Writers of Pro Football Prospectus 2008

Most Recent FO Features

HurtJ.J.jpg

» The Truth Behind Rising Injury Rates

Zachary O. Binney tries to explain the rise in NFL injury rates over the past decade. Are players getting hurt more often? Or are teams doing a better job of diagnosing and reporting injuries?

15 Oct 2014

FEI Week 7: Game Factors

by Brian Fremeau

The top three teams in the FEI ratings last week all lost over the weekend. Arizona (No. 1 last week) missed a two-point conversion attempt late in the fourth quarter and a field-goal attempt as time expired to fall to USC for their first loss of the year. UCLA (No. 2 last week) lost to Oregon 42-30 in a game where the Bruins trailed by 32 points with 10 minutes left in the fourth quarter. And TCU (No. 3 last week) lost to Baylor 61-58, giving up 24 straight points to the Bears over the final 11 minutes of the game.

The weekly ratings were rocked once again as a result. The average team in the FEI top 25 changed seven spots each from last week to this week, due not only to upsets but also to the elimination of preseason projection data from the FEI formula. From this point forward, only 2014 game data is included in the FEI calculations. Instability may still reign, of course. Through the first seven weeks of the season, there have been 36 upsets against teams that were favored by at least a touchdown heading into the game, a 50 percent increase in such upsets over the same span last season.

Beginning this week, I've also published opponent-adjusted offensive efficiency (OFEI) and opponent-adjusted defensive efficiency (DFEI) data. The defensive leaderboard doesn't feature many major surprises. Louisville leads all teams in DFEI, and also leads most unadjusted measures including first down rate (allowing a first down on only 38.8 percent of opponent drives), available yards surrendered (21.2 percent), explosive drive rate (1.2 percent), and value drive rate (15.7 percent). The Cardinals rank second nationally in points allowed per drive (0.67) and No. 1 in points allowed per value drive (2.55).

The offensive leaderboard is a bit more startling. The UCLA Bruins rank No. 1 in OFEI through the first seven weeks of the season, despite ranking only 28th in yards per game and 34th in scoring per game, 48th in available yards, and 63rd in value drive percentage. What does the FEI formula like about those numbers? As with many FEI questions, it all comes down to the opponent adjustments.

UCLA has faced the No. 3, No. 9, No. 21, No. 31, No. 34, and No. 54 defenses so far, a group of opponents that represents the No. 2 offensive strength of schedule to date. Four of UCLA's raw offensive efficiency performances to date rank among the top 200 offensive games played in college football (out of 678 total performances). When adjusted for opponent, all six of their offensive performances to date rank among the top 200, and four rank among the top 30. Washington State is the only other team with four games ranked in the top 30 of OFEI. Only three other teams (Arizona, Miami, and North Carolina) have at least two such games.

Every offensive, defensive, and overall single game performance is ranked in the FEI Game Factors data I publish on my site. In addition to the raw and opponent-adjusted data and game ranks, the relevance of each data point in the FEI formula is also provided. For UCLA, more weight is given to their upset loss to Utah (19.9 percent of the Bruins' FEI rating) than to their win over Texas (11.5 percent). Game weights are unique to each team and are a function of the relative strength of opponents faced. Mississppi's win over Alabama (25.6 percent of the Rebels' rating) is given nearly five times as much weight as its win over Louisiana Lafayette (5.5 percent).

Unlike Game Splits, which represent measures of value generated by offense, defense, special teams, field position, and turnovers, the Game Factors data changes as the season progresses. As UCLA's opponent ratings change, the relative success the Bruins had in their games against those opponents will also change.

Last year, the national champion Florida State Seminoles finished the season with eight of the top 100 opponent-adjusted single game performances of the year. Five teams each have four top-100 GFEI performances to date so far, and all five teams rank among the top 15 in FEI: No. 2 Ole Miss, No. 3 USC, No. 9 Stanford, No. 11 Oklahoma, and No. 13 Auburn. Six others have each recorded three top 100 GFEI performances to date: No. 4 TCU, No. 7 Georgia, No. 8 Baylor, No. 20 Clemson, and No. 34 Washington State. The lowest ranked team with at least two top 100 GFEI performances to date is Texas. Both of those games were close losses to highly ranked opponents (UCLA and Oklahoma). Four other Longhorns games were average or below-average performances.

Game Factors are the opponent-adjusted guts of each team's FEI rating. If you aren't sure why FEI thinks your team is overrated or underrated, the answer can be found in the Game Factors data. Opponent adjustments are very meaningful in this system, and some individual games carry significant weights at this point in the year.

FEI 2014 Week 7 Ratings

The Fremeau Efficiency Index (FEI) is a college football rating system based on opponent-adjusted drive efficiency. Nearly 20,000 possessions are contested annually in FBS vs. FBS games. First-half clock-kills and end-of-game garbage drives and scores are filtered out. Game Efficiency (GE) is a function of the starting field position and outcome of non-garbage possessions. Opponent adjustments are calculated with special emphasis placed on quality performances against good teams, win or lose.

Strength of Schedule (SOS) is calculated as the likelihood that an "elite team" (two standard deviations above average) would win every game on the given team's schedule. SOS listed here includes all regular season games scheduled. A multifaceted approach to measuring schedule strength is available here and visualized here.

Mean Wins (FBS MW) represent the average total games a team with the given FEI rating should expect to win against its complete schedule of FBS opponents. Remaining Mean Wins (FBS RMW) represent the average total games a team with the given FEI rating should expect to win against the remaining opponents on its schedule.

Offensive FEI (OFEI) is opponent-adjusted offensive efficiency. Defensive FEI (DFEI) is opponent-adjusted defensive efficiency. Special Teams Efficiency (STE) is the composite efficiency of field goal kickoff, punt, and return units. Field Position Advantage (FPA) is the average share of field position value generated by offense, defense, and special teams efficiency.

These ratings are exclusively produced from 2014 game data and are not influenced by preseason projections. Complete ratings and ratings splits for all 128 FBS teams are available here. Supplemental data including points per drive, game splits, and game factors are available as well.

Rk Team FBS
Rec
FEI LW GE GE
Rk
SOS Rk
FBS
MW
FBS
RMW
OFEI Rk DFEI Rk STE Rk FPA Rk
1 Oregon 4-1 .279 5 .173 15 .068 13 8.6 4.7 .628 5 -.436 21 .478 53 .537 24
2 Mississippi 6-0 .271 6 .320 2 .190 45 9.3 3.9 .432 19 -.663 7 1.355 33 .567 9
3 USC 4-2 .266 12 .106 28 .035 4 8.9 4.5 .613 6 -.264 40 -.445 78 .493 68
4 TCU 3-1 .249 3 .165 17 .289 69 9.6 6.5 .477 16 -.423 22 3.081 7 .570 8
5 Arizona 5-1 .242 1 .090 31 .044 8 8.6 4.2 .526 13 -.276 37 -.370 77 .548 18
6 Mississippi State 6-0 .241 10 .246 7 .154 36 8.9 3.6 .360 23 -.626 8 .213 58 .535 27
7 Georgia 5-1 .234 22 .266 4 .205 47 8.8 3.8 .606 7 -.204 46 2.706 11 .602 1
8 Baylor 5-0 .230 19 .284 3 .280 64 9.2 4.9 .377 20 -.577 12 -.558 83 .539 23
9 Stanford 3-2 .224 18 .118 25 .027 2 6.8 3.4 -.002 64 -.773 2 1.579 25 .561 12
10 Utah 3-1 .220 8 .148 21 .045 9 7.0 3.9 -.247 95 -.731 3 5.041 1 .598 2
11 Oklahoma 5-1 .217 4 .224 8 .237 54 9.9 5.0 .367 21 -.372 25 4.294 3 .555 16
12 Florida State 5-0 .212 14 .184 13 .237 55 8.6 4.3 .524 14 -.344 30 .284 57 .494 66
13 Auburn 5-1 .211 7 .193 12 .057 12 7.4 2.7 .534 9 -.303 35 1.400 32 .542 22
14 UCLA 4-2 .209 2 .051 49 .039 6 7.3 3.4 .916 1 -.096 56 -.461 81 .511 49
15 Alabama 5-1 .190 13 .180 14 .113 26 7.5 3.0 .294 31 -.548 14 -2.254 114 .455 112
Rk Team FBS
Rec
FEI LW GE GE
Rk
SOS Rk
FBS
MW
FBS
RMW
OFEI Rk DFEI Rk STE Rk FPA Rk
16 Georgia Tech 4-1 .188 11 .085 34 .167 40 7.9 4.1 .763 3 .118 77 2.864 10 .537 25
17 Notre Dame 6-0 .184 15 .204 10 .108 22 8.5 3.4 .228 40 -.585 10 1.218 36 .513 46
18 Colorado State 4-1 .184 34 .073 40 .360 79 9.0 5.5 .497 15 .010 66 1.554 27 .492 69
19 Virginia Tech 3-2 .182 17 .080 37 .203 46 7.9 4.3 -.086 76 -.705 6 1.625 21 .517 42
20 Clemson 3-2 .169 36 .081 36 .112 25 7.5 4.4 .247 34 -.584 11 -.452 80 .511 48
21 Louisville 4-2 .164 25 .107 27 .156 37 7.4 3.1 -.086 75 -.869 1 .213 59 .458 107
22 Washington 4-1 .163 49 .160 18 .056 11 7.2 3.5 -.116 79 -.554 13 .445 56 .581 5
23 Boston College 3-2 .161 37 .075 38 .127 28 7.0 3.6 .247 35 -.400 24 -1.057 92 .505 55
24 Arizona State 3-1 .160 24 .025 58 .043 7 5.5 3.3 .532 10 -.342 31 -1.844 105 .449 115
25 Ohio State 4-1 .159 16 .248 6 .281 66 9.1 5.1 .307 29 -.270 38 2.582 13 .589 4

Posted by: Brian Fremeau on 15 Oct 2014

9 comments, Last at 17 Oct 2014, 11:59pm by Will

Comments

1
by burbman :: Wed, 10/15/2014 - 2:50pm

FEI clearly ranks the PAC12 conference too high because everyone knows that computers harbor a West-Coast bias, due to the influence of Silicon Valley, and the *fact* that only the SEC could possibly field that many top programs simultaneously. Randomly drawing paper slips from a hat is way better than this.

Removing tongue from cheek now, I am surprised that the PAC12 is showing 8 of the top 25 rankings this far into the season. How much of the SOS advantage is due to the incestuous nature of conference schedule?

3
by Brian Fremeau :: Wed, 10/15/2014 - 3:17pm

Incestuous conference schedule definitely helps. FBS non-conference schedule has helped too, not only with several key wins but also not too many poor opponents. Pac-12 has played four total games against American, Conference USA, MAC, and Sun Belt (winning all four), and most of the nation's worst teams play in those conferences.

SEC is 22-1 against those four conferences.
Big Ten is 19-3.
Big 12 is 11-0.
ACC is 15-4.

2
by zenbitz :: Wed, 10/15/2014 - 3:13pm

I find this really interesting because there is so little data to base the relative conference rankings. The matrix is just too sparse. It's probably being determined by something like UCLA playing Virginia, Memphis and Texas as it's non-conference opponents. Usually they toss in a UC Davis or something in there.

4
by Kal :: Wed, 10/15/2014 - 3:55pm

Yes, and all the other Pac-12 teams. Oregon playing and dominating MSU. Arizona State playing well against LSU. WSU playing Rutgers and Nevada. Arizona playing Nevada. Stanford playing Notre Dame.

And because of how FEI works, those successes get spread to their opponents. Oregon played MSU well, so Arizona beating Oregon must make Arizona better. Oregon dominated UCLA which played well (kinda) against Memphis, Virginia and Texas, so Oregon must be better - and for that matter, Virginia, Texas and Memphis are all good. That everyone in the PAC-12 appears to be similar in strength also helps, as FEI rewards harder schedules and good games more than S+P does.

5
by stanbrown :: Wed, 10/15/2014 - 8:08pm

Pac 10 wins over Idaho St, Weber St, Fresno, New Mexico, Portland St, Army, UC Davis, E Washington just scream quality.

6
by Brian Fremeau :: Wed, 10/15/2014 - 8:43pm

FCS opponents not included (though even if I did, Pac-12 hasn't played more of those games than the other conferences either).

7
by olivertheorem :: Wed, 10/15/2014 - 9:46pm

Something's up with the overall FEI rankings. The list maxes out at 125 schools, but some offense/defense/ST rankings go higher than that. I think 3 or so schools got cut off.

8
by Brian Fremeau :: Thu, 10/16/2014 - 8:29am

Thanks, fixed the problem.

9
by Will :: Fri, 10/17/2014 - 11:59pm

Are my eyes deceiving me, or is Ohio State the only ranked B1G team at #25?

Will