Writers of Pro Football Prospectus 2008

Most Recent FO Features

HarvinPer09.jpg

» Impact of the NFL's Kickoff Rule Change

After three NFL seasons of kicking off from the 35-yard line, what has been the impact on touchbacks, returns, field position, scoring and injuries? Also, is this rule responsible for a record number of big comebacks?

09 Jan 2013

Final 2012 FEI Ratings

by Brian Fremeau

Alabama played its best game of the season on Monday night in a 42-14 victory over Notre Dame, but that’s understating it a bit. The Crimson Tide played the best college football game of the 2012 season period; the second straight year they have accomplished such a feat in the BCS title game. Yet, that’s still selling the win short. In the 2012 BCS championship match-up against the Fighting Irish, the Crimson Tide recorded the most dominant single game opponent-adjusted performance of the last ten years.

Like many others, I expected the BCS championship to be a low-scoring defensive battle between two of the nation’s best defenses. I expected both teams to move the ball reasonably well from their own end of the field, but to struggle to turn those drives into points as the end zone drew near. I expected to dissect each team’s execution by first-down series field position as a critical, measurable storyline of the game. Alabama obliterated my expectations.

The Crimson Tide scored a touchdown on their first three possessions and six of their first eight drives. Notre Dame had only surrendered seven touchdown drives of 50 or more yards all season heading into the title game, and Alabama nearly doubled that total in one night. Before Monday, the Irish had faced 122 opponent first-down series in Notre Dame territory, and only 18 percent of those situations led to a touchdown. In the title game, Alabama had 14 first-down series start in Notre Dame territory, all of which eventually led to the end zone.

Maybe we can look through the rearview mirror and wonder why we didn’t see that kind of performance coming all along. Alabama has a recent history of dominating performances in big games under Nick Saban, and as Matt Hinton described it yesterday, the Crimson Tide are a Death Star, a machine, and a dynasty built to last.

That said, the only game in which Alabama's offensive efficiency was comparable this season was in a 49-0 beatdown of Auburn at the end of the regular season. That Auburn team ranked among the worst teams in college football, No. 116 in FEI, No. 119 in OFEI, and No. 100 in DFEI. In the BCS championship game, Alabama's offense made one of the best defenses in the country look like one of the worst.

Alabama dominated Georgia's defense in the SEC championship game to the tune of 350 yards rushing, 512 total yards, and 7.1 yards per play. But against Notre Dame, the Crimson Tide machine generated more than twice as much offensive value in half as many possessions. That's ridiculous efficiency against any opponent, but this came against one of the best defenses in the country. Notre Dame didn't get on the scoreboard until garbage time hit.

Nick Saban's program in Tuscaloosa is the best in the country. The best players in the best system have proven themselves against most of the best teams over the last four years. It is especially remarkable that they have played their best against the best competition so consistently.

  • 2009: Alabama crushed Florida (FEI No. 2) 32-13 in the SEC championship game, Florida's only loss of the year.
  • 2010: Alabama crushed Michigan State (FEI No. 39) 49-7 in the Capital One Bowl, Michigan State's second loss of the year.
  • 2011: Alabama crushed LSU (FEI No. 1) 21-0 in the BCS championship game, LSU's only loss of the year.
  • 2012: Alabama crushed Notre Dame (FEI No. 10) 42-14 in the BCS championship game, Notre Dame's only loss of the year.

These games were either the final game or the second-to-last game played by Alabama in each of the last four seasons. All of them were remarkably dominant performances. We have just completed the first decade of the FEI era (2003-present), and there have been 7,036 FBS games played in that span. That's 14,072 individual game performances in the last decade. The four Alabama games listed above all rank in the top 20 of opponent-adjusted game efficiency performances, in the 99.9 percentile of the last decade. Monday night's game was the single-most dominant game of the last ten years.

No, FEI didn't see it coming. Alabama was ranked No. 4 heading into the bowl season and the Irish were ranked No. 3. I didn't expect that the winner of the BCS championship would have been able to leap ahead of the winner of the Fiesta Bowl (then-No. 1 Kansas State against then-No. 2 Oregon). But the best game played among the last 14,000 did the trick. Alabama is No. 1.

This marks the seventh straight year in which the FEI No. 1 and No. 2 teams did not play for the BCS championship (though they did meet in the SEC championship game in 2006). The current system is going away soon, and the field will be expanded to include many of the best teams that were left out. The No. 2 Oregon Ducks would have been a good candidate to challenge the dynasty, but who knows if it would have made much of a difference. As the Crimson Tide proved once again, knocking off Alabama in an end-of-year championship environment requires a once-in-a-decade level performance.

FEI Final Top 25

The Fremeau Efficiency Index (FEI) rewards playing well against good teams, win or lose, and punishes losing to poor teams more harshly than it rewards defeating poor teams. FEI is drive-based and it is specifically engineered to measure the college game. FEI is the opponent-adjusted value of Game Efficiency (GE), a measurement of the success rate of a team scoring and preventing opponent scoring throughout the non-garbage-time possessions of a game. FEI represents a team's efficiency value over average.

Other definitions:

  • SOS: Strength of schedule, based on the likelihood of an elite team going undefeated against the given team's entire schedule.
  • FBS MW: Mean Wins, the average number of games a team with the given FEI rating would be expected to win against its entire schedule.
  • OFEI: Offensive FEI, the opponent-adjusted efficiency of the given team's offense.
  • DFEI: Defensive FEI, the opponent-adjusted efficiency of the given team's defense.
  • STE: Special Teams Efficiency, the scoring value earned by field goal, punt and kickoff units measured in points per average game.
  • FPA: Field Position Advantage, the share of the value of total starting field position earned by each team against its opponents.

These FEI ratings are a function of results of games played through January 7th. The ratings for all FBS teams, including FEI splits for Offense, Defense, and Special Teams can be found here. Program FEI (five-year weighted) ratings and other supplemental drive-based data can be found here.

Rk Team FBS
Rec
FEI LW GE GE
Rk
SOS Rk FBS
MW
OFEI Rk DFEI Rk STE Rk FPA Rk
1 Alabama 12-1 .324 4 .342 1 .100 31 11.6 .533 5 -.657 4 1.320 28 .555 6
2 Oregon 11-1 .319 2 .334 2 .137 35 10.7 .521 6 -.587 8 1.682 18 .543 11
3 Texas A&M 9-2 .295 7 .221 4 .075 19 9.2 .806 2 -.412 20 -.800 89 .504 57
4 Kansas State 10-2 .279 1 .200 8 .050 5 9.5 .337 21 -.563 10 3.577 1 .577 1
5 Florida 10-2 .259 6 .122 25 .062 11 9.4 .144 43 -.795 1 3.224 3 .536 17
6 Oklahoma 9-3 .243 5 .131 20 .052 6 8.7 .535 4 -.339 25 1.061 34 .512 44
7 Stanford 12-2 .234 8 .124 24 .056 8 11.0 .190 34 -.670 3 1.924 12 .553 8
8 Georgia 11-2 .227 10 .222 3 .083 25 10.3 .445 14 -.475 17 .347 53 .526 30
9 Ohio State 12-0 .216 9 .153 14 .246 61 10.0 .470 10 -.482 15 -.551 82 .505 55
10 Notre Dame 12-1 .215 3 .145 17 .061 10 9.7 .454 12 -.476 16 -.822 90 .492 74
11 South Carolina 10-2 .215 12 .165 10 .115 34 9.2 .244 28 -.629 5 -.624 86 .506 52
12 LSU 9-3 .199 13 .112 27 .048 4 8.5 .125 46 -.605 6 2.195 9 .556 5
Rk Team FBS
Rec
FEI LW GE GE
Rk
SOS Rk FBS
MW
OFEI Rk DFEI Rk STE Rk FPA Rk
13 Florida State 10-2 .195 16 .214 6 .360 80 9.9 .163 41 -.571 9 1.223 31 .539 14
14 Oklahoma State 7-5 .185 15 .131 21 .086 28 8.4 .505 7 -.277 30 1.982 10 .491 75
15 Wisconsin 7-6 .183 11 .130 22 .105 32 9.1 .233 30 -.492 12 -.225 72 .552 10
16 Oregon State 8-4 .182 14 .065 39 .070 17 7.9 .329 22 -.384 23 .250 54 .501 62
17 Texas 9-4 .173 20 .065 40 .066 14 8.5 .400 16 -.198 38 1.280 30 .534 21
18 Baylor 7-5 .172 19 .068 38 .086 27 7.7 .846 1 .238 85 -1.536 102 .510 49
19 Nebraska 9-4 .147 17 .027 54 .074 18 7.6 .504 8 -.239 32 -1.544 103 .457 110
20 Cincinnati 8-3 .146 18 .137 18 .488 98 8.6 .246 27 -.491 13 .014 60 .534 22
21 Northwestern 9-3 .143 22 .075 35 .243 58 8.6 .239 29 -.165 45 2.895 4 .535 18
22 TCU 6-6 .142 24 .010 59 .083 26 6.9 -.103 71 -.561 11 .226 55 .519 38
23 Michigan State 7-6 .142 23 .043 48 .147 37 8.3 -.056 66 -.683 2 .436 50 .499 65
24 Utah State 10-2 .139 29 .210 7 .321 73 9.5 -.038 63 -.491 14 .984 38 .527 29
25 Clemson 10-2 .135 32 .148 16 .291 69 9.0 .446 13 -.058 55 1.404 26 .513 43

Posted by: Brian Fremeau on 09 Jan 2013

23 comments, Last at 11 Jan 2013, 2:05pm by cfn_ms

Comments

1
by sundown (not verified) :: Wed, 01/09/2013 - 6:43pm

"The Fremeau Efficiency Index (FEI) rewards playing well against good teams, win or lose, and punishes losing to poor teams more harshly than it rewards defeating poor teams."
________________________________________

Florida loses soundly to an unranked Louisville team. Somehow, this results in Florida moving UP a spot to #5 while Louisville still can't crack the top 25 despite the win. (That's some punishment for losing!)

Notre Dame loses to the team who played "the best college football game in history" and plummets from #3 to #10 in the final poll. Yet, in contrast, Oklahoma loses by almost the identical score to a poorer-ranked A&M team and drops a single spot from #5 to #6.

Michigan State beats TCU 17-16 and holds steady at #23...only to see TCU leapfrog them in defeat to finish at #22.

5
by Tony D. (not verified) :: Wed, 01/09/2013 - 8:09pm

The bottom of the rankings is even better. Boston College is at #86 at 1-10, three spots ahead of Purdue who was 6-7 with close losses to Notre Dame and Ohio State. Kansas was 0-11 against the FBS but finished ahead of 31 teams, including multiple schools that went 7-5.

10
by Brian Fremeau :: Wed, 01/09/2013 - 9:34pm

Kansas went 0-11. All 11 games came against teams ranked among the top-80. 10 of them came against teams ranked among the top-60. 6 of them came against teams ranked among the top-25 including two single score losses.

Navy has seven wins but is ranked below Kansas. Six of Navy's seven wins came against teams worse than the worst team Kansas played.

East Carolina has seven wins but is ranked below Kansas. All seven wins came against teams worse than the worst team Kansas played.

11
by JonFrum :: Wed, 01/09/2013 - 10:02pm

Given that teams don't play the same schedule, it's perfectly reasonable that a team with a bad record should rank ahead of a team with a good record. The bottom feeders of FBS cant' be considered in the same league as the top programs. Boston College plays quality opponents every year, and has fallen into a hole. That doesn't make them worse than the Little Sisters of the Poor.

14
by Kyle D. (not verified) :: Thu, 01/10/2013 - 5:25pm

But Georgia Tech isn't exactly the Little Sisters of the Poor. And Georgia Tech actually defeated USC in their bowl game. And the FEI loved USC, putting them in the top 30.

And Louisville fails to crack the top 25 despite only losing twice all season AND defeating Florida? That's tough to defend. Every system will have anomalies, but this thing is nothing but anomalies.

18
by Andrew Clayton (not verified) :: Fri, 01/11/2013 - 12:47am

It's the huge disparity in schedules that makes this such a wasted exercise. Teams have no control over their conference games and that's 3/4 of their games. If you play in a major conference with major powers you have a big leg up. If you don't there's nothing you can do. This is a noble effort but chases its tail trying to compare apples to oranges. And based on how the BCS bowls went it's a lousy predictor.

21
by cfn_ms :: Fri, 01/11/2013 - 12:15pm

I don't particularly agree with either of your points:

1) The team with the better schedule strength won more often than not (using Sagarin as a reference http://usatoday30.usatoday.com/sports/sagarin/fbt12.htm - he's a fairly independent source, and uses the simple averaging approach that most associate with SOS calcs):

- FSU (66) over NIU (121);
- Stanford (22) over Wisconsin (30);
- Bama (19) over Notre Dame (21);

2) Teams have a LOT of control over their OOC schedules. Many teams from non-power leagues CHOOSE to avoid challenging themselves in OOC play, typically signing up one tough game where really they should make ALL the slots tough games if they want schedules that are at all comparable to the big boys.

13
by Kal :: Thu, 01/10/2013 - 12:43am

Yeah, this seems more like a feature than a bug. The whole point of DVOA, FEI, S&P is that the record of a team does not capture how good a team actually is, and there are a lot of reasons for a team to lose that do not accurately illustrate the competitiveness of that team. It may not be perfect, but the whole point is to show that teams like Notre Dame with perfect records are not necessarily the best team in the nation, or that a team that is 9-3 could be much better than a team that is 12-0.

15
by Kyle D. (not verified) :: Thu, 01/10/2013 - 5:29pm

I'd agree if that's as far as it went, but it's telling you Louisville wasn't a top 25 team. Anybody out there really believe that? Or that if Boston College played Georgia Tech tomorrow that Georgia Tech wouldn't be the likely winner?

16
by cfn_ms :: Thu, 01/10/2013 - 6:08pm

Do I believe that the team who got hammered by Syracuse, lost to UConn, and barely squeaked by UNC, FIU, USM (0-12!!!), Pitt, USF, Cincy and Rutgers... do I believe that team isn't top 25? Actually... yeah. I do believe they're not top 25.

17
by Andrew Clayton (not verified) :: Fri, 01/11/2013 - 12:31am

Your opinion is as well reasoned as the FEI. And likely just as hard to change. A Louisville could win BCS bowls year after year and never crack the top 25 on this thing unless their conference got stronger. When 75 percent of the games are locked in it's a pointless exercise. Big conference schools are good because they play other big conference schools. That sums up the FEI.

19
by Brian Fremeau :: Fri, 01/11/2013 - 9:18am

Louisville would be ranked highly in FEI if it dominated its weak opposition, but it did not. FEI was not as good a predictor as some other systems this year, systems like Jeff Sagarin's predictor rating which placed Louisville No. 42 after the bowls:
http://usatoday30.usatoday.com/sports/sagarin/fbt12.htm

20
by cfn_ms :: Fri, 01/11/2013 - 12:09pm

I have to agree with that point. It's not "big conference bias" that caused Louisville to barely squeak by FIU and especially a truly atrocious USM team. Those are big negatives that woudl affect ANY team who performed similarly in those games. Ditto losing to UConn and getting hammered by Cuse.

Also, FWIW, Louisville CHOSE to not push themselves hard in OOC play. They could have tried signing up multiple mid-level AQ's as opposed to just UNC (obviously UK is worse than a mid-level AQ). Or they could have tried signing up relatively good non-AQ's (they did try this with USM, but if you're in the Big East [esp. with WVU gone] you can't sign up a AA team and FIU and expect SOS respect, you NEED to fill those spots with better opponents, maybe BYU, maybe other quality non-AQ's).

22
by Kyle D. (not verified) :: Fri, 01/11/2013 - 1:35pm

Sagarin's predictor also had USC in the top 25. It suffers from many of the same problems as FEI. Maybe it's worse, but faint praise there.

What seems to happen is too much weight is put on margin of victory and on opponents' results. When Florida is losing a bowl game by multiple scores to a team FEI thinks is weak but still moving up thanks to other SEC teams winning, there's a problem.

And USC is a great example of how a single game, even a loss, seems capable of really skewing things. Both FEI and Sagarin place USC over UCLA. UCLA won head-to-head, had a better record overall, and had to play Stanford twice which were two of their losses. There's nothing that could possibly give USC an edge other than non-conference schedule and possibly some weird stuff going on with margins of win/loss. UCLA played Nebraska, Rice and Houston non-conference. No AA teams, not a horrible schedule (especially since Nebraska was in the top 20 at the time), won them all. USC had Hawaii, Syracuse and Notre Dame. Beat two but lost to ND. And that's enough not just to rank them ahead of UCLA, but by multiple spots. I'm sure there's some explanation in the data for it, but it won't be anything you could sell to anybody who watched the teams play this year.

23
by cfn_ms :: Fri, 01/11/2013 - 2:05pm

Or maybe, just maybe, USC actually WAS a top 25 team. To review:

1) They won at Washington (which was a road trip that claimed both Stanford AND Oregon St, two very good teams)
2) They were pretty competitive against both Oregon and Notre Dame
3) They won by 13 at Syracuse; at the time, this was considered a negative, but the Cuse was a Big East co-champ, hammered Louisville (if we're comparing Lville to USC, there's one common opponent with VERY divergent results; yeah, you don't want to overrate just one game, but that's still a pretty strong data point), hammered WVU in the bowl game... and lost to Minnesota and NW (which was part of why it was considered a negative). Still a better showing than you'd think.
4) ALL of their losses, except the bowl disaster vs GT, were fairly close to very close, against a pretty nasty strength of schedule. Is that a top 25 team? If you're doing a power ranking, there's actually a pretty decent case for it.

wrt UCLA/USC specifically:

1) There's still 17-43 at Cal, a horrible, HORRIBLE loss;
2) USC had a much nastier in-league schedule than UCLA, getting 5 road games (Victory Bell was at UCLA this year) and drawing Oregon/Washington (compared to Oregon St / Wazzu for UCLA);
3) Losing by 23 to Baylor isn't really much if any better than losing by 14 to GT.

I'm pretty sure my own numbers have USC rated above UCLA; I know pre-bowl they did and I don't think bowl season gave much reason to change it.

8
by mm (old) (not verified) :: Wed, 01/09/2013 - 9:16pm

The system didn't just look at the Sugar Bowl game since the last standings; it looked at all of the bowl games played and used those to re-evaluate every game played this past season.

Florida certainly didn't benefit from losing to Louisville, but all of its results against SEC teams look better after the bowl season and its win against FSU looks better after the ACC had a solid bowl campaign. Having all those games look better was enough to offset the lousy Sugar Bowl performance.

9
by Brian Fremeau :: Wed, 01/09/2013 - 9:21pm

Florida's GFEI (opponent-adjusted game efficiency) for the Louisville game hurt them a little, but the bowl season overall was kind to their season-long performances. The Gators beat Texas A&M, whipped South Carolina, beat Florida State, beat Vanderbilt, and lost a single-score game to Georgia. The net effect of those teams' bowl success kept them from falling far.

For Louisville, it was almost the opposite. The Cardinals second best win of the year according to GFEI came against Pittsburgh, followed by narrow wins over Rutgers and Cincinnati, all of whom underperformed FEI expectations in the bowl season.

Notre Dame's GFEI against Alabama was an albatross, and many of its opponents also no-showed in the postseason (USC, Navy, Purdue, Pittsburgh, Oklahoma). Stanford, BYU, Michigan, and Michigan State did fine, but nothing to boost up the Irish.

Oklahoma's final score may have been similar to Notre Dame's, but the game efficiency wasn't anywhere nearly as devastating. The Sooners were down 34-13 through 18 game possessions when garbage time kicked in (-.333 GE) while the Irish were down 35-0 through 13 game possessions when garbage time kicked in (-.769 GE).

I will say that FEI had more love for the Big 12 overall than I expected (Kansas State and Oklahoma didn't drop the way I thought they would). TCU was a beneficiary of that Big 12 love (best win was a 49-21 drubbing of Baylor). Michigan State's hold had a lot to do with, you guessed it, the performance of their opponents (ND, Michigan, Nebraska, Wisconsin).

2
by Scott C :: Wed, 01/09/2013 - 6:47pm

What is important isn't whether the top two final teams meet in the championship game, but whether the top two teams prior to bowl selection do. It is unlikely that the former will happen because a strong victory in the championship game is very likely to knock the loser down, and a victory for the #3 or #4 team to bump them up past the prior #2.

A playoff system will help, but the effects of strong teams playing each other at the end of the year will always be highly likely to cause the final rankings to differ from the final results of such a playoff.

3
by usctrojan11 (not verified) :: Wed, 01/09/2013 - 8:04pm

im curious as to how arkansas was no.2 despite losing every game of consequence in 2006, including a 50-14 stomping by usc in ark

6
by Tony D. (not verified) :: Wed, 01/09/2013 - 8:21pm

Where are you getting 2006 rankings? I only see them back to 2007 on the site.

But that's no more curious than your 7-6 Trojans coming in at #30 this season ahead of two conference foes that beat them and finished with better records. (UCLA and Arizona.)

12
by Brian Fremeau :: Wed, 01/09/2013 - 10:12pm

Overall FEI ratings from 2003 to 2012 are posted here: http://bcftoys.com/fei

2006 Arkansas ranked No. 10, not No. 2.

4
by bucko (not verified) :: Wed, 01/09/2013 - 8:05pm

Pretty sure Wisconsin was 8-6

7
by Tony D. (not verified) :: Wed, 01/09/2013 - 8:22pm

They only show records against FBS opponents.