Oklahoma Looks Good, Georgia Looks Great: Our Model’s Read on Week 3

Week 3 is in the books, and the College Football Playoff picture’s clarity remains…let’s call it translucent. We received further confirmation that Georgia is very good. Alabama, Ohio State, and Michigan all dominated, as expected. We lost BYU and Michigan State as (varying levels of) fringe candidates, but we may have gained Washington and Oregon.

We’ll have a more thorough recap of the week posted tomorrow, and you can find the full list of our model’s probabilities, ratings, and rankings for FBS teams on this page, but for now, we’re going to look at the biggest movers and the big picture.

Entering the week, our playoff probabilities looked like this towards the top:

They now look like this (right-hand column is this week):

Some takeaways:

Oklahoma and Oklahoma State effectively flipped places, with our model going from being skeptical of the Sooners relative to its own preseason expectations to being impressed by their beatdown of Nebraska. Margin of victory is an important input to Movelor, our rating system, and while it does account for diminishing returns, being able to blow a team out is significant. It says a lot about both your offense and your defense. Blowing out a team like Nebraska—not a good team, but better than most of the other blowout recipients this weekend—shows something.

The threshold we arbitrarily chose for inclusion on these tables is 4.0%, and we’re down from 18 teams sitting above that threshold entering the week to just 15 this morning. Not all of this was due to losing, or even due to playing badly. Michigan State and BYU did fall off because of losses. But while Wake Forest and Arkansas—our others who departed—each struggled (Wake Forest escaped Liberty at home on a stopped two-point conversion, Arkansas couldn’t pull away from Missouri State until the fourth quarter in Fayetteville), those struggles were exacerbated by other performances within their conferences. NC State was impressive against Texas Tech, and got a win that was not guaranteed, driving up their own playoff probability but also their own Movelor rating, and thereby increasing the difficulty of Clemson and Wake Forest’s paths. Pitt doesn’t make the list—they’re at 3.7%–but they also impressed on the week, turning on the jets late against Western Michigan. Kentucky slipped in part because Georgia was so dominant against South Carolina, and because Tennessee had no trouble against Akron, mediocre though Akron may be.

We mentioned Oregon and Washington above, as they beat BYU and Michigan State, respectively (and did so firmly). Neither crosses 4.0% (Oregon’s at 3.3%, Washington’s at 1.2%), but each rose. We’ll talk more about Washington tomorrow, I’d guess, but they have a remarkably clear path, akin to Clemson’s all those years. Oregon, meanwhile, gave credence to the theory that their disaster in Atlanta had little to do with them and a lot to do with Georgia, and while they have a loss—none of the fifteen teams above 4.0% has a loss besides Utah, our model’s Pac-12 favorite—it’s about as good a loss as you’re going to see.

On the topic of Clemson: The model is low on the Tigers. It’s low on the ACC as a whole, but it’s especially low on the Tigers relative to public perception. Why? Well, the ACC was pretty bad last year, partially because Clemson wasn’t great. And so far, nothing’s been shown to change that. Pitt’s been fine, NC State has been bad and then good, Florida State’s been encouraging, but it’s hard to take any of those three seriously on the national scene, and the same goes for UNC, Wake Forest, and Miami. Miami received USC hype with Mario Cristobal entering the program, and so far they’ve done little to impress, looking impotent last night in College Station against a team who lost to truly mediocre Appalachian State (I love the mystique too, but that team couldn’t beat UNC and they shouldn’t have gotten away with it against Troy yesterday). More than anything, though, Clemson’s problem is itself. There are still big questions about the offense. After last night, there are some questions about the pass defense. Georgia and Alabama and Michigan and Ohio State are out there winning by seven or eight scores. Clemson is winning against oftentimes worse competition by four scores. If Clemson goes 12-1 and there’s an 11-1 Michigan or Ohio State, the conference championship will weigh heavily in the Tigers’ favor, but it’s fairly evident, to Movelor and to us and to betting markets, that the hypothetical Big Ten runner up will be the better team. Also, Clemson does have to play at Notre Dame, to whom Movelor is adjusting in the wake of their early struggles but whom Movelor still believes, rightly or wrongly, is the eleventh-best team in the nation, with Clemson fifteenth-best (but only a tenth of a point behind the Irish).

Penn State showed a lot to like and grabbed a respectable win, much like what NC State did but on a grander scale. Minnesota, meanwhile, has quietly been stomping bad teams after finishing last year somewhat hot. This program won eleven games in 2019 and nine last season, playing in the second-best conference in the country. Movelor likes that a lot, and our model likes their path given they play in the Big Ten West.

Finally, USC: Our model is coming around on the Trojans, but they had a steep hill to climb because our model doesn’t account for transfer talent (we have no real precedent for USC’s roster), and have they really been that impressive? They’ve beaten Rice, Stanford, and Fresno State. Movelor sees Fresno State as the 72^nd-best team in the FBS. We’ll learn a lot more when they face Oregon State on Saturday, whom Movelor ranks 29^th. Only more learning from there.

**

Again, more thorough recap of the week tomorrow. Maybe we’ll talk about sleepers.