They put out the final rankings yesterday, and we didn’t see many surprises. Alabama passed Michigan for number one, which was the slightest surprise to our model but not actually all that surprising overall. Michigan State stayed ahead of BYU, which our model expected. Utah and Pitt jumped into the top twelve, which was surprising but meaningless. Oklahoma State, Oregon, Iowa, and Wake Forest hung up higher than expected, which was surprising but meaningless. San Diego State stayed in the rankings, which was surprising but not outlandish. Overall, it was a whole lot of nothing. But now we have the final top 25, so for the sake of closure, here’s what our model thinks of it (and here’s how the model works):
| Ranking | Team | Ranking w/o FPA | Est. Ranking Score | FPA | LW FPA | New FPA |
| 1 | Alabama | 2 | 100.0 | 2.7 | 2.5 | 0.1 |
| 2 | Michigan | 1 | 99.8 | 1.5 | 1.6 | -0.1 |
| 3 | Georgia | 3 | 96.8 | 0.0 | 0.0 | 0.0 |
| 4 | Cincinnati | 4 | 95.3 | 1.5 | 1.5 | 0.0 |
| 5 | Notre Dame | 5 | 89.9 | -1.9 | -1.9 | 0.0 |
| 6 | Ohio State | 6 | 89.4 | -0.3 | -0.3 | 0.0 |
| 7 | Baylor | 7 | 87.2 | 0.6 | 0.6 | 0.0 |
| 8 | Mississippi | 14 | 85.9 | 4.6 | 3.9 | 0.7 |
| 9 | Oklahoma State | 8 | 85.8 | 0.7 | 1.5 | -0.7 |
| 10 | Michigan State | 10 | 83.9 | 0.0 | -0.4 | 0.4 |
| 11 | Utah | 13 | 83.1 | 1.8 | 0.4 | 1.3 |
| 12 | Pittsburgh | 9 | 83.0 | -1.9 | -0.7 | -1.2 |
| 13 | Brigham Young | 11 | 82.5 | 0.1 | 0.7 | -0.6 |
| 14 | Oregon | 24 | 78.8 | 4.9 | 3.7 | 1.2 |
| 15 | Iowa | 19 | 78.0 | 1.0 | -0.9 | 1.9 |
| 16 | Oklahoma | 12 | 78.0 | -3.7 | -1.8 | -1.9 |
| 17 | Wake Forest | 20 | 77.2 | 0.4 | -1.6 | 2.1 |
| 18 | NC State | 22 | 77.0 | 0.7 | 2.2 | -1.5 |
| 19 | Clemson | 15 | 76.9 | -3.6 | -2.2 | -1.3 |
| 20 | Houston | 23 | 75.6 | 1.5 | -1.8 | 3.3 |
| 21 | Arkansas | 28 | 75.5 | 4.6 | 6.4 | -1.8 |
| 22 | Kentucky | 25 | 75.3 | 1.6 | 1.5 | 0.1 |
| 23 | Louisiana-Lafayette | 16 | 75.1 | -4.6 | -3.7 | -0.9 |
| 24 | San Diego State | 26 | 73.1 | -0.1 | -2.5 | 2.4 |
| 25 | Texas A&M | 27 | 73.0 | 2.1 | 5.4 | -3.3 |
| NR | Wisconsin | 17 | 72.8 | -5.8 | -4.2 | -1.6 |
| NR | UTSA | 18 | 72.7 | -4.6 | -4.5 | -0.1 |
| NR | Purdue | 21 | 72.6 | -3.9 | -2.7 | -1.2 |
The big takeaway here is that it doesn’t hurt you very much to lose your conference championship. A conference championship loss, it seems, may be less damaging than an identical loss would have been in the regular season, especially if you aren’t in contention for a playoff or New Year’s Six berth. With the short turnaround on the rankings from Saturday night to Sunday morning, this is arguably fair. We probably don’t actually want the committee pausing playoff résumé dissection to decide who jumps into the top 25 with San Diego State dropping out.
Personally, I have few complaints overall. Were there no FPA, the entire New Year’s Six lineup would have stayed the same with the exception of Alabama and Michigan flipping, and one important caveat about FPA is that it’s possible the metric is reflecting trends our model is missing from the past in addition to how the committee’s deviating from precedent. Conference-based FPA adjustments may be in the works for our model next year (we need to iron out some wrinkles in FPA overall, too), but the fact the committee was consistent enough, relative to these last seven years, for our model to accurately predict the New Year’s Six participants, nearly perfectly predict the bowl lineup, and accurately designate at least 25 of the committee’s top 28 teams without taking any prior rankings into account speaks well of the system’s consistency, which one would think is a key piece of any fair system. Does this mean this system is fair? No. It’s flawed. But there’s little if anything to sincerely complain about this year, and that’s kind of nice.
Now, to watch it all play out. Plenty more to come on that front.
