Writers of Pro Football Prospectus 2008

Most Recent FO Features

FitzgeraldLar04.jpg

» Scramble for the Ball: The DVOA Schism

Mike and Tom try to figure out what kind of secret sauce Arizona is feeding the media to sit at the top of the power rankings and in the middle of our DVOA rankings.

06 Dec 2012

FEI Week 14: Wildcat Strike

by Brian Fremeau

Five weeks ago, four teams had established themselves as the clear contenders in the BCS championship race. Alabama and Oregon sat atop the major polls, Kansas State and Notre Dame had won over the computers, and it appeared that we were in for a fiercely contested debate down the stretch to determine the national championship game participants among those four contenders.

FEI agreed, and from that point until today, those four teams have remained at the top of the FEI ratings each week. Alabama lost to Texas A&M, Oregon lost to Stanford, and Kansas State lost to Baylor. Those results and others shuffled the top four rankings each week, but the four remained the same. Those that follow FEI know that this isn’t an unusual phenomenon. Though the opponent-adjusted drive efficiency calculations I run each week are complex, I’m rarely surprised by the FEI output.

And yet, this year’s final regular-season ratings do challenge assumptions I’ve made about FEI, starting at the top. The Kansas State Wildcats are ranked No. 1 this week, a leap up from No. 4, following a victory over the Texas Longhorns last weekend. I expected that kind of jump might have been in store for Alabama after their SEC championship game win over Georgia, but it didn’t work out that way. What is it about the Wildcats that FEI finds to be so special?

The short answer is that FEI finds the Big 12 to be pretty special, and Kansas State earns a lot of credit for winning the conference, and winning in dominant fashion against many of its league foes. The Wildcats rank among the top 20 in Offensive FEI, Defensive FEI, Special Teams Efficiency, and Field Position Advantage, and KSU ranks in the top 5 in three of those measures. Alabama is the only other team to rank in the top 30 in those four measures. Stanford is the only other team to rank in the top 40 in those four measures. In other words, Kansas State distinguished itself as the best team in the country according to the primary efficiency components of FEI.

And yet, the Wildcats defied one of the primary philosophical components of FEI by getting blown out at Baylor by 28 points back on November 17. The Bears ended up with a very respectable No. 19 ranking in FEI, so the loss itself shouldn’t have been crippling, but it is extraordinarily rare for a team to lose badly and not have that significantly impact their FEI rating. Should that loss carry more weight than I am already assigning to it? That’s a question I’m going to have to explore in the offseason.

Speaking of the offseason, as you know, I have been using the Program FEI ratings (five-year weighted drive efficiency) as the foundation of my preseason projections. The Kansas State Wildcats were ranked No. 48 in PFEI heading into 2012, and if their FEI 2012 ranking holds, they will have made the biggest leap to No. 1 by far for any team I have ever tracked. Prior to this year, the lowest ranked PFEI team to ascend to No. 1 in next-year FEI was the 2003 LSU Tigers (PFEI No. 26 heading into 2003). Every other team in the last decade to ascend to No. 1 in FEI was ranked among the top 20 in PFEI. Kansas State is only the third team in the last 10 years to ascend to the top 5 in FEI from outside the top 40 in PFEI.

The bowl season is about to challenge other assumptions I have made about FEI. This is the first season since I’ve been calculating FEI that neither of the BCS championship game participants are ranked No. 1 in FEI. Neither is ranked No. 2, either. Kansas State’s leap and Alabama’s non-leap this weekend means that the Fiesta Bowl, not the BCS championship game, is the marquee matchup according to FEI.

The FEIsta Bowl, if you will, features Kansas State against the Oregon Ducks, another team to challenge assumptions I’ve made throughout the year. Oregon was held back below the other contenders for most of the season due to schedule strength, and though they did play their toughest opponents down the stretch, I wasn’t sure it would be enough to move ahead of the others. Not winning their division due to the loss to Stanford, and not playing in the Pac-12 title game was probably going to hurt their final FEI rating, I assumed. Instead, I think not playing the game might have helped.

Remember that I include a "relevance" factor in weighting the value of each game to calculate opponent adjustments. Oregon didn’t play many top teams, but the relevance of those games actually works in their favor. Beating Oregon State in the final week of the year may not have been enough to bump them forward if they had a few other wins against top-20 teams. But that data point is so much more significant than their collection of wins against very weak teams, it did provide the bump Oregon needed to leapfrog Notre Dame and Alabama.

There’s no intention to reward a weak schedule in FEI, and for the most part, FEI does the opposite. But Oregon is a unique case -– No. 1 in raw Game Efficiency against many weak teams, big wins against two top-25 teams (USC and Oregon State), and an overtime loss to a top-10 team. Those three results suggest to FEI that the Ducks are definitely one of the nation’s best, but the relative insignificance of most of their other wins gives them a boost. Should that be the case? Again, this is something I’m going to look more closely at in the offseason.

The way the FEI ratings currently shake out, I fully expect that the winner of the Fiesta Bowl will be ranked above the winner of the BCS championship game in the final FEI ratings to be published after the bowls. The top four teams will almost certainly remain the top four teams at the end of the year. But since I’ve had many other assumptions challenged this year, I guess I should hold off on making any guarantees.

Week 14 Revisionist Box Scores

This weekly feature identifies the games played each week that were most impacted by turnovers, special teams, field position, or some combination of the three. The neutralized margin of victory is a function of the point values earned and surrendered based on field position and expected scoring rates.

Week 14 Games In Which Total Turnover Value Exceeded Non-Garbage Final Score Margin
Date Winning Team Non-Garbage
Final Score
Losing Team TTV
+
TTV
-
TTV
Net
TO Neutral
Score Margin
11/29 Louisville 20-17 Rutgers 8.6 4.1 4.5 -1.5
11/30 Stanford 27-24 UCLA 7.6 0.0 7.6 -4.6
12/1 Boise State 27-21 Nevada 6.3 0.0 6.3 -0.3

Week 14 Games In Which Special Teams Value Exceeded Non-Garbage Final Score Margin
Date Winning Team Non-Garbage
Final Score
Losing Team STV
+
STV Neutral
Score Margin
11/29 Louisville 20-17 Rutgers 4.1 -1.1
11/30 Stanford 27-24 UCLA 3.8 -0.8

Week 14 Games In Which Field Position Value Exceeded Non-Garbage Final Score Margin
Date Winning Team Non-Garbage
Final Score
Losing Team FPV
+
FPV
-
FPV
Net
FPV Neutral
Score Margin
11/29 Louisville 20-17 Rutgers 23.1 18.5 4.6 -1.6
11/30 Stanford 27-24 UCLA 23.0 14.6 8.4 -5.4
12/1 Baylor 41-34 Oklahoma State 29.0 18.4 10.6 -3.6
12/1 Tulsa 33-27 Central Florida 33.6 26.6 7.0 -1.0

2012 totals to date:

  • Net Total Turnover Value was the difference in 113 of 696 FBS games (16.2 percent)
  • Net Special Teams Value was the difference in 56 of 696 FBS games (8.0 percent)
  • Net Field Position Value was the difference in 70 of 696 FBS games (10.1 percent)
  • Turnovers, Special Teams and/or Field Position was the difference in 158 of 696 FBS games (22.7 percent)

2012 Game Splits for all teams, including the offensive, defensive, special teams, field position, and turnover values recorded in each FBS game are provided here.

FEI Week 14 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 Pvs: Strength of schedule to date, based on the likelihood of an elite team going undefeated against the given team's schedule to date.
  • SOS Tot: Strength of schedule, based on the likelihood of an elite team going undefeated against the given team's entire schedule, including bowl games.
  • 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.
  • FBS RMW: Remaining Mean Wins, the average number of games a team with the given FEI rating would be expected to win against its remaining 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 December 1st. 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
Pvs
Rk SOS
Tot
Rk FBS
MW
FBS
RMW
OFEI Rk DFEI Rk STE Rk FPA Rk
1 Kansas State 10-1 .305 4 .243 3 .111 22 .050 6 10.0 0.5 .335 20 -.598 5 4.529 1 .591 1
2 Oregon 10-1 .299 1 .342 1 .246 56 .107 30 10.4 0.5 .501 10 -.587 7 .807 41 .542 13
3 Notre Dame 12-0 .289 2 .193 10 .129 28 .068 14 10.9 0.5 .518 6 -.750 2 -.768 89 .495 68
4 Alabama 11-1 .279 3 .319 2 .219 50 .107 29 11.1 0.5 .364 19 -.583 8 1.245 30 .554 7
5 Oklahoma 9-2 .273 5 .163 14 .073 9 .046 4 9.2 0.6 .539 5 -.501 16 .919 38 .512 47
6 Florida 10-1 .260 6 .143 19 .139 32 .124 32 9.9 0.9 .102 50 -.762 1 3.165 3 .548 10
7 Texas A&M 8-2 .242 7 .212 7 .192 42 .105 27 8.6 0.4 .669 2 -.273 31 -.595 85 .505 58
8 Stanford 11-2 .229 8 .127 26 .053 2 .038 2 10.5 0.6 .165 39 -.705 4 1.862 12 .559 4
9 Ohio State 12-0 .224 9 .153 17 .222 52 .222 56 10.0 - .508 9 -.504 15 -.551 83 .505 57
10 Georgia 10-2 .199 11 .230 6 .164 35 .128 34 10.1 0.6 .368 17 -.381 23 .676 48 .531 26
11 Wisconsin 7-5 .198 19 .146 18 .141 33 .093 23 9.3 0.4 .251 29 -.555 10 -.086 66 .551 9
12 South Carolina 9-2 .191 13 .177 12 .169 36 .138 39 9.0 0.6 .126 43 -.587 6 -.201 73 .504 59
Rk Team FBS
Rec
FEI LW GE GE
Rk
SOS
Pvs
Rk SOS
Tot
Rk FBS
MW
FBS
RMW
OFEI Rk DFEI Rk STE Rk FPA Rk
13 LSU 9-2 .187 14 .126 27 .118 25 .103 26 8.7 0.7 .124 44 -.540 12 1.973 10 .553 8
14 Oregon State 8-3 .187 15 .076 36 .102 20 .081 19 8.1 0.6 .407 13 -.376 24 .134 58 .499 66
15 Oklahoma State 6-5 .184 12 .095 32 .061 6 .060 10 8.3 0.9 .517 8 -.263 33 1.538 20 .475 91
16 Florida State 9-2 .169 16 .210 8 .442 87 .374 80 9.6 0.6 .078 52 -.522 14 1.705 18 .543 12
17 Nebraska 9-3 .163 10 .043 50 .093 16 .066 11 7.9 0.4 .548 4 -.321 28 -1.574 104 .450 113
18 Cincinnati 7-3 .155 18 .136 22 .515 100 .507 100 8.9 1.0 .242 32 -.526 13 .841 40 .541 15
19 Baylor 6-5 .155 23 .050 46 .069 8 .057 7 7.1 0.6 .836 1 .341 99 -1.236 96 .518 44
20 Texas 8-4 .154 17 .067 40 .064 7 .048 5 8.1 0.4 .364 18 -.128 48 1.472 23 .540 17
21 Michigan 8-4 .143 21 .132 23 .061 5 .044 3 7.9 0.4 .289 26 -.416 20 .614 49 .483 81
22 Northwestern 8-3 .142 22 .069 39 .243 55 .237 59 8.5 0.8 .268 28 -.156 41 3.058 5 .532 24
23 Michigan State 6-6 .140 26 .046 47 .118 24 .097 25 7.9 0.5 -.057 63 -.711 3 .075 61 .495 67
24 TCU 6-5 .139 20 .012 61 .081 14 .067 12 6.9 0.5 -.137 78 -.543 11 .289 54 .524 34
25 USC 7-5 .137 25 .102 30 .075 10 .072 15 8.1 0.8 .281 27 -.188 37 1.540 19 .519 43

Posted by: Brian Fremeau on 06 Dec 2012

10 comments, Last at 09 Dec 2012, 10:30pm by Brian Fremeau

Comments

1
by RickD :: Thu, 12/06/2012 - 12:26pm

I love that these rankings have a bunch of 5- and 6- loss teams and doesn't feel the need to include Louisville simply because they won the Big East.

2
by Will Allen :: Thu, 12/06/2012 - 1:15pm

Wouldn't it be fun if, on Saturday, Wisconsin was playing at Notre Dame, Florida State at Alabama, Georgia at Stanford, and Oregon at Kansas State? Yeah, I know Florida is getting left out, and Wisconsin has 5 losses, but if we accept the reality that the major conferences (Big 10, Pac 12, SEC, ACC, Big 12) are going to demand at least one representative in a tourney, and have a rule where you don't penalize a division winner for losing a conference championship game, to favor a team the division winner beat in the regular season, and limit each conference to two participants, it seems like an entertaining slate of games.

I think you could auction the t.v. rights to four games like that, on the 2nd Staurday in December, with two semi-finals on January 1st at two of the major Bowl sites, with a Championship 7 to 14 days later, at whatever suitable city bid successfully for the last game, for a pretty fair sum. Just a guess.

4
by Mood_Indigo (not verified) :: Thu, 12/06/2012 - 1:24pm

The Stanford players have to study for their final exams.

5
by Will Allen :: Thu, 12/06/2012 - 1:35pm

Is that considered when Stanford's basketball team takes a month to play in their conference tourney, and then the NCAA tourney? Or are athletes who play with a brown oblong ball just considered intellectually and academically different than athletes who play with a round orange ball?

8
by My name is my password (not verified) :: Fri, 12/07/2012 - 10:11am

With your parameters, why does Oregon make it in? They didn't win their division. Thus, if Stanford would have lost their title game, you would be comfortable leaving Oregon at home?

Putting in Georgia over Florida basically rewards weak schedules. One cannot say Georgia>Florida from the win, as Georgia>Florida>USC>Georgia circle cancels any transitive property assumption

9
by Will Allen :: Fri, 12/07/2012 - 10:22am

Because the reality of the situation is that a majority of the major conferences will likely demand that no single conference have more than two participants, just like they will demand that each major have at least one participant. Personally, I don't care which third place finisher gets shafted.

3
by DisplacedPackerFan :: Thu, 12/06/2012 - 1:19pm

I'm not convinced that your #1 and #2 are wrong either.

Last year at this time 1-4 in FEI were at 0.318, 0.297, 0.255, 0.252
Those 4 teams ended up at 0.318, 0.311, 0.235 (#5), 0.306(#3). The new #4 was 0.240. LSU, OK State, and Alabama were great teams last year.

This year they are 0.305, 0.299, 0.289, 0.279. #5 and #6 are 0.273 and 0.260.

These are good teams, the gaps just aren't as big. It's possible that you are right and that a little bit of noise is shuffling them around more than it should be, or that they really are really good to great teams and trying to put them in the proper order is hard.

OK I'm just gonna treat them like batting averages and drop the decimals.

I wasn't able to quickly find the pre-bowl numbers for 2010, but the final numbers looked more like 2011, a couple 300 plus teams, and then a clear drop off. 2009 was even more stark in the drop offs.

It seems pretty rare to have more than 4 teams above a 260. (3 in 07, 4 in 08, 2 in 09, 4 in 10, 3 in 11) This year there are currently 6. It also seems rare for the top team to end up under 340 (07 it was 282, 08 was 356, 09 was 342, 10 was 348, 11 was 318). So the "best" isn't as dominate this year either. Now the bowl games are going to provide some bump too, but I don't see any team getting up past 315.

I still think these rankings are solid and more accurate than many others out there. It's good to challenge assumptions, but don't assume they are wrong. It's possible that the rest of the conventional wisdom that has Notre Dame and Alabama as #1 and #2 is wrong. Most of the other metrics agree that all four of these teams are really good and most other metrics have had failures as well in the past "getting it right".

6
by Kal :: Thu, 12/06/2012 - 4:50pm

What is the actual predictive value of special teams and field position as far as winning games go? I know that DVOA ranks special teams as 1/7th of the value; I don't quite know what field position does to that overall value other than giving some amount of expected points.

Similarly, it sounds like both special teams and field position are considered 'lucky' - otherwise the revisionist box score wouldn't be the way it is. Is that the case? How much can we expect special teams and field position to continue on a game to game basis? how much variance is there?

7
by Brian Fremeau :: Thu, 12/06/2012 - 5:01pm

The relationship between efficiency stats and winning percentage is illustrated in this chart: http://www.bcftoys.com/home/2011/12/18/fei-data-update.html

It only recently occurred to me, but I could run a set of FEI ratings based on the revisionist box scores. There could be a turnover-neutral FEI, a special teams neutral FEI, a field position neutral FEI. I think I ought to do that this off-season and find out if those ratings (or maybe a combination of alternate FEIs) might be more predictive.

10
by Brian Fremeau :: Sun, 12/09/2012 - 10:30pm

I'll definitely devote another column altogether to this idea, but I did run a turnover-neutral, special-teams-neutral, and field-position-neutral FEI. Each of them had the same top-6 teams, just in a different order for each:

Turnover-neutral
1. Oklahoma
2. Notre Dame
3. Alabama
4. Oregon
5. Texas A&M
6. Kansas State

Special-Teams-neutral
1. Oregon
2. Notre Dame
3. Kansas State
4. Oklahoma
5. Alabama
6. Texas A&M

Field-Position-neutral
1. Notre Dame
2. Oregon
3. Oklahoma
4. Kansas State
5. Alabama
6. Texas A&M