Will Adrian Peterson leave Minnesota for a warmer climate in 2015?
17 Oct 2007
by Brian Fremeau
The Fremeau Efficiency Index (FEI) principles and methodology can be found here. Like DVOA, FEI rewards playing well against good teams, win or lose, and punishes losing to poor teams more harshly than it rewards defeating poor teams. Unlike DVOA, it is drive-based, not play-by-play based, and it is specifically engineered to measure the college game.
FEI is the opponent-adjusted value of Game Efficiency, a measurement of the success rate of a team scoring and preventing opponent scoring throughout the non-garbage-time possessions of a game. Like DVOA, it represents a team's efficiency value over average.
Only games between FBS (Division 1-A) teams are considered. The Week 7 Ratings represent the results of all games played through Sunday, October 14, 2007.
We have reached halftime of the 2007 college football season, the bands are trotting out onto the field, and most fans across the country are just trying to catch their breath. Four big-time programs -- USC, LSU, Oklahoma, and Florida -- appeared virtually unstoppable in late-September and then dropped a combined five games in three weeks, three to previously unranked opponents. Zero AP preseason top ten teams remain unbeaten. At least three teams have recorded their highest poll ranking in decades, if not ever. Three teams each arguably suffered the greatest upset of all time. Are we witnessing the most superlative college football season in history?
Maybe. Or maybe we just need a little more perspective. As shocking as Stanford's defeat of USC on October 6 may have been to everyone from oddsmakers to pollsters, should the ultimate value of an upset be judged exclusively against the expected game result? When the No. 1 team in the country loses to a previously hapless team, what have we learned? Were we more wrong about the strength of the top team or the hapless one? A bit of both? Or were we wrong about neither team, and truly just witnessed something historic?
One of the criticisms of voted polls relates to the notion of poll-anchoring -- individual voters treating previous-week ballots as rigid rankings to be adjusted only as needed once new isolated (upset) results are recorded. USC's fall from the top to midteens in the polls more likely represents a consensus "punishment" for losing to Stanford than it does a revised evaluation of the strength of the Trojan team. Likewise, the ascension of Ohio State and Boston College in the polls can be more heavily attributed to similar punishments doled out to other "upset" victims over the first seven weeks than to the actual merits of the Buckeyes and Eagles.
Computer ranking systems generally do re-evaluate past data along with new data each week, resulting in rankings less stable than the polls, particularly early in the season. The ranking logic of a given computer system is quite stable, however. Many of the computer evaluations of each of USC's results -- almost exclusively recorded against relatively underwhelming competition -- have treated the Trojans less like a victimized elite team and more like the dozens of other teams that populate the top third of college football. As a result, the Stanford loss at this stage in the season measures as an upset, but not necessarily a historically significant one.
What, though, do we make of the seemingly endless series of upsets that have rocked the college landscape in this wild season to date? Kenneth Massey's College Football Ranking Comparison includes more than 100 human and computer ranking system results, and produces an average consensus ranking for each team. Massey calculates a "Ranking Violation Percentage" for the consensus, representing the percentage of games played in which, retrodictively, a lower ranked team defeated a higher ranked team. To date in 2007, the consensus Ranking Violation Percentage is 12.8 percent. Through the first seven weeks in 2006, the consensus Ranking Violation Percentage was 12.6 percent. By this measure, the upset frequency in college football has been severely exaggerated thus far this season by the media. Perhaps at season's end we will be able to view these results in the proper context.
FEI's most significant differences from the newly-released AP poll and BCS ratings lie with the treatment of zero-loss and two-loss teams. The voter's inclination to elevate undefeated teams and demote two-loss teams based on record alone is a safe decision, if not a particularly thoughtful one. FEI considers some blemished resumes as more impressive than some unblemished ones, for now. If Ohio State and Boston College impress in the coming weeks, I expect them to move up in the rankings. Several poll-underrated and -overrated teams are worth examining further.
Auburn (5-2; No. 17 BCS, No. 18 AP, No. 3 FEI)
The Tigers are winners of four straight, have defeated No. 7 Florida, No. 14 Kansas State and No. 24 Arkansas, and suffered single-score losses to No. 2 South Florida and No. 45 Mississippi State. A showdown with No. 1 LSU awaits this weekend, and a victory would put Auburn in the driver's seat for the SEC West title. Regardless of the outcome on Saturday, expect Auburn to be one of several dangerous teams with blemished records in the top 10.
Kansas State (3-2; NR BCS, No. 25 AP, No. 14 FEI)
Losing only to No. 3 Auburn and No. 11 Kansas, Kansas State is primed to steamroll opponents throughout the next month as they did against Colorado and Texas earlier this year, particularly now that Nebraska has fallen of the face of the earth. A November 17 matchup with No. 17 Missouri could have Big12 title game implications if Kansas falters.
Texas (5-2; No. 22 BCS, No. 19 AP, No. 37 FEI)
Just as they did last season, Texas has feasted on NCAA bottom-half opposition, and have often looked ordinary in the process. Their explosive offensive games came against No. 116 Rice and No. 110 Iowa State. In the Longhorns' losses to No. 14 Kansas State and No. 5 Oklahoma, they were outscored 69-42. Texas shouldn't face a significant challenge until No. 26 Texas Tech visits on November 10 -- a stretch that will likely inflate their record and ranking to unsustainable levels.
Georgia (5-2; No. 20 BCS, No. 21 AP, No. 38 FEI)
The meat of Georgia's schedule will come down the stretch after an off-date on Saturday, with matchups against No. 7 Florida, No. 3 Auburn, and No. 4 Kentucky all looming in the next month. They played No. 10 South Carolina close, but a beatdown at the hands of No. 21 Tennessee and a near disaster at No. 74 Vanderbilt mark an alarming trend. I don't expect a great final record out of the Bulldogs, but the strength of the SEC will likely keep Georgia's FEI rating from plummeting.
I suggested in the FEI Season Preview article two months ago that the change in college football's kickoff yard line from the 35 to the 30 might significantly alter average starting field position and offensive efficiency for all teams. Thus far, average starting field position following kickoff returns has advanced three yards, from a team's own 26-yard line in 2006 to its own 29-yard line in 2007. Overall field position has advanced from a team's own 31-yard line to its own 32-yard line. In several cases, individual team field position results have been much more dramatic.
Illinois (4-2) and Indiana (4-2) each have improved their total field position advantage (the difference between a team's average starting field position and its opponents') by 13 yards per possession from 2006 to 2007, topping all other teams. On the flip side, Rutgers (3-2) and Southern Mississippi (2-3) have given up more than 14 yards in field position advantage per possession since last season. Here are the most advantaged and disadvantaged teams by field position thus far in 2007.
|Field Position Advantage (FPA) Through Week 7|
|TEAM||2007 W-L||2007 FPA|
|North Carolina State||(0-5)||-7.4|
|New Mexico State||(1-4)||-9.5|
Good field position is a function of special teams execution, offensive drive sustainability, and protecting the football. As most of the bottom ten on this list can attest, significant field position deficits can be absolutely crippling.
36 comments, Last at 22 Oct 2007, 8:41am by John Grenci