2023 Initial CFP Committee Rankings (reaction): The AP Poll is good enough for the Committee

The first Committee Rankings are out, and...they basically just copied the AP poll, with a few minor adjustments. Over the years they've been getting less independent but this year is the worst. The top 8 are almost exactly the same. In fact, using my formula (which failed to predict the top ten very well at all) but doubling the weight of the AP poll, I managed to match the top 9 places. So compared to the average year, the Committee was twice as unoriginal this time around.

Here's the Committee Top 25, with the divergence from the AP poll in parens:

1. Ohio State (up 2): This was what we (and most everyone) predicted, but in practice it wasn't due to their two big wins necessarily. See below
2. Georgia (down 1): The Committee basically copied the top 8 of the AP poll but had to dock Georgia a spot for their terrible SOS.
3. Michigan (down 1): Ditto with the awful triple digit Wolverine SOS. Basically last year is as important as this year, is what the Committee is saying
4. Florida State (even): Looks like the Clemson win isn't much of a plus any more.
5. Washington (even): No real credit for the Oregon win, except being ahead of Oregon
6. Oregon (even): Yeah the Washington game was very close, but Ohio State is 10 spots ahead of Penn State
7. Texas (even): Their win over Alabama is much better than anything Oregon can claim, but whatever
8. Alabama (even): If they had lost to Texas on the road instead of at home, I swear the Committee would rank them higher than the Longhorns. They want to so bad.
9. Oklahoma (up 1): They give the Sooners a little more credit for beating Texas than the AP does
10. Mississippi (up 1): I figured they'd put Ole Miss here since they have the Tulane win and Penn State has no top 25 wins
11. Penn State (down 2): Their résumé is very similar to Oregon's: loss to undefeated team, win over 6-2 team, but here they are 5 spots lower
12. Missouri (up 2): Tigers are given credit for beating K-State, and only loss is to LSU
13. Louisville (up 2): Interesting, I'm glad the Committee put 1-loss Louisville over 2-loss Notre Dame whom they beat. I didn't think they would.
14. LSU (down 1): Tigers are the first 2-loss team...but should be ahead of Missouri
15. Notre Dame (down 3): The biggest downgrade from the AP so far. The Duke and USC wins are waning in value.
16. Oregon State (even): beat both Utah and UCLA, so they are top dog in the 2nd-tier of the Pac-12
17. Tennessee (up 2): Interesting that the Vols get an upgrade from their AP #19. They have beaten three 5-3 teams I guess.
18. Utah (even): Beat both UCLA and USC so they are properly ahead of them
19. UCLA (up 1): Résumé very similar to USC
20. USC (up 4): Biggest upgrade of the lot. Why? They've beaten one winning team. People who think the Committee looks at components like offense and defense and shuns teams with bad defense, are clearly mistaken; maybe they love strong offense more?
21. Kansas (up 1): Arguably could be behind Oklahoma State (who beat them) but they have the Oklahoma win, so...
22. Oklahoma State (up 4): Another big upgrade, and we had the Cowboys at #24 (by formula) or #22 (my guess). They beat Kansas and K-State but have 2 poor losses too
23. Kansas State (up 2): The Wildcats are a tough team to rank, they are hot and cold. This seems right, behind OK State who beat them.
24. Tulane (down 3): Tulane's SOS is around #100, and the Committee always hates non-Power Five teams if they can. But Tulane proved themselves last year by finishing in the top 25 so they get a pass.
25. Air Force (down 8): By far the biggest downgrade in the list, and I suspected it would happen. Our formula said #20, and I had #23 but pondered that the Falcons' #129 SOS might keep them out altogether. But as I noted, there aren't many candidates beyond these teams. I thought maybe Arizona would get in instead of Tulane, but it was neck and neck.

So there you have it. Clearly SOS only matters really if you're a non Power Five team, otherwise it's A-ok to play nobody and end up #2 or #3. Not that this matters in the long run, as Power Five teams all end up playing tough teams. Michigan and Georgia have about the toughest remaining schedules of any teams, while Air Force's slate doesn't get much harder.

Like in previous years, the Committee re-shuffling of the AP top 25 isn't very well thought out, though commentators treat it like the Bible. USC up four spots for playing the worst defense of any top 25 team? That's not insight, it's just random. They did put Louisville ahead of Notre Dame which I commend, but then why isn't Oklahoma still ahead of Texas? 

And there's the Penn State vs. Oregon discrepancy. Both have a road loss to an 8-0 Power conference team (Ohio State; Washington), and a dominant win over a 6-2 team (Iowa; Utah). Utah is ranked while Iowa isn't, but other than that, Oregon has 2 wins over 4-4 teams while Penn State has a 4-4 win and a 5-3 win. Then they each have an FCS win and three wins over losing teams. Is that five spots worth of difference? They've been equally dominant in their wins, pretty much. 

The bottom line? The AP puts Oregon at #6 and Penn State at #9, so the Ducks were going to be ahead in the Committee unless they went radical. Instead, they pushed them farther apart, which is the wrong direction. The AP has it better, three spots apart not five.

So the question is, is the Committee's top 25 "better" than the AP's? It should be given its importance, but I'm not sure it is. It's a mixed bag:

• Having Georgia #2 is better than having them at #1 right now, but not by much. An entirely new, from scratch re-evaluation of the season with no looking at last year's results would have the Bulldogs lower than that.
• Most of the top ten is just a rip-off of the AP anyway, so what's the point?
• They increase the AP's error in having Oregon too far ahead of Penn State; the AP's discrepancy is defensible, the Committee's isn't
• Having Missouri (7-1) above LSU (6-2), who beat them, is questionable. The AP has LSU first, and logically it works since LSU lost to #4 FSU and #10/11 Ole Miss, both higher than Missouri. Putting LSU ahead of Missouri makes all their results consistent. So I give this one to the AP, big time. The Committee is the one that is supposedly analyzing these things as a group, while the AP is just a collection of lists! These things are why I have no confidence that the Committee will get things right, especially early on when you really have to think.
• They do have Louisville (7-1) ahead of Notre Dame (7-2), which is good
• They have the minor Pac-12 teams in the right order (Oregon State, Utah, UCLA, USC), but so does the AP. The Committee does have UCLA and USC back to back, which is better than the AP which has four spots between them when their losses are very similar (I suspect it was just random though—they were talking Pac-12 and put all the teams in a cluster).
• They put Oklahoma State, who beat Kansas and Kansas State, in the top 25 while the AP has them just outside. But that's not the AP's fault as they rank 8-0 James Madison, who isn't eligible for the Committee's top 25. 
• The Committee puts Tulane (7-1) and Air Force (8-0) at the bottom, when the AP has them #21 and #17. Is this good? If you value SOS, yes. But they didn't really follow through with Georgia and Michigan. You can say those teams are proving their worth with their gameplay, but Air Force is a solid team by that standard too. It really makes it look like anti-Group of Five bias, which the AP poll has too, but it's not as bad as the Committee. It looks like they got to #23 and realized hey! We only have Power Five teams ranked so far! Better throw in a couple of Group of Fivers! And rank them by SOS to show how important SOS is! Which, by the way, the only reason Tulane's SOS is better is the inclusion of Ole Miss—and they lost that game. A 7-0 Tulane has a schedule that is barely better than Air Force's—so they are getting a lot of credit for playing a game and losing it.
• Overall the average deviation from the AP rank is 1.64 places, and for the entire top 11 you can argue there are just two moves: Ohio State from #3 to #1, and Penn State from #9 to #11. For the top 19, the average deviation is 1 place.

So I'd say the Committee did not improve on the AP's ranking all in all. Tasked with starting from scratch, they punted and made obvious and sometimes erroneous minor moves. Take any random 13 votes from the AP and you might get something similar. Which is disappointing considering the Committee actually meets and discusses these rankings in depth, whereas the AP voters just vote (presumably while eating breakfast on Sunday, watching NFL previews). 

But luckily these early issues tend to go away. The AP will copy most of the Committee's moves. Why, I don't know, but watch for Air Force to inexplicably fall in the AP, Ohio State to pass Michigan, and if Georgia looks week at all against Missouri the Buckeyes will be #1 on the 5th. And the two lists will march in lockstep as the season goes on, each influencing the other here and there until the final week when they're almost the same. 

This year, though, the Committee should have just copied and pasted the AP top 25, would have saved them a lot of time and ended up just as good, right from the start.