| Rank | Team | FPI Rank | CHAMP Prob | POFF Prob | Change |
| 1 | Ohio State | 4 | 17.338% | 51.830% | 4.1% |
| 2 | Clemson | 8 | 14.366% | 51.504% | 17.2% |
| 3 | Baylor | 1 | 20.154% | 45.901% | -12.5% |
| 4 | Iowa | 29 | 3.351% | 43.962% | 0.1% |
| 5 | Texas Christian | 3 | 11.732% | 34.038% | -3.8% |
| 6 | Michigan State | 22 | 3.088% | 27.632% | 2.6% |
| 7 | Louisiana State | 7 | 7.155% | 25.080% | 5.7% |
| 8 | Notre Dame | 9 | 5.065% | 19.007% | -0.1% |
| 9 | Alabama | 6 | 3.740% | 12.410% | 0.8% |
| 10 | Oklahoma State | 19 | 1.671% | 12.408% | 3.3% |
| 11 | Stanford | 11 | 2.706% | 12.027% | 1.3% |
| 12 | Florida | 13 | 2.136% | 11.715% | -0.6% |
| 13 | Oklahoma | 2 | 4.187% | 11.463% | 4.2% |
| 14 | Utah | 21 | 1.212% | 10.022% | -10.9% |
| 15 | North Carolina | 24 | 0.534% | 5.527% | 0.7% |
| 16 | Florida State | 17 | 0.622% | 4.417% | -13.7% |
| 17 | Duke | 36 | 0.183% | 4.300% | 3.0% |
| 18 | Houston | 38 | 0.097% | 2.326% | 0.9% |
| 19 | Toledo | 47 | 0.073% | 2.242% | -0.5% |
| 20 | Wisconsin | 28 | 0.172% | 2.218% | 0.9% |
| 21 | Brigham Young | 48 | 0.059% | 1.876% | 0.7% |
| 22 | Memphis | 34 | 0.079% | 1.730% | 0.4% |
| 23 | Pittsburgh | 41 | 0.039% | 1.106% | 0.5% |
| 24 | Michigan | 15 | 0.141% | 0.954% | -0.1% |
| 25 | Mississippi | 10 | 0.201% | 0.840% | 0.2% |
| 26 | Texas A&M | 16 | 0.115% | 0.779% | -3.7% |
| 27 | Temple | 49 | 0.021% | 0.725% | 0.3% |
| 28 | Georgia | 14 | 0.088% | 0.581% | 0.0% |
| 29 | UCLA | 18 | 0.059% | 0.431% | 0.2% |
| 30 | Mississippi State | 20 | 0.049% | 0.373% | 0.2% |
| 31 | Northwestern | 57 | 0.005% | 0.258% | 0.2% |
| 32 | California | 25 | 0.012% | 0.133% | -1.0% |
| 33 | North Carolina State | 32 | 0.005% | 0.092% | 0.0% |
| 34 | Penn State | 53 | 0.002% | 0.082% | 0.0% |
| 35 | Washington State | 61 | 0.000% | 0.013% | 0.0% |
1. You may notice a new column in this week's rankings: championship probability. This is a pretty simple calculation, where I multiply the playoff odds by the odds of winning two playoff games. I determine the odds of winning a playoff game by calculating the spread of each team versus the average playoff team, and then doing the same for the final (the team in the final is generally expected to be a little bit stronger, because better teams are more likely to win in the semis). This results in some interesting things such as the wide gap between Baylor and Iowa, where they both have roughly the same odds of making the playoff, but then a much difference chance of success once they actually get there. By my model the Big 12 has a better than 1 in 3 chance of ending its decade-long title drought.
1b. I added the championship probability calculations largely because I wanted to start tracking something cool: The odds of having a "new" champion. As there are many such teams near the top of the rankings, I thought this might be an interesting thing to track. Since 1936 (the year the AP poll began), there have been 30 champions crowned by one of the major polls.* This suggests a rich diversity in the sports' history of champions. While that may be true, the recent past hasn't been as welcoming to new champions, with 1996 Florida being the most recent first-time winner. There have been close calls (1999 VT, 2010 and 2015 Oregon, 2011 Oklahoma State), but this year might represent the new blood's best chance yet. Utah's loss and Baylor losing their QB hurt the odds a bit, but we're still looking at a 30.5% chance of a new champion being crowned in January, with Baylor, Iowa, and Stanford leading the way.
*If you add in the discontinued FWAA poll, you would add three more champions to the list: Ole Miss, Arkansas, and Iowa. Most people don't recognize that poll as "official," so I'll just use the standard 30 teams for this exercise.
2. The action of week 8 eliminated another six teams from playoff contention. Pour one out for Kentucky, Miami (FL), Illinois, Arizona, Texas Tech, and your pre-season SEC champ Auburn. USC remains the only FPI top-ten team (5th) eliminated from these rankings.
3. If you add up individual team probabilities, then it appears as though the Big Ten is the most likely conference to sneak two teams into the playoff. That said, I don't necessarily think that is the case. What I think the model is currently getting at is that it's so sure that the Big Ten champion will make it while being really unsure of other conferences, that it's overstating the odds*. It will probably come down to how favorably the committee treats a decent but underwhelming Michigan State team. In other news, the Pac-12 is down to its last legs, as it probably needs Stanford to win out to have a realistic shot. In a year with fewer unbeaten teams in other conferences the odds wouldn't be so dire, but as the likelihood of finishing the year with 2-4 unbeatens rises, the Pac-12 is more and more likely to get shut out.
*I may try to build something into next year's model that accounts for this in-conference clustering. Or I may not. There's nothing saying that the committee will inherently be adverse to selecting multiple teams from the same conference. We'll see.
| Conference | Exp Playoff Teams |
| Big 10 | 1.269 |
| Big 12 | 1.038 |
| ACC | 0.669 |
| SEC | 0.518 |
| Pac 12 | 0.226 |
| Independent | 0.209 |
| American | 0.048 |
| MAC | 0.022 |
4. Clemson is the big upward mover of the week, having viciously deposed of Al Golden. Baylor's first human outing sinks them about as much as first-time losers Utah and FSU. The bye-week effect knocks TCU down a bit and keeps Iowa, ND, and Florida roughly even with the previous week. A bunch more prominent teams have byes this week (Baylor, Ohio State, LSU, Bama, MSU), but since the other top teams don't exactly play difficult opponents, I'm guessing there won't be a lot of shuffling next week. Notre Dame continues to lead the pack of one-loss teams, and will probably continue to do so, as Temple and Pitt both present decent challenges while still being beatable. Wisconsin is the king of the two-loss teams with a silly November schedule that should get them to 10-2. Houston takes the lead of the four "group of five" teams, although that race could still easily go any direction.
Week 9 Preview
| Home | Away | Home Win Prob | Playoff Teams Lost |
| NC State | Clemson | 26.11% | 0.045 |
| Florida | Georgia | 53.95% | 0.030 |
| Temple | Notre Dame | 20.95% | 0.022 |
| TCU | West Virginia | 81.37% | 0.021 |
| Texas Tech | Oklahoma State | 39.42% | 0.016 |
| Iowa | Maryland | 89.40% | 0.016 |
| Pittsburgh | North Carolina | 41.44% | 0.015 |
| Washington State | Stanford | 17.26% | 0.010 |
| Duke | Miami (FL) | 68.04% | 0.007 |
| Auburn | Ole Miss | 29.72% | 0.002 |
As promised in the season preview, this is your week to catch up on the non-football things you've been putting off. Sure, if you turn on your TV at 3:30, you'll probably be entertained for a solid 8 or 9 hours. But the odds of the season being turned on its head this week are quite low. Clemson is the only top team with a serious threat, and I think FPI is underrating the Tigers, so it's not even as big of a threat as my model thinks it is. Oklahoma State could face a major look-ahead threat a week before facing TCU, but I don't think they're much a playoff threat anyway, so that doesn't move the needle much. I honestly might be most excited for the Sneaky Good Bowl (Pitt-UNC), which is as good of a summary of the weekend as anything.
No comments:
Post a Comment