College football’s tenth full week was a sparse one. Few contenders played, and even fewer played meaningful competition. Still, things happened, so as far as the College Football Playoff picture goes, things changed. Here are those things, with insights from our model:
The SEC East Was Pretty Much Decided—But the Scene’s Still Grim for Georgia
Florida isn’t entirely out of the SEC picture following their loss to Georgia. The Dawgs still have to host Mizzou and Texas A&M while traveling to Auburn. The Gators have Vanderbilt at home and Missouri on the road to wrap up conference play. It’s possible that Georgia will drop two games, Florida will take care of business, and Dan Mullen will get his first shot at an SEC title.
It isn’t likely.
With the victory on Saturday, Georgia climbed to 29.1% likely to win the SEC. Florida sits at only 3.0% likely. But while the Bulldogs hold the rights to play Alabama or LSU, our model is skeptical that even a victory over one of those two teams would be enough to put UGA back into the playoff field. Georgia’s playoff probability is only 3.8%. Assuming they need to win the SEC to have any shot of making the playoff, this implies only a 13% probability that Georgia cracks the field even if they do, indeed, win the SEC.
Why?
Our model expects the College Football Playoff Selection Committee to punish Georgia much more significantly for their home loss to South Carolina than, say, the AP Poll is punishing them. Georgia is sixth in the AP Poll. Our model expects them to only be ranked 14th in the first official rankings, which will be released tomorrow night (more on those below). While our model doesn’t treat the playoff race as the horse race that is traditional college football polling (our research inclines us to believe the committee’s approach is more formulaic than horse race), it’s still not very good to be 14th at this stage in the game, and it’s reflecting our model’s belief that if the CFP Rankings extended far enough, the South Carolina team that beat UGA at home would be ranked 41st. That’s a bad loss. A loss bad enough to make overcoming it fairly doubtful, even if the Dawgs end up with a road win over Auburn and a neutral-site win over Alabama and LSU.
Working in Georgia’s favor, though, is that many of the instances in which our model sees them winning the SEC and not making the playoff include a second regular season loss. The aggregate ratings our model employs currently have Georgia’s trip to Auburn rated a tossup. Make the assumption that nearly all simulations in which Georgia loses a second time end with the Dawgs not in the playoff, and that 13% number from earlier denoting how likely it is that Georgia will make the playoff if they win the SEC jumps to something more like 30% or 35% in situations in which they win out. It’s still not likely, but it’s a few orders of magnitude better.
For those who are still skeptical, a good comparison to this situation is 2016. That year, Penn State won the Big Ten. The Big Ten made the playoff. Penn State did not. The committee elected to go with an Ohio State team that had lost to Penn State head to head and did not have a conference championship over the two-loss Nittany Lions, one of whose losses came against Pitt.
The comparison isn’t apples-to-apples. Georgia, unlike Penn State that year, came into the season with high expectations, which may impact the committee (on our list for offseason research is how indicative things like the preseason AP Poll are of how significantly the committee over or under-ranks teams relative to their average formula). Georgia would only have one loss in this situation, not two (Penn State’s other loss that year came on the road to Michigan). Still, it’s not hard to imagine the committee looking at, say, an 11-1 LSU (or Alabama) and deciding they’re more deserving of a playoff berth than a 12-1 Georgia. It’s more difficult, but not impossible, to imagine the committee looking at a 12-1 Alabama (or LSU) who lost to Georgia in the SEC Championship and deciding they’re more deserving of a playoff spot than the Dawgs. 2016 set the precedent (which can, of course, be overruled) that head-to-head results and conference titles only go so far.
We’ll learn more about the Dawgs’ chances when we get those first CFP Rankings tomorrow evening, and we’ll be updating our model in the wake of that news, as it will indicate for us how impressive the committee thinks the wins over Florida and Notre Dame are, and, more importantly, how damaging that loss to the Gamecocks is going to be.
The Pac-12 Is Alive
One of the more eye-opening changes in our model’s readouts this week is that the Pac-12 is now ahead of the Big 12 in probability of sending a team to the playoff.
The importance of Utah’s victory over Washington is naturally going to be understated. Washington, after all, was unranked. It may have been essential, though, to the playoff case of the eventual Pac-12 Champion. If Utah wins the conference, they’ll need to still be at only one loss to have any sort of chance. If Oregon wins it, they might need Utah to be a top ten team to give their résumé some added heft. Prior to the week, Utah only beat Washington in 53% of our model’s 4,000 simulations, so this wasn’t just another game against an unranked opponent. It was a potentially lethal tripwire. Utah came away unscathed.
For Oregon’s part, the Ducks were only a 61% favorite down in Los Angeles, and they put a whooping on USC. Oregon’s got the Pac-12 North all but wrapped up now, with a two-game lead and only three contests remaining.
The Pac-12 has work to do, but with Oklahoma holding a worse loss than that of Oregon (at this point), and Baylor expected to be an underdog in two of their final five games, things aren’t hopeless. They may get more hopeful still if the rankings come out more bearish on Penn State than our model expects. Currently, our model is rather sold on the idea that the committee would include an 11-1 Penn State, with the only loss coming in Columbus, over a 12-1 Pac-12 or Big 12 champion. Our model also believes, though, that Penn State will be ranked second tomorrow night and Ohio State will be first. If Penn State’s résumé-to-date isn’t as impressive to the committee as our model thinks it will be, and if Oregon and Utah’s are more impressive than our model expects, it will further inflate the Pac-12’s growing hopes. They’ll still, though, have the possibility of an 11-1 Alabama or LSU to worry about.
Goodbye, Group of Five
SMU’s trip to Memphis was thrilling, and our model expects four AAC teams to make the cut in tomorrow night’s top 25, more than the ACC and Pac-12 combined. Still, SMU and Appalachian State’s losses cut into the Group of Five’s already miniscule playoff hopes. They’ll be resurrected if the committee shows a change of heart when it comes to the AAC, but even then, with our model expecting the best-ranked Group of Five team to come in 18th overall tomorrow night, it’s safe to write off the Group of Five yet again.
Our Model’s Expected Rankings
We published these for the first time last Tuesday, but here they are with another week of results feeding into them. We’ve included a “score” our model gives teams to represent how close it thinks they are to one another in the committee’s eyes. It’s scaled such that the best FBS team/résumé is at 100.0 and the worst (UMass) is at 0.0.
1. Ohio State (100.0)
2. Penn State (91.6)
3. Alabama (91.0)
4. Clemson (89.3)
5. LSU (89.2)
6. Baylor (82.3)
7. Auburn (81.2)
8. Oregon (80.5)
9. Michigan (79.9)
10. Oklahoma (79.3)
11. Wisconsin (79.3)
12. Minnesota (78.1)
13. Utah (77.5)
14. Georgia (75.2)
15. Florida (75.1)
16. Notre Dame (75.0)
17. Iowa (73.2)
18. Cincinnati (72.8)
19. Kansas State (71.5)
20. Navy (69.7)
21. SMU (68.7)
22. Texas A&M (67.8)
23. Indiana (67.2)
24. Oklahoma State (67.0)
25. UCF (66.7)
We’ll be updating our model after the rankings are released with early impressions of how much the committee is punishing and forgiving particular teams, so keep your eyes out for our recap on that front come Wednesday. The release of the first rankings may shake things up a lot more than this week did. If you’re using these to follow along, pay a lot of attention to that number on the right. Penn State being ranked fifth is more expected by our model than Ohio State being ranked second.
The Best Teams
For the last piece of insight, here are the teams the aggregate ratings our model employs view as the best teams in the country, along with the number of points by which each would be expected to lose to Ohio State on a neutral field. These don’t directly impact our model’s rankings or playoff probabilities, but they are what’s used for all the simulations of remaining games, so they indirectly impact our model in a big way.
1. Ohio State (0.0)
2. Alabama (2.4)
3. Clemson (5.8)
4. LSU (7.6)
5. Penn State (9.6)
6. Oklahoma (10.1)
7. Georgia (11.0)
8. Auburn (13.5)
9. Oregon (13.5)
10. Utah (14.0)
11. Wisconsin (14.4)
12. Michigan (15.1)
13. Florida (15.4)
14. UCF (18.3)
15. Washington (18.5)
16. Minnesota (19.8)
17. Iowa (19.9)
18. Iowa State (19.9)
19. Texas A&M (20.3)
20. Baylor (20.4)
21. Notre Dame (20.6)
22. Washington State (23.0)
23. Oklahoma State (23.1)
24. Memphis (23.3)
25. Texas (23.4)
26. Kansas State (24.0)
27. USC (24.2)
28. Michigan State (24.3)
29. Indiana (24.8)
30. TCU (24.8)
31. Cincinnati (24.9)
32. Missouri (25.1)
33. Navy (25.6)
34. North Dakota State (25.9)
