicking offensively after throttling Tulane last week. Our model average for this game is Ole Miss -5.1 with a variance of just 2.4 points, with all models aligning on the Rebels to cover the number. FPI has the strongest take on the Rebels, projecting Ole Miss -6.4, while even the most conservative model, Sagarin, has Lane Kiffin’s squad by 4. I’d expect LSU to be a pretty public team this week, but the data supports Ole Miss.
College Football Pick: Ole Miss -1.5
Penn State (-3.5) vs Oregon, O/U 51.5
Perhaps the single biggest game of the weekend (I’d hear an argument for Georgia/Bama), this clash of the titans in Happy Valley is going to be must-see-TV in a rematch of last year’s Big Ten Championship Game. Our models all see this one playing out similarly, with Oregon covering the 3.5, if not winning outright. The average projection from the models on this game is Penn State -1, with a variance of just 3.5 points.
Sagarin is the only model that even has Penn State winning by a field goal, at -3.1, while TSI and FPI make Penn State a 0.5-1 point favorite and SP+ actually favors Oregon by a half point. I’m surprised the hook is still available, but I definitely like the Ducks here, and the models unanimously agree.
College Football Pick: Oregon +3.5
Model Disagreement
Old Dominion (-16) vs Liberty, O/U 51.5
Not many teams have been more disappointing relative to preseason expectations than Liberty. Think of them as G6 Clemson. Truthfully, I don’t think TSI has caught up to how much worse they are than anticipated, and it seems our other models also are having a hard time gauging this matchup because Old Dominion is probably better than expected in conjunction with Liberty flaming out (excuse the pun).
Our model average projection for this game is Old Dominion -12.5 with a variance of 11.8 points. FPI and Sagarin both like the Monarchs to cover the 16 here, projecting ODU -17.5 and -17.1, respectively. On the other hand, TSI and SP+ project ODU -10.2 and -5.7, respectively, so the models are all over the place on this game. Based on what I’m seeing in the on-field numbers, I think ODU is the play, but with so much mixed data it’s probably best to just stay away.
San Diego State (-2.5) vs Northern Illinois, O/U 43
San Diego State came out of nowhere and walloped Cal last week, shutting out the Golden Bears in impressive fashion. That type of performance, along with their better-than-expected performances before that, have led to SDSU being improved by 15 points in TSI ratings this season. With that comes some model disagreement because it’s the question facing all predictive modeling: how much do you weigh preseason expectations vs what you’ve seen on the field in a small sample?
Our models average SDSU -3.2 with a variance of 10.9 points. TSI is leading the charge for the Aztecs, projecting them as almost 9-point favorites here, while Sagarin actually favors NIU by 2 points. FPI and SP+ are torn also, projecting SDSU -1.5 and SDSU -4.5, respectively. Transparently, I bet SDSU based on the on-field data I’m seeing, but I wouldn’t blame you if you passed based on the lack of model consensus.
Keep an eye out for this weekly feature on Thursdays and also my Monday early-week college football bets.
