Beavers, Tigers, and Dukes: Last Night’s CFP Rankings, the Playoff Picture, and the New Year’s Six

College football’s postseason is taking shape day by day these days. Last night, we got new College Football Playoff rankings. Tonight, we’re reportedly (per Pete Thamel) going to get the NCAA’s decision on James Madison and Jacksonville State’s bowl eligibility.

To rehash, on that second point: James Madison’s likely to make a bowl no matter what. So is Jacksonville State. It’s unlikely there will be enough bowl-eligible teams, and that’s after considering possible 6–7 teams and even throwing a bone to Army and Navy if they finish 5–6 before playing one another (we don’t know what priority those two would get compared to JMU and JSU—we know the NCAA would prioritize 6–7 teams, or at least that they were doing that as of 2019, but we do not know about 5–6 Army and Navy). What these waivers would do, then:

• Convince the College Football Playoff committee to consider JMU for a ranking.
• Convince the Sun Belt to allow JMU to play in its conference championship.
• Make JMU eligible for the Group of Five’s spot in a New Year’s Six bowl (they would need to be the top-ranked Group of Five conference champion, so this is a product of the last two bullet points).
• Make Conference USA decide whether or not to allow Jacksonville State to play in its conference championship after New Mexico State appeared to clinch the second spot last weekend.

This doesn’t affect the playoff, but it’s looming large in the postseason picture. What does affect the playoff?

The committee really likes Oregon State.

To be clear, Oregon State’s very unlikely to make the College Football Playoff. But by ranking Oregon State ahead of Penn State and Mississippi, the committee showed the Beavers’ current résumé a lot of respect. With games remaining against Washington and Oregon, plus a hypothetical rematch with one of those two in the Pac-12 Championship (or a rematch with Arizona or Utah), the Beavers could finish the season with three more ranked wins, some of them over teams with Heisman contenders and national championship aspirations. They would need a lot of help, but the committee put them near the front of its list of two-loss teams, and unlike Missouri, who leads that list, Oregon State has a lot in front of them to accomplish.

More thoughts below. Let’s get to the rankings.

Our model reacts every week to the committee’s updated rankings. It looks at what it expects from the rankings, compares that list to the real rankings, and moves teams around using a variable we call FPA—Forgiveness/Punishment Adjustment. The most succinct explanation we’ve managed for FPA is that it measures how far from precedent the committee is deviating with respect to a given team, and that it sticks with teams in future weeks and in our model’s simulations of the rest of the season. Here are the rankings, how much FPA everyone is receiving, and how much new FPA the committee added this week. We included Liberty because they would be ranked if not for FPA.

Thoughts, observations, explanations:

• I’m not sure Georgia is really as far ahead of Ohio State as our model thinks. Our model doesn’t take away FPA once it’s granted unless the team in question falls out of the rankings or the committee shows that FPA is gone. Our thought with Georgia previously was that the committee was baking in some expected performances like Saturday’s, and that this was why Georgia was receiving so much benefit of the doubt. Now that they’ve proven it, a case could be made that there’s less benefit of the doubt to give, that they might be closer to Ohio State than our model believes, and that a 12–1 Georgia has a worse chance to make the playoff than our model thinks it has.
• The committee, as expected, seemed to use Ohio State’s win over Notre Dame as the tiebreaker over Michigan even after the Penn State victory became a common one, whether they said so explicitly or not. Fair is fair, but if Michigan beats Ohio State, they could conceivably jump Georgia by virtue of being artificially held back by their early schedule. That would probably only be a consideration for which teams play in which semifinal, but it’s worth thinking about, because playing Florida State or Oregon or Washington rather than the Big Ten or SEC Champion would be nice for whoever draws that straw.
• Missouri got the headlines (and you could take that ranking as the committee making a point of how Louisville isn’t going to get that Pitt loss completely washed away), but again, Oregon State is the story. Oregon State is now likelier, in our model’s simulations, to make the playoff than Louisville is. It’s only about a 1-in-40 chance, but in simulations in which Oregon State finishes 11–2, it’s nearly 50/50.
• Arizona’s rise is funny, because it comes after an ugly win over a pretty bad Colorado team. But! Good for them, and—I hate to say this—good for Iowa. Both of those are good teams playing hard and playing well, even if Iowa is doing it while self-handicapped.
• Oklahoma State is still ranked despite a 26-point loss to South Alabama and a 42-point loss to UCF, and I think we should remember that the next time Oklahoma State complains of SEC bias in their world.
• Liberty remains unranked, which is fine—the pit that is Conference USA makes it hard to evaluate their best teams’ performance in conventional ways, and the committee is very justified here in departing from our model’s judgment of precedent—but makes Tulane’s continued presence in the rankings (while the Green Wave scuffle) kind of funny. It feels like the committee feels an obligation to have somebody from the Group of Five ranked, to make bowl projections easier for those who do them based on where teams currently stand. Considering future opponents, what we know of the committee’s evaluation of Liberty and Tulane, and how the committee has ranked teams in the past, our model still has Liberty as, on average, the eventual top-ranked Group of Five champion, followed by Troy. If James Madison gets full bowl eligibility tonight, they would jump Liberty in final average ranking, but beating Troy in the Sun Belt Championship is a tough thing to do, so we probably still wouldn’t project them into the New Year’s Six spot ourselves just yet.

Updated bowl projections through last night’s games and rankings are below, with the probabilities below them (for archival purposes). A few notes:

• To err on the side of not overstating JMU and JSU’s bowl probabilities, we’re assuming they’ll be granted their waiver *after* the NCAA grants waivers to hypothetical 5–6 Army and Navy. We aren’t sure it would work out this way (we don’t think it will, in fact), but it seems like the safer side to err on, especially with JMU and JSU both *still* north of a 90% bowl probability even after making the change. We say all this to say: Navy’s projected to make a bowl game in these projections, but there are probably a few other teams who are likelier and aren’t included because we haven’t gotten to the APR rankings yet.
• The only changes today don’t come from the rankings, but from MACtion and from our adjustment to the Navy situation. With Northern Illinois beating Western Michigan, Northern Illinois is in. With Navy in, Rice is out (though, again, we’re very much not confident in our guess about Navy’s waiver priority at 5–6, and Rice has the better APR if Navy is 5–6 and Rice is 5–7 and APR’s the tiebreaker). Those switches make for some shuffles in the mid-major bowl matchups. All Power Five teams stay put, though.