College Football Morning: Favorites, Contenders, and Factors
A three-category national championship picture.
One of our preferred frameworks at The Barking Crow, when it comes to capturing the current championship picture in a given sport, is to separate teams into three categories, with an implicit fourth beyond the three. We have our favorites, the teams likeliest to win it all. We have our contenders, teams who could believably move into that “favorite” tier. We have our factors, teams who aren’t serious threats to win a title, but who are impacting or will impact the race. The implicit fourth category is all the teams beyond the “factors,” your Texas Techs and Michigan States and Louisianas Monroe. In Major League Baseball, those are often the bad teams. In college football, it’s a different situation.
We’ve spent a lot of time this year focused on who’ll make the College Football Playoff, and understandably so. Since the playoff was first instituted a decade ago, making the playoff has become its own badge of honor. Even in the four-team era, playoff selection was an achievable goal for most FBS programs. (Yes, most. That’s a topic for another day.) With the expansion to the twelve-team field, social media staffs and recruiting coordinators are salivating, as is whoever makes those big letters which hang in football stadiums to honor championships and retired numbers. Only 15 programs have ever made the College Football Playoff. That number will expand to something like 19 or 20 this year. It’ll remain an exclusive club, but it’s easier ground than ever in which to plant one’s flag. Combine this attainability with the novelty of the new format, and of course college football fans are talking mainly about the playoff.
For a minute, though, let’s discuss the national championship. As always, probabilities come from our model.
Favorites (54.8% National Championship Probability)
Our “favorites” category is for the teams it wouldn’t be surprising to see win it all. In college football, this means not only being in position and being good enough, but having the talent reserves which make us believe the team can remain good enough as these next nearly three months of football play out. Who’s in position? Who has the athletes for January? Three teams:
Ohio State (20.7%)
Texas (17.9%)
Georgia (16.1%)
If you’re the betting sort, or even just a probabilistic thinker, you’ve noticed that our model says there’s only a little better than a 50/50 chance that one of these three teams does ultimately win the national title. Yet here we are, refusing to grant even Oregon access to this tier. What gives?
I don’t have a perfect, objective answer for you. These are subjective categories. Texas has nearly twice as many former five-stars as Oregon. Ohio State and Georgia have even more. I don’t fully trust our model, strong as I believe it is. We’ve wanted to better account for talent in our model for years, and we’ve recently begun wondering whether we should build in some assumptions about how teams will weather the storms as a season progresses. Our model is very simple. I do think it could be underestimating these three teams’ championship chances. I don’t know the margin by which it might be doing that.
In November of 2020, Notre Dame beat Clemson, scoring 26 offensive points in regulation. Six weeks later, Clemson pounded Notre Dame in the ACC Championship, holding the Irish to ten. Notre Dame went from 6.6 yards per play against Clemson to 4.5 yards per play in the span of a month and a half. James Skalski missed the first meeting, but I have a hard time believing that was the full story. Whether it’s how they train physically or how they approach x’s and o’s or simply a matter of depth, these truly elite teams, the serious national title threats, seem to play their best in December and January.
Maybe this is selection bias. Of course national champions play their best in the national championship. But I’d like to do more research on whether talent or a program having “been there before” can predict late-season power ratings separation. When Movelor, our model’s rating system, says Georgia is 0.2 points worse than Oregon right now, I trust it. But Movelor isn’t saying Oregon’s going to be the better team in January, and neither am I. The reason the probability’s only 54.8% is that Movelor also isn’t saying Oregon’s going to be the worse team in January. I am.
So: These three teams plus Alabama are the most talented in the country. Alabama’s in a bad position. These three teams are in a good position. I think this is our serious list.
Contenders (30.2% National Championship Probability)
Going down into the contenders, then, those teams we think could move into favorite status:
Oregon (14.9%)
Notre Dame (7.7%)
Alabama (5.1%)
Clemson (2.6%)
Why these four? Well, they’re four of the next five in our model’s championship probabilities, but that comes from somewhere. Oregon’s playing slightly better than Georgia right now. Notre Dame’s playing better than Alabama. Alabama and Clemson both have more talent than Oregon. (Oklahoma and LSU are the other teams ahead of ND in the composite this year, where ND lacks five-stars but has the most four-stars in the country.)
Clemson’s defense is concerning, but it isn’t a glaring flaw, and with the Tigers not often tested by their ACC schedule, it’s likelier that their Movelor rating is artificially deflated than it is that Tennessee’s is too low. Tennessee has to play their best ball right now. Clemson doesn’t. We’re theorizing that this might be showing up. More than that, though, this is about talent, and about the uncertainty introduced by a longer single-elimination format. We don’t think Clemson’s ceiling is national championship high, but we’re wary of a world where Ohio State, Texas, and Georgia somehow end up on the same side of the bracket. There’s also more reason to believe in Cade Klubnik in a big game scenario than in Nico Iamaleava.
Again, though, this is subjective. These categories are a subjective attempt to account for the known imperfections of our model.
Factors (15.1% National Championship Probability)
The factors? The teams who’ll impact the championship but we really don’t think will win it? (Again: Maybe we’re too confident and our model is telling us something.) Let’s do those in two groups:
Tennessee (4.1%)
Mississippi (2.3%)
LSU (2.5%)
Penn State (2.4%)
Kansas State (1.3%)
Miami (0.8%)
This first group has something significant in common, which is that everyone’s at least in Movelor’s top 14. We’ve listed them in Movelor order, to signify that these teams are probably the best ones not listed yet and the teams we’d peg as likeliest to climb up into the previous tier. Still, they each have such a significant red flag that it’s hard to include any in that “Contenders” category. For Tennessee, it’s depth and Iamaleava, who isn’t bad but is severely unpolished. For Mississippi and LSU, it’s the hole they’ve already dug. For Kansas State, it’s their inopportune brand of inconsistency and their major, major, major talent deficit. For Miami, it’s the defense. For Penn State, it’s the perennial bridesmaid identity, one driven by the program’s failure to win a single huge game in the last eight years despite plenty upon plenty of opportunities.
These teams are on our radar, and in the four-team era, we’d probably stop with them plus a handful from this next group. They’re significant characters, and they’re good teams, but they’re not great teams, and only Tennessee is in great position among them.
The next group:
BYU (0.3%)
Iowa State (0.4%)
Boise State (0.1%)
SMU (0.5%)
Indiana (0.1%)
UNLV (0.0%)
Army (0.0%)
Pitt (0.0%)
Tulane (0.0%)
Texas A&M (0.1%)
Navy (0.0%)
We’ve listed these in order of playoff probability, and it’s a varied list. Texas A&M’s the most talented of the bunch. SMU’s probably the best team. BYU and Iowa State have both the best playoff shot and the best chance at securing a bye. Ultimately, this is everyone with at least a 1-in-20 playoff chance, per our model. The first team out? Colorado. Maybe that’s unfair, but we had to draw the line somewhere, and Navy’s almost twice as likely as the Buffaloes to make the 12-team field. Include Colorado, and it’s hard to exclude Western Kentucky. Western Kentucky is not a factor in the national championship picture. At least Navy has a chance to upend Notre Dame.
Where does all this leave us? It’s a 24-team picture. There are, to be fair, there are ten simulations from our latest ten thousand in which someone outside these 24 wins the title. Five of those belong to Mizzou, where Movelor doesn’t know the Tigers lost Brady Cook for this weekend. The other five simulations belong to Virginia Tech, South Carolina, Louisville, Wisconsin, and…Michigan, about whom we’ve spoken enough this week.
The list should be culled more tonight and tomorrow. But entering the weekend, it’s a 24-team picture.
**
Enjoying College Football Morning? Subscribe to receive it every day in your inbox.
**
How about Pitt?
In yesterday’s Week 9 preview, we said the Panthers would be 10% playoff-likely if they beat Syracuse. The number is 16% instead. What happened?
Well, Pitt won huge. 10% was the number in their average victory, one by 9.9 points. (Movelor had the spread at 6.1 points, but in simulations where Pitt won, 9.9 was the average margin.) Instead, Pitt won by four touchdowns and four extra points, aided by three interceptions returned for touchdowns.
Plenty of people who know college football data well will tell us our model shouldn’t put so much stock in this particular 28-point win because returning three picks for touchdowns isn’t replicable. This is fair. Movelor’s simple, and we acknowledge that. I’d imagine Movelor’s is a bigger reaction than SP+ or FPI will have. But Movelor does pretty well nonetheless, and I’m not too worried about it accidentally overestimating a team who’s now 6–1 against the Vegas spread this year. It might overestimate this 28-point win, but it’ll underestimate others. Most of this sort of imperfection comes out in the wash.
Exciting stuff on tap tonight, more in Las Vegas than Los Angeles. Louisville/Boston College should be interesting as well. Louisville’s the one team in the ACC who’s played a difficult schedule. Have the losses been enough to break them? We’ll see you on the other side.
**
Weekend resources:
Movelor’s Week 9 spreads
Our model’s playoff probabilities
Bark.
**
This post was also published at www.thebarkingcrow.com, where you can always find all of Joe Stunardi, Stuart McGrath, and NIT Stu’s work.
 | A guest post by
|