The First CFP Rankings Are On Their Way: Who’s Number Five?

There are three things we’re going to learn from this evening’s release of the season’s first College Football Playoff rankings. The first doesn’t really matter. The third might matter, but in a diluted way. The second matters the most.

The first is who’s number one. It would be unsurprising to see any of Ohio State, Georgia, and Tennessee momentarily hold the crown. Ultimately, though, aside from the element where the committee’s nod towards Ohio State or an SEC team might foreshadow the same nod in December, and thereby a different semifinal opponent (something of dubious impact anyway, because if Clemson’s 13-0 it’ll be better to be the 2-seed and play them than to be the 1-seed and not play them), this is a pretty small deal. The playoffs are mostly one of those things where you’re in or you’re out.

The second is the big one, and it’s the order in which Michigan, Alabama, TCU, Clemson, USC, Oregon, and UCLA line up. We can be fairly assured at this stage that Tennessee will be ahead of Michigan tonight—Michigan’s going to need to make its bones later, or more realistically, get some help—but we don’t know how exactly these seven will shake out, the seven remaining who are either undefeated themselves or lost their lone loss to a well-respected foe. We talked yesterday about playoff lanes. These seven currently share five of them: the second Big Ten lane, the third SEC lane, the Big 12 lane, the ACC lane, and the Pac-12 lane. How the committee views them now is important. Our model, in simulations so far, has been inserting a loud random variable: FPA (Forgiveness/Punishment Adjustment, our measure of how much the committee is deviating from pure precedent in dealing with each team). Tonight, we learn where FPA currently actually stands, which will make FPA a lot quieter from here on out. This is a big shift. We get a lot of information from these first rankings.

The third is who all’s in the top 25, something that matters because the top 25 can be a bit sticky and because the committee could theoretically lean on the top 25 as a demarcation for quality victories. This is often circular, and therefore maddening, but it does kind of theoretically matter. If Mississippi State is the three-loss team in the field (and there should be at least one), that helps Alabama’s second-or-third-SEC-team lane. If a two-loss NC State or Wake Forest or Syracuse is held out, that’s a sign Clemson shouldn’t be given any room for error (though sometimes Clemson is treated differently by the committee from the rest of the ACC). If Tulane’s in there, maybe that’s good for the Big 12 in a world where things get really crazy, because Tulane is one of the teams who’s beaten Kansas State. Like we said: It’s diluted.

Here’s how our model sees the top 30, and how big each gap is if we project our formulaic estimation of the committee’s thoughts onto the whole of the FBS (100.0 = 1^st-ranked team; 0.0 = 131^st-ranked team, which is Akron at the moment, since I know you were curious):

1. Ohio State (100.0)
2. Tennessee (98.7)
3. Georgia (97.9)
4. Michigan (95.1)
5. Alabama (93.4)
6. TCU (91.8)
7. Clemson (91.8)
8. USC (85.2)
9. Oregon (83.2)
10. UCLA (82.0)
11. Penn State (81.9)
12. Illinois (80.2)
13. LSU (79.1)
14. Mississippi (78.7)
15. Utah (77.7)
16. Kansas State (77.7)
17. Oklahoma State (77.3)
18. North Carolina (77.1)
19. Syracuse (75.9)
20. Tulane (72.0)
21. NC State (72.0)
22. Texas (71.8)
23. Maryland (71.8)
24. Wake Forest (71.7)
25. Oregon State (70.1)
26. Liberty (68.6)
27. Notre Dame (67.9)
28. Mississippi State (67.1)
29. Purdue (66.6)
30. Washington (66.4)

It’s possible someone outside our expected top 30 will make the committee’s top 25, but it isn’t particularly likely. There are 24 Power Five teams with zero to two losses, and while our model doesn’t think Washington’s done enough for the committee to put it in, that does kind of tell the story. It’s very easy to make a top 25: Take the zero-to-two-loss Power Five teams, take Tulane or Liberty, maybe throw in a three-loss team like Texas or Mississippi State and kick someone out. It’s possible we’ll be wrong, but that would require a major deviation from precedent by the committee, and if I know the committee, it’s usually doing the easiest thing (which I’d argue is a good feature, but we can hash that out some other time).

Since we have an interesting argument for Number 1, we might as well talk through it, and I’d like to talk through the rest of the top ten and Mississippi as well, while you’re here. Starting from the back:

Mississippi is 8-1, plays in the SEC West, and somehow doesn’t have a single win our model evaluates as being among the top 99 in the FBS season so far. They beat Troy by eighteen points at home. That’s the big accomplishment. They have wins over Kentucky and Texas A&M, but they each came by only a field goal, and we don’t expect either of Kentucky or Texas A&M to be ranked (Kentucky is a funny sleeper, because picking the SEC team with the most hype as the three-loss team is something the committee would do, but it would be a reach). It is near magical that Mississippi drew this particular schedule in a sport that rewards scheduling your best games last (and thereby getting more facetime in the polls before you sink). Their loss—to LSU, by 25—isn’t terrible, but it kind of tells the story of what happens when you match these guys up with someone capable. Maybe the committee will do the easy thing and place them 11^th. This makes them a candidate for the biggest FPA boost of the week. Exciting!

USC, Oregon, and UCLA are a tight little pack, and kind of a fun one because they’ll shake themselves out, or Utah will shake them out on behalf of the rest of us, so we can argue and argue and all our arguments will never matter one single bit. USC’s loss is the best of the three, coming by a point on the road to Utah. Oregon has the best win, the beating of UCLA in Eugene. UCLA has the best second-best win, the win over Washington in Los Angeles that kind of put them on the map. USC has the best APD (Adjusted Point Differential), which is a measure of the margin by which teams beat their opponents relative to each opponent’s average margin, and is a messed up but useful proxy for the eye test and more legitimate evaluations of team’s actual capabilities. Right now, APD and the best loss are enough in the model’s eyes to have USC at the front of this pack, but it’s tight. Penn State could conceivably break things up, too, its two losses both among the 21 best in the country, even with Michigan laying the whooping on them in Ann Arbor.

I’m guessing Michigan will stay ahead of all three of Alabama, TCU, and Clemson, unless the committee gets onto the hamster wheel and says Alabama’s good because they lost to Tennessee who’s good because they beat Alabama. Between Clemson, TCU, and the Tide, Alabama’s the only one who’s lost, but they have a much, much better APD (second-best in the country, better than those of both Tennessee and Georgia), which the model thinks will be enough, with all three programs’ best wins similar in quality.

Finally, then, Ohio State and Tennessee and Georgia. Tennessee has the best top-two-wins combination, having both beaten Alabama and smoked LSU. Ohio State has the best APD, routinely routing fools. Georgia is the reigning champion and has had a rough game here or there but also beat Oregon—likely a top ten team—by 46 points. If the committee wants to steer into the moment, they could put Tennessee and Georgia 1 and 2, but that gets at the heart of what’s happening here, which is that this is a stylistic choice.

**

If you’re wanting a refresher on how the model works, here is said refresher. APD, FPA, how we calculate win quality and loss quality…all the good stuff. New probabilities tomorrow, barring the unexpected, like the committee ranking ME fourth in the country, which would be an honor but very hard to put into the model.