I don’t want to focus too much on a small possibility, but it does lead us into more consequential points, so let’s start with this: Alabama has a chance to make this year’s College Football Playoff. Our model, which has never missed a playoff team in three years of practice and five years of backtesting before that, says the probability of the Tide being selected to play in the national semifinal is 1.4%, or roughly 1-in-70. That is a small, small, small possibility. But. It’s possible.
Why is it possible? Let’s talk about last night’s rankings.
The Top Four
Nothing to see here. These four were going to stay in order, they were going to stay in this particular order, we learned nothing. We still don’t know how close Michigan is to TCU, or how close Michigan is to Ohio State, or how close Ohio State is to Georgia, or how close TCU—at the bottom—is to LSU. Our model *thinks* there’s a big gap between the undefeated Big 12 team and the two-loss SEC team who lost by 27 at home and couldn’t beat Florida State in their second home, but we don’t know that for sure.
The Question With LSU, the Answer About USC
What we do know is that LSU is, right now, the committee’s fifth-best candidate. This means something, but not so much for LSU as for USC.
For LSU, we still think that beating Texas A&M and then—more importantly—beating Georgia to win the SEC would probably be enough to crack the field provided a few probable things break the Tigers’ way. We don’t know what happens in one specific scenario therein, though, one in which TCU wins out and the Ohio State/Michigan game is very, very close. Does hypothetical 11-2 LSU jump the loser? That’s up in the air.
For USC, we got an important data point, and that is this: The Trojans’ eleven performances so far do not measure up to LSU’s in the eyes of the committee. This doesn’t bar USC from the playoff by any means—beating Notre Dame would matter, winning the Pac-12 would really matter—but it shows that their ceiling is fairly low. Using our model’s scoresheet, the best guess at the gap between USC and TCU is that it’s larger than the gap between TCU and Georgia right now. This matters in the same scenario in which we ask the LSU question: If the Michigan/Ohio State game is very, very close and TCU wins out, locking themselves in and making the question (in the USC version) whether 12-1 USC or 11-1 Michigan/Ohio State is better…the hill is tough to climb. At this moment, our model leans towards the committee choosing that Michigan/Ohio State loser.
Alabama Might Be Alive, Clemson Might Be Dead
This is a good way to badly use data. The fact Alabama’s chance is higher than expected—1.4%, when we thought it was just 0.6% at this time yesterday—and the fact Clemson’s is lower—18.3%, compared to 26.8%–obscures how 18.3% is still thirteen times larger than 1.4%.
But.
Clemson’s still currently behind Alabama, and in Clemson’s version of the USC/LSU question from above, this is not a good sign for the Tigers. Just as USC sitting behind LSU shows that the ceiling on the Trojans is only so high, Clemson occupying the place it does signifies that the Tigers are deep in it. Conference championships matter a ton—that’s why Clemson’s thirteen times likelier to make this thing than Alabama is—but a hole is a hole, and this is a big one. Also, there’s a chance the committee is in most scenarios done with Clemson. It’s possible Clemson has simply performed too poorly for the committee to consider them in any scenario outside of mass chaos. Because behind Clemson comes Oregon, and Oregon’s already got those two losses.
For Alabama, there’s a friendly little element here where we don’t know how far ahead of Clemson the Tide currently sit. Our model assumes the gap to be as small as possible, but it could be larger, and while we really doubt a 10-2 Alabama will stay ahead of a 12-1 Clemson, there are worlds in which that is what we get.
Last thing here: Clemson’s ranking is a good thing for TCU, whose shot as a 12-1 team, conference title or not, isn’t quite zero. But it’s mostly just a bad thing for Clemson.
Oregon, But Mostly Washington
Oregon’s chances of making the playoffs jumped a lot as a result of the rankings, both in our model’s eyes and to the human observer. On the model side, it’s rather simple: Oregon’s ranked ahead of Tennessee already, and Tennessee’s got a solid team sheet! They beat Alabama! They smoked LSU! If Oregon’s work to date is better than Tennessee’s, Oregon’s looking lively.
On the human side, it’s very heartening for the Ducks to see Washington ranked so highly this week. The fact the Huskies are 13th in the eyes of the committee means Oregon’s losses are better than those of LSU: The blowout loss came against a higher-ranked team (Georgia > Tennessee), the close loss came against a higher-ranked team (Washington > Florida State), this is all good for an Oregon team who, if USC beats Notre Dame, will get a chance to not only win a Power Five championship but notch a top-ten win. Oregon isn’t going to jump an 11-2 LSU, and they aren’t going to pass 13-0 TCU, but jumping 12-1 Clemson or 12-1 TCU or 11-1 Michigan is within the realm of possibility. There’s a lot of work to do to get to 11-2, and it starts with a tough one against Oregon State this weekend, but Oregon is alive. Very alive.
Do Tennessee and Penn State Have a Chance?
Our model has Tennessee at 0.3% playoff-likely and Penn State at 0.1% playoff-likely. Are these real possibilities? Let’s try to see, and let’s use the premise that if we can get Penn State in the playoff, we can get Tennessee in. What, then, happens if…
- Alabama loses to Auburn and Tennessee loses to Vanderbilt, making Penn State the lone prominent 10-2 team, and…
- TCU loses to Iowa State and then loses to Kansas State or Texas, giving the Big 12 at best a three-loss champion, and…
- South Carolina beats Clemson, putting the Tigers into an even bigger hole, and…
- Notre Dame beats USC badly, but USC turns around and wins the Pac-12, and…
- Georgia smokes LSU, or LSU just loses to Texas A&M.
Does that get Penn State in the playoff? I think it might. Is it 1-in-1,000 likely? I’m not sure it is, but that isn’t that many results, so maybe I’m wrong. Anyway, yes, I think Penn State could theoretically still make the playoff. I don’t think our model’s totally nuts on that. But it’s fourteen times less likely than Alabama making it, and we’re already not *really* thinking Bama will crack the top four.
The AAC Thing
What the fuck?
The rankings are not a horse race the way the AP Poll often is. They are not some exercise in starting with the last week’s rankings and then adjusting them to reflect the wins and losses since last week. They’re a comprehensive new top 25 produced every single week. Or at least, that’s how it’s supposed to work. Either way, though, how is UCF ranked 22nd after losing to Navy?
One interpetation of this is that the committee did the AAC a favor. The AAC, hopefully infamously (they should get all the shit in the world for this), built tiebreakers that don’t work. They contradict. Incomprehensibly stupid. Entirely unnecessary. But, here the AAC is.
In a comic twist, this isn’t how the committee helped the AAC. The committee choosing to rank UCF actually made the theoretical contradiction real, pulling it to the surface and casting light all over it for anyone who cares enough to look (Tulane should care to look). What should have happened is what we outlined Monday, which is that UCF should have finished unranked and the AAC’s chosen computer rankings should have decided any tiebreakers that couldn’t be broken by head-to-head results. Instead, UCF was ranked, and the AAC jumped on the opportunity, declaring last night that as long as UCF wins against South Florida, they get to play in the AAC Championship, even though that contradicts one literal reading of their tiebreaker rules (or the ones we can find, anyway—adding to the ridiculousness of this, you have to dig back to 2020 to find public tiebreakers from this conference).
How does this help the AAC? Well, it gave them the opportunity to make their declaration, using the other literal reading of their tiebreaker rules to keep the decision out of the computers’ hands. It also saved them from what would have likely been a back-to-back rematch between Tulane and Cincinnati played in the exact same stadium as this weekend’s game. That would have been fair and just, but it would have felt kind of dumb.
I’m not sure this interpretation—that the committee wanted to do the AAC a favor—is correct. I also don’t think a large part of the committee didn’t notice UCF losing to Navy, though that could be possible, and would be both hilarious and bad. What I’d guess is going on here is that the committee wanted to make clear that the AAC champion is going to be the highest-ranked Group of Five champ no matter what happens from here. Sorry, Boise State. Sorry, Coastal Carolina and Troy. It’s the AAC champion. Bowls, plan accordingly.
FPA
The full FPA list (FPA measures how far the committee deviates from precedent with its rankings), from most forgiveness to most punishment:
- UCF: +9.0
- LSU: +5.0
- Oregon: +3.9
- Cincinnati: +3.7
- Tulane: +3.1
- Kansas State: +2.4
- UCLA: +2.0
- TCU: +1.8
- Washington: +1.7
- Mississippi: +0.6
- Notre Dame: +0.4
- Georgia: 0.0
- UNC: -0.2
- Louisville: -0.4
- Ohio State: -0.7
- Clemson: -0.9
- Michigan: -1.1
- Mississippi State: -1.1
- Iowa: -1.3
- USC: -1.5
- Oregon State: -1.7
- Utah: -1.8
- Wake Forest: -1.8
- Florida State: -2.0
- Alabama: -2.8
- Penn State: -2.9
- Tennessee: -3.7
- Texas: -9.6
Thoughts on this:
The committee is holding pretty close to precedent at the top, with the notable exceptions of Alabama, LSU, Tennessee, and Oregon. With Tennessee, I wouldn’t be surprised if this is the Hendon Hooker piece. We don’t really have a good way to account for that in our model, because it hasn’t been tested in precedent. With Oregon, I think there’s a bit of recency and a bit of wanting to reward the Ducks for going to Atlanta to play Georgia. That was a terrible loss even if it did come against the top-ranked team. They lost by more than forty points. But, if the committee takes the institutionalist approach and wants to encourage teams to play big nonconference games, it has an interest in turning a blind eye to forty-point losses to Georgia in Atlanta.
The LSU/Alabama thing is likely head-to-head, but then with LSU ahead of USC…I’m not sure what to think. One interpretation is that in a shift, margin doesn’t matter as much as it has before. Alabama has margin loudly on its side: It’s won big, it’s lost small. LSU does not: It’s won small, it’s lost big. Maybe, in a similar vein to Oregon, the committee doesn’t want to punish LSU for scheduling Florida State at the superdome? I’m not sure. It’s tricky.
The AAC love as a whole isn’t as bad when you consider that our model assumes every Group of Five program gets a gigantic punishment right off the bat. Some of this is fair—the model is big into looking at things like point differential, and with only three or four nonconference games per team, this can twist in odd ways if a team’s schedule is easy enough. Some of it, though, discounts these teams more than they deserve, and the wholesale embrace of the AAC really signifies that to the committee, the AAC is in the middle ground between the Power Five and the rest of the Group of Five. That tracks.
I don’t know why the committee hates Texas. It’s comic. Texas has a couple bad losses, sure, but this is absurd. It could be changing the playoff race, too, and it could really change it if Texas ends up being TCU’s opponent in the Big 12 Championship. Beating Texas twice would be a great accomplishment. It might not be viewed as such.
FPA-Less Rankings
What would the top 25 look like if precedent were winning out? It would be similar:
1. Georgia
2. Ohio State
3. Michigan
4. TCU
5. Alabama
6. USC
7. Tennessee
8. Clemson
9. Penn State
10. LSU
11. Oregon
12. Utah
13. Texas
14. Florida State
15. Kansas State
16. Washington
17. Notre Dame
18. Oregon State
19. UNC
20. Mississippi
21. UCLA
22. Wake Forest
23. Iowa
24. Mississippi State
25. Tulane
The only meaningful difference, then, is LSU. And if they lose to Georgia, which is likely, that won’t matter.
