Hello.
I am very sorry about our continued absence from this website. Our plan had been to at least keep the college football model’s outputs updating after the baby was born, but we did not do that. The baby is happy and healthy, and we his parents are getting healthier by the day, but it’s been a bit of a weird ride these last four weeks, and as our regular readers are used to us doing, we dropped the ball.
We’ve been working on posts, and once we have two days stored up (we’re trying to better insulate our content stream from external shocks, protecting against that ball-dropping habit), we’ll start publishing regularly again. I’m really hoping that happens this week. In the meantime, though, some thoughts on this morning’s playoff selection show before it happens, in case anyone’s visiting the site and wondering either 1) what our College Football Playoff selection model says or 2) what I think, as a guy who maintains a College Football Playoff selection model.
SMU vs. Alabama
The model says Alabama. Now, the model would not say Alabama if the model didn’t know what the committee had already told us with its previous rankings. But based on the rankings so far, the model says Alabama.
It’s very close. SMU’s losses are each better than Alabama’s two worse defeats, at least the way our model measures them, and crucially, SMU has one fewer loss than Alabama. But between Alabama’s better wins, Alabama’s more impressive performance (as measured by adjusted point differential, our metric for how dominant teams have been), and what the committee has already said about each team’s résumé through twelve games, our model believes the committee will choose Alabama.
I don’t have a particularly strong opinion here when it comes to what the committee should do. I understand the argument that teams shouldn’t be penalized for losing in their conference championship game, because it feels like they’re being penalized for making their conference championship game, but I don’t think it’s all that strong an argument on its own merits. After all: Teams can be rewarded for their conference championship game performance, and the conference championship is part of most teams’ schedules, at least in their ideal world. SMU played Clemson in the ACC Championship. Alabama played Western Kentucky in nonconference play. I don’t have a problem with both of those results mattering.
Alabama’s the better team, if you want to go that route. SP+, FPI, Movelor…they all agree. Alabama’s better at football, based on what they’ve shown on the field this season. Alabama also is probably more accomplished. SOR, which I believe is FPI-based, says it’s harder to go 9–3 against Alabama’s schedule than it is to go 11–2 against SMU’s schedule. But it’s close enough that I wouldn’t have a problem with the committee viewing it differently.
What will the committee do? I tend to think they’ll take SMU. The committee doesn’t like pissing people off. Even last year, when the committee pissed a lot of people off, an argument can be made that it only did that to avoid pissing even more people off. It would have been hard to justify a playoff without including the 12–1 champion of the conference who dominated the last four playoffs. The SEC might complain this time around, but I don’t think the committee has to worry about neutral parties feeling like Alabama got screwed, or caring that Alabama got screwed if they do feel like Alabama got screwed. The committee definitely doesn’t have to worry about this playoff feeling illegitimate because Alabama isn’t in it.
The committee mostly consists of people whose daytime jobs involve being extremely sensitive to public opinion. My guess—only a guess—is that they’ll avoid causing themselves a headache.
One thing people are overlooking in this debate is that it isn’t going to exactly go down as “SMU vs. Alabama” in the committee room. Or at least, it shouldn’t. Theoretically, the committee should be ranking SMU, Indiana, Boise State, Alabama, Miami, Mississippi, South Carolina, Arizona State, and Clemson in some order. That’s the process they say they follow. For our purposes, we can exclude Miami, Mississippi, and South Carolina from that mix—the committee has said that those teams’ cake is baked, and if ASU’s behind Boise State or Alabama, well, it doesn’t affect anything—but the fact this is a broader debate than simply “SMU vs. Alabama” could prove a complicating factor.
So…
SMU vs. Indiana vs. Boise State vs. Alabama vs. Arizona State vs. Clemson
These six teams plus the three we’ve excluded will, together, be ranked between 8th and 16th. Clemson will, at the very least, pass Iowa State.
In what order do these six go?
Our model says Boise State will be 8th, Indiana will be 9th, Arizona State will be 10th, Alabama will be 11th, SMU will be 13th, and Clemson will be 14th, with Miami wedged in there between Alabama and SMU.
Were this still the four-team era, I’d be inclined to trust our model on everything but SMU vs. Clemson, where I’d be inclined to expect the committee to go by head-to-head. This is not the four-team era, though. There are two big differences at work here:
The first is that the playoff’s twelve teams big, and that’s brought with it this current of “don’t punish teams who lost in their conference championship.”
The second is that the Big Ten and SEC are better than they were in the four-team era, relative to the rest of the power conferences, and that the power conference designation is no longer a formal one. A huge question about the committee’s rankings tomorrow is how differently conference championships will matter. In the four-team era, winning the Mountain West meant next to nothing, while winning the Big 12 and ACC meant a lot. Our model assumes this will still be the case. Will it? Or will the Big 12 and ACC titles mean less? And could the Mountain West title mean more?
With Boise State in the lead, I don’t think we need to worry about the Mountain West piece. I think the bigger question is how much winning the Big 12 matters for ASU. Do they really jump both SMU and Alabama, plus those other three teams in between? Most importantly, do they stay ahead of Clemson even though Clemson beat a team in SMU of whom the committee thinks more highly than Iowa State?
With ASU’s victory so dominant, and with our model—which, again, spits at the MWC title and throws a parade for the Big 12 champ—expecting ASU to be ranked between Boise State and Clemson, this probably doesn’t matter. Boise State will in all likelihood be the 3-seed, and Arizona State will in all likelihood be the 4-seed. The only way in which ASU matters is if ASU gets caught in the middle of Alabama vs. SMU and throws off some votes, the way Zack Britton helped swing the 2016 AL Cy Young from Justin Verlander to Rick Porcello.
More likely to get caught in the middle of Alabama vs. SMU is Clemson. Clemson just beat SMU head-to-head and won the pair’s conference title. Clemson’s only losses came to Georgia and to two teams right around the edges of the committee’s top 25. (We can assume Louisville is close to the top 25.) Will the committee get twisted into some logic where Clemson needs to be ahead of SMU because of head-to-head, but Clemson can’t be ahead of Alabama because that’s ridiculous? If so, it’s bad news for SMU.
I don’t think the committee will handle it this way. I think they’ll either 1) do the ridiculous thing and rank Clemson ahead of Alabama, 2) put Alabama ahead of both and kick SMU to the curb independently of their protocols and process, in a binary Alabama vs. SMU debate, or 3) keep Alabama between SMU and Clemson and avoid those annoying people who say, “HOW IS TEAM A ONE SPOT BEHIND TEAM B WHEN TEAM A BEAT TEAM B HEAD TO HEAD??” as though all other criteria disappear once two teams are in close enough proximity for magnets to work.
If the committee goes Route 2, we’ll never know they did it. We may think they got twisted into the logic described above, but we won’t know for sure that they did. But again, I kind of think they’re going to keep SMU ahead of Alabama and Alabama ahead of Clemson, resulting in Clemson being the 12-seed. I’m more interested in whether they go to the trouble of dropping SMU behind Indiana. They should—it was a weird ranking last week which put SMU ahead of the Hoosiers—but I don’t know that they will. I don’t know whether SMU or Indiana will be the 10-seed.
This leaves us with our top four seeds most likely settled (Oregon, Georgia, Boise State, Arizona State, in order), and our 10–12-seeds at least opined about. Moving on, then, to the 5-seed through the 9-seed:
Texas vs. Penn State vs. Notre Dame vs. Ohio State vs. Tennessee
Our model expects Notre Dame to get the 5-seed, Texas to get the 6-seed, Ohio State to get the 7-seed, Penn State to get the 8-seed, and Tennessee to get the 9-seed. It was waffling between Penn State and Tennessee as Penn State fell behind Oregon and drew back closer. That one’s tight, in our model’s eyes.
Again, in the four-team era, this would make sense. Oregon would be in. Notre Dame would probably be in. Georgia would probably be in. The fourth spot would probably come down to Texas, Boise State, and Indiana, with an Ohio State case made and an Arizona State case made and probably a Tennessee case made.
In the 12-team era, though, this is a surprising list. It’s easy to picture it happening this way—if the committee does move teams who lost, it would make more sense to put Ohio State ahead of Penn State than the other way around, given Ohio State’s head-to-head win and better performances against common opponents (among other, more holistic things)—but it’s not exactly what I expect to happen. I tend to think Texas will be the 5-seed because they hung so close to Georgia and in this instance, the committee has reasons to treat that as less of a loss. I tend to think Penn State will be the 6-seed because of a similar phenomenon regarding their performance against Oregon. I tend to think Notre Dame will therefore be the 7-seed. I don’t have a problem with any of that. It’s really coming down to hair-splitting. The only thing I feel confident about beyond the top four seeds is that I’m pretty sure Tennessee will be the 9-seed.
**
So, two brackets. The first going with our model. The second going with my intuition.
Model
First Round: 9-seed Tennessee at 8-seed Penn State
First Round: 10-seed Indiana at 7-seed Ohio State
First Round: 11-seed Alabama at 6-seed Texas
First Round: 12-seed Clemson at 5-seed Notre Dame
Rose Bowl: 1-seed Oregon vs. PSU/Tenn
Sugar Bowl: 2-seed Georgia vs. OSU/IU
Fiesta Bowl: 3-seed Boise State vs. Texas/Bama
Peach Bowl: 4-seed Arizona State vs. ND/Clemson
Intuition
First Round: 9-seed Tennessee at 8-seed Ohio State
First Round: 10-seed Indiana at 7-seed Notre Dame
First Round: 11-seed SMU at 6-seed Penn State
First Round: 12-seed Clemson at 5-seed Texas
Rose Bowl: 1-seed Oregon vs. OSU/Tenn
Sugar Bowl: 2-seed Georgia vs. ND/IU
Fiesta Bowl: 3-seed Boise State vs. PSU/SMU
Peach Bowl: 4-seed Arizona State vs. Texas/Clemson
Maybe I’m getting too cute here, but when forcing myself to make a guess, I think the committee will want to move somebody who lost last night, to avoid setting a precedent that they can’t move anyone. I also think they’ll want to avoid an Ohio State/Indiana rematch, even if they deny considering such things. I don’t think they’d go so far as to push Arizona State ahead of Boise State for the sake of a hometown Fiesta Bowl, but I do think they might wiggle Indiana and SMU in order to avoid Ohio State and Indiana meeting again in the first round. Plus, this way they’d get an intrastate matchup in South Bend.
**
Again, we’re hoping to have more about the playoff itself come tomorrow, when the playoff matchups are set. After that, there’s plenty of college basketball to get caught up on, and the FCS playoffs are getting into their good part. We hope to enjoy all of it together very soon. Thank you for your patience, and for coming back even when we’ve been away so long. It means a lot.
Bark.
