Writers of Pro Football Prospectus 2008

04 Oct 2017

FEI Week 5 Ratings

by Brian Fremeau

Regular readers familiar with the development of the FEI ratings over the years may have made note of the change in the way strength of schedule is measured this season. For years, I have positioned my strength of schedule data from the perspective of an elite team (two standard deviations better than average) rather than as a simple average of the power ratings of a given team's opponents as many other systems do. That's because perspective matters with schedule strength ratings. Top teams will have a much easier time winning a pair of games against two mediocre opponents than against one fellow elite opponent and one weaker one. A schedule strength rating that simply calculates an average opponent strength might not distinguish between the two.

Until this season, I have primarily represented schedule strength as the likelihood than an elite team would go undefeated against a given schedule. This was useful shorthand data, especially when used in the context of the BCS or College Football Playoff selection processes. But it also heavily weighted a schedule as reasonably challenging simply by playing a single elite opponent. A Sun Belt team, for instance, that played Alabama and a bunch of scrubs would sometimes measure as having a good schedule -- one that was difficult to go undefeated against, even for an elite opponent, even though it would not be particularly challenging to win 10 games against.

What I began to work with last year and have committed to this year is a different version of schedule strength data. I still prioritize the perspective of an elite opponent, but instead of measuring the likelihood that an elite team would go undefeated against a given schedule, I'm measuring the average number of losses an elite team would expect to have against a given schedule. This results in a more satisfying balance between recognizing the difficulty of playing one or more top opponents and the depth of a given schedule.

It's important to recognize also how schedule strength shifts throughout the year. The current FEI ratings have a .822 correlation with the preseason FEI projections posted way back in August, and preseason projection data still accounts for 29 percent of each team's rating. The projections and weekly adjustments have been reliable to date. My win likelihood data has anticipated a projection accuracy of 77.9 percent of games thus far, and FEI has correctly projected 77.6 percent of game winners through five weeks.

But there have been significant shake-ups in the ratings thus far as well. Teams have moved by an average of 17 ranking positions from the preseason until now, and some have seen dramatic shifts -- UCF is up 48 spots from No. 67 preseason to No. 19 this week, LSU down 50 spots from No. 9 preseason to No. 59 this week. And as a result of these team ranking shifts, schedule strength ratings have shifted dramatically as well.

Prior to the start of the year, SEC teams laid claim to seven of the eight toughest projected schedules. This week, Auburn (No. 13 SOS) is the only remaining SEC team with an overall schedule strength ranked in the top 25. The Big 12 and ACC lead the way now in the overall schedule strength ratings because, quite simply, they'll play multiple highly ranked opponents.

The Texas Longhorns have the nation's top-ranked regular season schedule this week, with games against three top-10 opponents on their slate, including USC, Oklahoma, and TCU, none of which are true home games for Texas. Florida State has the nation's second-toughest schedule with non-home games against Alabama and Clemson, and six games total against top-30 opponents.

On the flip side of that coin is the SEC. My ratings are very high on both Alabama and Georgia, but Auburn and Florida are the only other SEC teams ranked in the FEI top-30 this week. At the beginning of the year, Auburn's schedule ranked as the toughest with four projected top-10 opponents and nine games against top-50 opponents. This week, they have only a total of three top-10 opponents on their regular season schedule and only five games against the top-50 on their slate.

The top SEC teams are still very formidable, but much of the rest of the league has dropped precipitously. Missouri has dropped 64 spots in the FEI rankings from the start of the year, Ole Miss has dropped 63 spots, LSU has dropped 50 spots, Vanderbilt has dropped 40 spots, Tennessee has dropped 33 spots, Arkansas has dropped 29 spots, and Texas A&M has dropped 26 spots. Seven of the 13 biggest ranking drops in FEI rank from the preseason to date have come from SEC teams.

Again, it's still early in the year, and several of these SEC teams may recover with a strong finish. And even if they don't, both Alabama and Georgia could be playoff contenders with an opportunity to carry the torch for the rest of the league. They just may not have the opportunity in the postseason selection debate to lean on the strength of that league in the same way they have in the past.

FEI 2017 Week 5 Ratings

The Fremeau Efficiency Index (FEI) is a college football rating system based on opponent-adjusted drive efficiency in FBS games. Preseason projected ratings are a function of five-year program ratings, recent recruiting success, and returning offensive and defensive experience, and account for 29 percent of this week's ratings.

Strength of Schedule (SOS) ratings are a function of the projected FEI ratings of a given team's schedule of opponents and the location (home/away/neutral) of each game, representing the average number of losses an elite team (two standard deviations better than average) would have against the schedule. Each team's strength of schedule of opponents played to date (PSOS) is also provided.

Ratings for all 130 teams can be found here.

Rk Team Rec FEI SOS Rk PSOS Rk
1 Clemson 5-0 0.271 1.18 44 0.56 23
2 Alabama 5-0 0.268 0.85 75 0.37 58
3 TCU 3-0 0.261 1.41 32 0.40 53
4 Georgia 4-0 0.252 1.09 53 0.47 43
5 Oklahoma 4-0 0.243 1.63 16 0.54 26
6 Ohio State 4-1 0.220 1.35 35 0.43 49
7 Penn State 5-0 0.216 1.10 52 0.18 92
8 Michigan 4-0 0.207 1.42 29 0.19 89
9 Florida State 1-2 0.204 1.94 2 0.72 9
10 USC 4-1 0.200 1.16 47 0.52 29
11 Notre Dame 4-1 0.191 1.59 18 0.50 31
12 Oklahoma State 4-1 0.183 1.26 40 0.62 19
13 Washington 4-0 0.180 0.88 71 0.07 120
14 Stanford 3-2 0.178 1.64 14 0.79 7
15 Miami 2-0 0.177 0.93 65 0.11 115
Rk Team Rec FEI SOS Rk PSOS Rk
16 Wisconsin 4-0 0.167 0.65 85 0.04 127
17 Oregon 3-1 0.153 1.16 48 0.15 102
18 San Diego State 4-0 0.152 0.44 105 0.34 62
19 Central Florida 3-0 0.151 0.45 102 0.13 110
20 Auburn 3-1 0.149 1.65 13 0.67 11
21 Washington State 4-0 0.133 1.25 41 0.29 68
22 Louisville 3-1 0.131 1.33 38 0.47 40
23 Houston 3-1 0.125 0.54 93 0.22 81
24 Texas Tech 2-1 0.125 1.66 11 0.45 45
25 Virginia Tech 3-1 0.108 1.08 54 0.47 38
26 Florida 3-1 0.103 1.33 37 0.47 37
27 UCLA 3-2 0.102 1.59 17 0.53 28
28 Arkansas State 0-2 0.094 0.33 117 0.16 97
29 North Carolina State 3-1 0.093 1.57 19 0.49 35
30 Memphis 2-1 0.091 0.72 79 0.38 57

Posted by: Brian Fremeau on 04 Oct 2017

3 comments, Last at 11 Oct 2017, 12:54pm by lsjiv1

Comments

1
by Nahoj :: Thu, 10/05/2017 - 10:20am

I don't pay a whole lot of attention to college football, but I usually glance over these rankings, and anyway... Arkansas State?!?

2
by schmoker :: Mon, 10/09/2017 - 8:36am

I don't pay any attention to college football, but I also gotta wonder ... Arkansas State?

3
by lsjiv1 :: Wed, 10/11/2017 - 12:54pm

Hi Brian: I am working on analysis from 2016 and was wondering if there is anywhere to see/get FEI stats by week for the 2016 season? I've looked around google, footballoutsiders.com, and http://www.bcftoys.com but can't locate a full set per week. This is the last piece to my puzzle so hoping you can point me in the right direction :)

Thanks and keep up the great work,

Chip