<
>

College football SP+ rankings after Week 13: Clemson moves up, the committee's BYU mistake and more

The biggest superpowers of the late-2010s are distancing themselves in 2020. Alabama and Clemson, winners of the 2015-18 national titles (among others, in the Tide's case), both looked spectacular on Saturday and inched further ahead of the pack in the SP+ ratings. They beat Auburn and Pitt (22nd and 34th in SP+ last week, respectively) by a combined 94-30, and they each gained about two adjusted points per game in the SP+ ratings. With Ohio State sidelined, the Tide and Tigers are now a semi-distant No. 1 and No. 2.

What is SP+? In a single sentence, it's a tempo- and opponent-adjusted measure of college football efficiency. I created the system at Football Outsiders in 2008, and as my experience with both college football and its stats has grown, I have made quite a few tweaks to the system.

More than ever, it's important to note that SP+ is intended to be predictive and forward-facing. It is not a résumé ranking that gives credit for big wins or particularly brave scheduling -- no good predictive system is. It is simply a measure of the most sustainable and predictable aspects of football. If you're lucky or unimpressive in a win, your rating will probably fall. If you're strong and unlucky in a loss, it will probably rise.

Those last two sentences are key. Week 13 might not have been the normal Rivalry Week in any way, shape or form, but it produced Rivalry Week-esque twist results.

According to my postgame win expectancy measure -- in which I take the key, predictive stats from a given game (the stats that eventually feed SP+), toss them into the air and say, "With these stats, you could have expected to win this game X% of the time -- 11 of the week's 44 games produced unlikely winners, especially in the Big Ten and Pac-12. Rutgers' win over Purdue came with a 12% postgame win expectancy, Iowa-Nebraska 16%, Stanford-Cal 34%, Oregon State-Oregon 37% and Washington-Utah 39%. Accordingly, a team like Nebraska rose in SP+, while Iowa fell slightly. Never let it be said that stats aren't incredibly anti-social.

Here are the full numbers.

This week's movers

Moving up

Eight teams moved up at least 10 spots from last week.

  • Liberty: up 17 spots from 49th to 32nd

  • UTSA: up 15 spots from 100th to 85th

  • Nebraska: up 14 spots from 56th to 42nd

  • Buffalo: up 14 spots from 63rd to 49th

  • Coastal Carolina: up 11 spots from 36th to 25th

  • Utah: up 11 spots from 37th to 26th

  • Colorado: up 11 spots from 69th to 58th

  • East Carolina: up 10 spots from 105th to 95th

Because of the total uniqueness of this season, I've been monitoring a lot of inputs I don't normally have to monitor in-season -- I've changed the home-field adjustment a couple of times as more data rolled in, for instance. But I've also been watching whether I should be weighting preseason projections differently this year than in previous years.

Based on the last few weeks of results, I decided to indeed lower the weight of said projections for teams who have played a healthy number of games. As you see above, that helped a few teams that have been dragged down by those preseason numbers (Liberty and Coastal, for instance). Other teams, like Nebraska or East Carolina, moved up on the weight of a strong Week 13 performance.

Moving down

Six teams moved down at least 10 spots as well, including a few that were being propped up by the preseason numbers (and a couple that got walloped by the aforementioned top two teams).

  • Auburn: down 16 spots from 22nd to 38th

  • Troy: down 15 spots from 64th to 79th

  • Kentucky: down 14 spots from 68th to 82nd

  • Pitt: down 10 spots from 34th to 44th

  • Ole Miss: down 10 spots from 43rd to 53rd

  • South Carolina: down 10 spots from 83rd to 93rd

SP+ vs. Résumé SP+

SP+ is indeed intended to be predictive and forward-facing. To aid in College Football Playoff discussion, however, I created what I call Résumé SP+ a few years ago. The goal was to assess your performance to date in a more résumé-related fashion.

Here's how I calculate Résumé SP+:

1. Project how the average top-five team (calculated by literally averaging the ratings of the current top five teams in SP+) would fare against your schedule in terms of an average projected scoring margin.

2. Compare said average projected margin to your own margin. (I install a 50-point cap on everything, so any win or projected win over 50 points just comes up as 50.)

The goal is to avoid simply judging a team by its schedule, instead looking at how it actually performed against said schedule. If you have a weaker slate -- like, say, BYU's, which ranks 108th to date according to SP+ -- you can still prove something by dominating that slate like a top-five team would. When it comes to the College Football Playoff committee's current discussions and methods, how you play doesn't matter nearly enough. It should matter far more than who you play.

The current Résumé SP+ top 15:

1. Alabama (+7.3 PPG over what the average top-five team would produce)
2. BYU (+0.3)
3. Clemson (+0.1)
4. Cincinnati (-0.3)
5. Wisconsin (-0.9)
6. Notre Dame (-2.7)
7. Ohio State (-3.4)
8. Oklahoma (-5.8)
9. Iowa (-6.3)
10. Marshall (-7.2)
11. Florida (-7.6)
12. Northwestern (-9.6)
13. Georgia (-9.6)
14. Indiana (-10.0)
15. Miami (-10.4)

Other key teams: Texas A&M is 19th (-13.4), Iowa State 20th (-13.5), USC 25th (-15.0).

(Reminder: Since we're setting the bar at "top-five quality," you would expect very few teams to finish with a positive average.)

This pretty clearly lays out the most confounding part of last week's initial CFP rankings release.

Anything you can say about the strength of BYU's schedule is probably correct. The Cougars have played one current SP+ top-50 team and have No. 45 SDSU on deck on Dec. 12. It's an awful slate. But (a) it's not their fault they've played an awful slate (their schedule disintegrated in August, and they had to scrounge for opponents), and (b) SP+ suggests that if an overall top-five team has played against this schedule, it wouldn't have fared any better than BYU has.

The Cougars are now seventh overall in SP+ and second in Résumé SP+. They have proven everything they possibly could against their schedule, and it's appalling that the committee, which continues to meet in person and discuss these rankings for hours, couldn't come up with an opinion more rigorous than they "ain't played nobody." Because six P5 teams canceled on BYU in August, the committee ranked the Cougars an astounding 14th -- I thought the absolute floor was 13th -- and decided they probably shouldn't even earn a spot in an NY6 bowl.

It's 2020. There are lots of good algorithms out there and lots of ways to judge teams in more rigorous ways than "ain't played nobody." But in its first ranking, the committee seemingly chose not to use them.