| Rank | Team | FPI Rank | CHAMP Prob | POFF Prob | Change |
| 1 | Clemson | 6 | 20.528% | 67.345% | 10.9% |
| 2 | Iowa | 29 | 3.797% | 49.791% | 7.3% |
| 3 | Ohio State | 5 | 15.077% | 47.413% | -1.3% |
| 4 | Baylor | 2 | 15.427% | 37.534% | -3.0% |
| 5 | Oklahoma State | 13 | 7.387% | 37.334% | 22.1% |
| 6 | Notre Dame | 9 | 7.391% | 26.781% | 4.9% |
| 7 | Alabama | 4 | 5.482% | 17.060% | 5.7% |
| 8 | Florida | 14 | 2.861% | 16.120% | -3.1% |
| 9 | Utah | 22 | 1.766% | 14.829% | 6.0% |
| 10 | Texas Christian | 3 | 5.155% | 14.515% | -24.4% |
| 11 | Stanford | 11 | 3.094% | 13.568% | 2.5% |
| 12 | Oklahoma | 1 | 5.551% | 12.948% | 2.4% |
| 13 | Louisiana State | 8 | 3.386% | 12.269% | -10.1% |
| 14 | North Carolina | 19 | 1.414% | 11.510% | 3.7% |
| 15 | Michigan State | 21 | 0.995% | 8.354% | -17.6% |
| 16 | Houston | 35 | 0.211% | 4.029% | 1.0% |
| 17 | Wisconsin | 24 | 0.275% | 2.795% | 0.4% |
| 18 | Brigham Young | 46 | 0.059% | 1.979% | 0.5% |
| 19 | Florida State | 18 | 0.189% | 1.323% | -2.9% |
| 20 | Michigan | 16 | 0.164% | 1.081% | 0.2% |
| 21 | Mississippi State | 15 | 0.085% | 0.531% | 0.2% |
| 22 | UCLA | 17 | 0.068% | 0.478% | 0.2% |
| 23 | Northwestern | 61 | 0.008% | 0.413% | 0.2% |
1. Lots of good teams lost in Week 10, especially those on the periphery of my playoff rankings. Memphis, Toledo, Duke, Pittsburgh, Texas A&M, Ole Miss, and Penn State all shuffled off this mortal coil with varying levels of surprise. Ole Miss is a particularly notable elimination, as they were as high as second on this list in late September.
2. The playoff picture as it relates to conferences couldn't be any clearer. The Big 12, Big Ten, and Clemson all have inside tracks at a spot, while the SEC winner and ND/Stanford winner will likely duke it our for the final spot. There's still plenty of room for goofiness: Clemson could lose, the Big 12 could all knock each other out, and Stanford plays four very losable games in a row. That said, everything is as clear as it reasonably could be, given how up in the air the season looked before yesterday.
| Conference | Exp Playoff Teams |
| Big 10 | 1.098 |
| Big 12 | 1.023 |
| ACC | 0.802 |
| SEC | 0.460 |
| Pac 12 | 0.289 |
| Independent | 0.288 |
| American | 0.040 |
3. FiveThirtyEight unveiled their updated playoff model this week. It's an interesting take on the proceedings, and is basically the polar opposite of my methodology. Whereas mine is as simple as I could make it, Nate and friends use a lot of granular details to make the model as exhaustive as possible. Even though there are definitely deficiencies in mine (see the next section), I think it is roughly just as good. In fact, most of the results of their model line up pretty well with mine, with two notable exceptions: Iowa and Alabama. I had Iowa just over 40% last week, while FiveThirtyEight had them at 7%. Conversely, Alabama was sitting at 11% here and 41% there. My model doesn't account for SOS or big wins, so it makes sense that it's overrating Iowa. It's also entirely dependent on FPI which is in a love affair with the Big 12, so that explains why Alabama is so low. While I agree that the extreme strength of schedule difference between those two teams should bring them closer together in actuality, I think the difference between the teams is extreme on the other end with the FiveThirtyEight model. As good as the Tide is, a one-loss team that had 3 or 4 difficult games left shouldn't have been that high in their playoff odds.
4. I'll write more on this at the end of the season, but I will be making a few tweaks to the model for next season. One obvious change would be incorporating other metrics into the team strength component. I like FPI, but it's had a few weird outliers this year, which is throwing everything off from the relative consensus of other models. I will likely add S&P+ and Sagarin's predictor next year to balance out some of the rough edges.
The other big change concerns a manual edit I made to this week's rankings. I noticed that Florida was ahead of Alabama in my initial standings, which seemed odd. The explanation was pretty simple: I needed to account for the SEC title game now that one of the participants was decided. I had been using a shortcut for most of the season that worked fairly well, assigning 0.4 extra losses to any team in a conference with a title game. Since Florida became the first team to clinch a division, that shortcut stopped working, because it implicitly assumed that they have a 60% chance of winning the title game (which they do not). Ideally, I would like to start predicting the participants in the conference title games, so I can assign the appropriate number of predicted losses to a team. And indeed, when the title game participants are sorted out, I will add those games to the simulation. Still, I can do better, and will figure out some method for doing this for 2016. In the meantime, I dropped Florida to 50% for the SEC title game, and will change that as needed.
5. I haven't commented on my disaster scenarios in a couple of weeks, so let's do that. The 2014 Florida State Watch is basically dead, with prime contenders Michigan State and TCU finally succumbing to the close game luck dragons. Ohio State was looking like this sort of team, but has enough dominant wins now to probably be "safe" from this insidious designation. On the other side of the coin, the 2007 Kansas Watch is alive and well, with Iowa just a few steps away from what might be an undefeated showdown in Indianapolis. So that's fun.
6. While some of the results of the weekend feel like the establishment getting stronger, the odds of a new champion being crowned actually went up, to 33.46%. One might assume Oklahoma State over TCU is the "have" beating the "have not", but the Cowboys are the team from that duo to have not won a title, so good for them I guess. A full 10 of the 23 remaining teams have not won a title.
Nebraska Bowl Watch
At the request of some fine*, young** doctors and pharmacists, I am adding a small section to track the progress of the much-beleaguered Cornhusker team. Their upset win over the Spartans allowed them to maintain faint bowl hopes, predicated on winning out. With the slight boost from winning such a big game, the Huskers are now at a 33.34% chance of going to a bowl, per FPI. Fainter still are their odds of making the playoff, but you never know if the committee will happen to listen to certain influential voices:
.@CFBPlayoff Nebraska is undefeated at night keep that in mind when you do your New Years Eve playoff picks
— Fake Bo Pelini (@FauxPelini) November 8, 2015
*Not really
** We're actually kinda old now
Week 11 Preview
| Home | Away | Home Win Prob | Playoff Teams Lost |
| Baylor | Oklahoma | 56.58% | 0.091 |
| Mississippi State | Alabama | 41.44% | 0.038 |
| Iowa | Minnesota | 82.43% | 0.029 |
| Arizona | Utah | 29.97% | 0.022 |
| Illinois | Ohio State | 13.12% | 0.021 |
| South Carolina | Florida | 23.87% | 0.019 |
| Iowa State | Oklahoma State | 13.96% | 0.017 |
| Syracuse | Clemson | 7.17% | 0.016 |
| Stanford | Oregon | 78.08% | 0.015 |
| LSU | Arkansas | 78.69% | 0.013 |
Week 11 is a lull of sorts between the carnage you just witnessed and the madness of the next two weeks. Still, it's a strong week that just happens to have the most impactful game of the year so far (which will probably be topped each of the next two weeks). In other news, FPI is a little bearish on Alabama, so it gives CLANGA State a pretty good chance to win a game in which the Tide find themselves on the inferior end of the quarterback spectrum (those are totally the games they lose). Everything else is basically an underdog with a ~20% chance of pulling something off. A week with a bunch of those type games is often awesome, so I'll be watching optimistically.
Nebraska hype is REAL
ReplyDelete