Another set of College Football Playoff Rankings are upon us, released last night to much uproar in the spheres of people who like to get upset about things.
I’ll confess. The Oregon ranking is pissing me off. They’ve got me there. Most of the time, you know the rankings will settle themselves out, but with Oregon, there’s a chance they won’t, and that’s frustrating, because the Ducks would be a three-score underdog against Georgia right now, and there’s a clear reason to not rank them fourth. A clear reason named Stanford.
Overall, though, the rankings aren’t the most meaningful thing in the world. They’ll likely settle themselves out, at least to an extent, and as the committee showed with Michigan and Michigan State, things can flip rather arbitrarily. Still, they’re a useful marker of where teams are at, and they’re a useful marker of how much benefit of the doubt or absence thereof the committee is giving teams, which is something our model is specifically designed to do through FPA.
The details on how our model, including FPA, works are here, but the quick summary is that the model’s formula is based on the seven years of precedent the committee has established, with FPA measuring where, by how much, and in what direction the committee is deviating from precedent. At this point, multiple sets of rankings in, FPA is measuring what’s been seen from multiple sets of rankings, not just the latest ones. Here’s what it makes of where we’re at:
Ranking | Team | Ranking w/o FPA | Est. Ranking Score | FPA |
1 | Georgia | 1 | 100.0 | 0.0 |
2 | Alabama | 3 | 93.1 | 2.0 |
3 | Oregon | 12 | 91.7 | 7.8 |
4 | Ohio State | 2 | 90.4 | -1.0 |
5 | Cincinnati | 7 | 88.9 | 2.5 |
6 | Michigan | 8 | 88.2 | 2.5 |
7 | Michigan State | 5 | 87.3 | -0.8 |
8 | Oklahoma | 4 | 87.2 | -1.8 |
9 | Notre Dame | 6 | 87.1 | -0.2 |
10 | Oklahoma State | 9 | 86.1 | 1.0 |
11 | Texas A&M | 17 | 79.7 | 3.2 |
12 | Wake Forest | 11 | 79.7 | -4.7 |
13 | Baylor | 14 | 77.7 | -0.5 |
14 | BYU | 19 | 76.9 | 0.8 |
15 | Mississippi | 21 | 76.1 | 0.7 |
16 | NC State | 18 | 75.8 | -0.7 |
17 | Auburn | 28 | 75.7 | 4.3 |
18 | Wisconsin | 15 | 75.5 | -2.5 |
19 | Purdue | 16 | 75.2 | -1.7 |
20 | Iowa | 13 | 75.1 | -3.3 |
21 | Pittsburgh | 20 | 74.3 | -1.2 |
22 | San Diego State | 23 | 74.0 | 0.5 |
23 | UTSA | 10 | 73.8 | -11.2 |
24 | Utah | 25 | 73.6 | 1.0 |
25 | Arkansas | 31 | 73.5 | 4.3 |
NR | Penn State* | 27 | 73.2 | 1.0 |
NR | Houston | 22 | 73.2 | -0.5 |
*Some of Penn State’s FPA comes from the Kelly Bryant Rule™, explained in the link above. With that FPA and no other FPA for anybody, Penn State would land between Mississippi and Houston.
Some notes:
Conference (and Division) Matters
Or at least, it seems to. The three ACC teams subject to FPA each are receiving a comparable ding against them. So are the three Big Ten West teams. Five of the six SEC teams receive a bump, with the lone exception being Georgia, who might be getting one we just can’t see because they’re so far ahead of the field that their place isn’t questioned.
This does not mean the committee is doing anything we’d characterize as “wrong.” Our model is the best one we know of, but it isn’t perfect, and one shortcoming we’re aware of is the element of it which evaluates teams in silos, to an extent. Because so little of the college football season is nonconference play, and because it comes at the beginning, and because so little of nonconference play is between Power Five teams, there aren’t great connecting data points between leagues. For the most part, our model views the ACC and SEC as comparably competitive, and while that doesn’t mean as much as it might sound like it means (the gap isn’t so big as to shake our model up very much, as our model’s results demonstrate), it’s a shortcoming. It’s justifiable to perceive the SEC to be better than the ACC. I think we can all acknowledge that the SEC is better than the ACC. Similarly, I think we can acknowledge that the Big Ten East features tougher top-line games than the Big Ten West. The committee baking this into their rankings is fair, and I’d venture this is one of the two biggest current sources of FPA, even accounting for Cincinnati’s sizable positive FPA and UTSA’s massive negative FPA, as our model, based on precedent, lumps Group of Five teams in as being treated equally to one another despite the AAC being significantly better than Conference USA.
Let’s Harp on Oregon Some More
I said that conference/division adjustments are probably one of the two biggest current sources of FPA. The other? Whatever is possessing these people to rank Oregon ahead of Ohio State. And Cincinnati. And Michigan. And Michigan State. And Oklahoma. And Notre Dame. And Oklahoma State. And…I mean, our model has UTSA and Wake Forest ahead of Oregon right now, but go back to the section on conference-by-conference FPA for that argument.
It’s fair for the best win in the country—and that win in Columbus is the best win in the country, to date—to mean something. But for it to mean enough to wipe out the entire rest of Oregon’s résumé, one headlined by a loss at currently 3-6 Stanford and marked by zero ranked wins aside from the Ohio State one (only twice has Oregon beaten an FBS team by more than ten this year, something even Oklahoma State has done more often)…that’s bad. That’s dumb. And this is why it matters:
Throughout the last seven years, things have tended to work themselves out, for the most part. There have been very few legitimate complaints about the final rankings. But it isn’t guaranteed that they’ll always work themselves out. And while Oregon’s résumé would certainly benefit from beating Utah twice, something they’ll likely have to do if they’re going to still be in the playoff conversation, those wins would each need to come by 28 points, if added in to current résumés, to get Oregon to fourth without their current FPA. Realistically, they may need to be bigger still, because the other teams in the mix are playing much better opponents than Utah themselves, meaning their own résumés may well improve substantially.
Oregon will probably lose. They’ll probably lose, and this will probably be fine, with someone maybe getting screwed out of a New Year’s Six bowl as some of the FPA holds but no program suffering an existential setback from it. But on the off-chance Oregon wins out…and on the strong chance within that off-chance that Oregon doesn’t blow Utah out both times…we could finally see the committee undisputedly fail at their core purpose.
There Are Two Packs; Georgia Is in Neither
Regardless of whether you rank Oregon 3rd or 11th, they’re part of the first pack, the one that spans from Alabama to Oklahoma State (or Oregon). The gap between Alabama and Georgia in our estimated ranking score is about the same as that between Alabama and Oklahoma State, that between Oklahoma State and Texas A&M, and that between Texas A&M and Clemson, who doesn’t show up on our table but would be 28th were our table 28 rows long. There’s Georgia, then there’s a pack of 1-loss Power Fives and Cincinnati, then there’s another pack, and it’s thick. Again, to jump from 28th to 11th, Clemson would need to do something equally impressive to what it would take for Texas A&M to jump from 11th to 10th. The same is true for Oklahoma State jumping to 2nd versus Alabama jumping to 1st. This is important to understand while gaming out playoff possibilities. Texas A&M and beyond might really be too far back at this point. Georgia is well upwards of 90% likely to make the field. Those teams between them are in a dogfight, and it’s often being fought by proxy.
***
We’ll have more on the games this week for you all…tomorrow, I believe, but what those in the mix are up to:
Michigan goes to Penn State.
Oklahoma visits Baylor.
Georgia plays at Tennessee.
Ohio State hosts Purdue.
Michigan State hosts Maryland.
Texas A&M visits Mississippi.
Notre Dame visits Virginia.
NC State visits Wake Forest.
Oklahoma State hosts TCU.
Oregon hosts Washington State.
Cincinnati visits South Florida.
Alabama is idle (hosts New Mexico State).