by Brian Fremeau

As the clock wound down in the fourth quarter in Eugene, Oregon, Saturday night, you couldn't help but be stunned by the unfamiliarity of it all. Pete Caroll appeared dumbfounded as he paced the sideline. The incessant cadences and refrains from the Trojans' band had been silenced. The USC players on the field hopelessly flailed at and chased after an Oregon attack running wild out of the backfield and through the secondary.

It was a cardinal and gold bloodletting, the first of its kind in over a decade. USC had delivered that kind of blow to opponents time and time again over the years but hadn't suffered one of its own since Matt Barkley was playing Pee Wee football. Sure, the Trojans had stumbled inexplicably against inferior opponents a handful of times under Carroll. But they hadn't ever been stomped. Before Halloween night, USC had lost only once by more than a single score, an 11-point loss to Notre Dame in 2001, Carroll's first season. The 27-point loss to Oregon concluded one of the most dominant runs in college football history.

In 1978, college football's Division 1 split into 1-A and 1-AA, now referred to as the Football Bowl Subdivision (FBS) and Football Championship Subdivision (FCS). In the 31.5 seasons since, no team has come close to matching USC's run of consecutive games played without ever losing by more than a single score.

Big deal? Yeah, kind of. USC has held the No. 1 spot in the Program FEI ratings every year since 2003. In every single one of the past six seasons, the Trojans have finished No. 1 or No. 2 in the final FEI ratings for that year. The have out-recruited, out-coached, out-swaggered, out-Heismaned, out-camaraderied, and out-performed the rest of college football since 2002. And until Saturday night, when they hiccupped, they still gave themselves a chance at victory by keeping the game close.

One of the key insights to understanding the FEI methodology is the way poor performances as measured by Game Efficiency (GE) can have a major impact on a team's rating. Those that follow the FEI ratings week to week have recognized the way its top-down approach to strength of schedule places a premium on strong performance against strong competition, win or lose. But big losses can have a big impact, especially at the top of the ratings. Why? Because elite teams almost never get stomped. ("Elite" teams are defined roughly as "top 5"; specifically, they are defined here as having an FEI rating at least 1.66 standard deviations above average).

The FEI ratings system has been processing data since 2003, and in that span only three "Elite" teams have ever suffered a worse defeat than USC (-.33 GE) did on Saturday:

• -.41 GE - 2005 Virginia Tech (11-2, No. 5 FEI) lost to Miami (9-3, No. 12 FEI) 27-7; 27-0 in 19 non-garbage possessions
• -.43 GE - 2005 LSU (11-2, No. 6 FEI) lost to Georgia (10-3, No. 7 FEI) 34-14; 34-7 in 18 non-garbage possessions
• -.52 GE - 2004 Oklahoma (12-1, No. 4 FEI) lost to USC (13-0, No. 1 FEI) 55-19; 48-10 in 21 non-garbage possessions

Plenty of non-Elite teams have suffered losses of that magnitude, which is precisely what makes that particular statistic significant for the Trojans. Unless USC can completely dominate the remainder of its schedule and Oregon can continue to ascend as an Elite opponent through its remaining slate, the -.33 Game Efficiency for the Oregon game will likely be an albatross for USC's FEI rating.

To better illustrate the relationship of Game Efficiency and opponent FEI rating, I plotted every game from 2003 to 2008 in Figure 1. Actually, every game appears twice, the GE (y-axis) against the opponent's FEI rating (x-axis) for each team in the match-up. The color of the dot represents the given team's own FEI rating (not the opponent's), broken down into general power categories. Strong performances against strong teams appear toward the upper-right portion of the graph.

Note that the vast majority of teams posting a -.33 rating or worse in any game over the last six years fall into the Average category or worse. Note also that the primary distinction between the "Elite" cloud (black dots) and "Very Good" cloud (purple dots) is the relative frequency of poor GE results. Perhaps it is common sense, but again, Elite teams almost never get stomped. Very Good teams do from time to time.

Now that USC's streak has been broken, which teams may be making a run of their own? Through last weekend, both Texas and Utah lead the way, each currently boasting a 29-FBS-game streak of never having lost by more than a single score. Florida (25), Boise State (20), Iowa (20), Ohio State (19), and TCU (15) are the only others in double digits. Pete Carroll's Trojans (0) can kick off their next streak Saturday at Arizona State (1).

Week 9 FEI Top 25

The principles of the Fremeau Efficiency Index (FEI) 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 (GE), 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. Strength of Schedule (SOS) is calculated as the likelihood that an elite team would win every game on the given team's schedule to date. SOS listed here does not include future games scheduled.

Only games between FBS teams are considered in the FEI calculations. Preseason projections are no longer included in the calculations. Current FEI ratings are a function of results of games played through Nov. 1.

FEI ratings for all 120 FBS teams are now listed in the stats page section of FootballOutsiders.com. Click here for current ratings; the pull-down menu in the stats section directs you to 2007 and 2008 ratings.

Posted by: Brian Fremeau on 04 Nov 2009

6 comments, Last at 07 Nov 2009, 11:29am by Brian Fremeau