FEI Week 15 Ratings

by Brian Fremeau

The Fremeau Efficiency Index principles and methodology can be found here [1]. 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. Strength of Schedule is calculated from a privileged perspective (explained here [2]) and represents the likelihood that an Elite team (top 5) would post an undefeated record against the given team's opponents to date.

The following ratings are calculated based on data from all FBS games played through Saturday, December 6. Only games between FBS teams are considered.

Adjusted Offensive Efficiency and Adjusted Defensive Efficiency are the opponent-adjusted values of Offensive Efficiency and Defensive Efficiency, explained here [3]. Like FEI, the multiple-order adjustments are weighted according to both the strength of the opponent and the relative significance of the result; efficiency against a team's best competition faced is given more relevance weight. AOE and ADE represent a team's value over/under average. Positive AOE and negative ADE are the most valuable.

The Week 15 FEI Ratings for all 120 FBS teams can be found here [4]. Expanded FEI Ratings data can be found here [5].

An Imperfect Consensus

Florida and Oklahoma are the top two teams in the BCS [6], the Associated Press Poll [7], the USA Today Coaches' Poll [8], the Harris Interactive Poll [9], the CBSSports.com BlogPoll [10], the College Football Ranking Comparison [11], and, for good measure, the Fremeau Efficiency Index. Before the BCS and under the old bowl structure, the Gators and Sooners would not have had an opportunity to play against one another in a bowl game. Meanwhile, the BCS itself is almost universally derided by folks on both sides of the college football playoff argument. The playoff proponents cannot fathom why college football does not function like every other professional and amateur sports league known to the Western world. And bowl purists blame the BCS for weakening if not ruining the traditions of New Year's Day.

Do we have a consensus at the top? Consider that the Texas Longhorns hold one of the top two spots on almost as many Associated Press final ballots as the Sooners. Texas has a top-two ranking in three of the six computer systems used by the BCS, whereas Florida is rated as a top-two team by only one of the BCS computers. Oklahoma arguably had the best win of the season in its blowout win over Texas Tech. Florida lost by the narrowest score margin to Mississippi. Texas lost its only game on the road at the last second, and won a head-to-head matchup on a neutral field with Oklahoma. Penn State, USC, Alabama, and Texas Tech all have compelling arguments as well, each having lost only once while navigating varying degrees of challenging BCS conference slates.

In other words, we can agree that the best two teams are playing for the championship and we can agree that the system that produced this game is flawed. We are college football and we are complicated.

One of the great obstacles in resolving the annual college football debate over "Who is No. 1?" is the insistence on determining an answer to the question with absolute certainty. Whether by tallying votes in an opinion poll or processing data through a complex algorithm, ranking systems (including FEI) assign a linear ranking number to each team upon publishing the final output, No. 1 through whatever. The ranking number is obviously more user-friendly and intuitive, even if the ratings themselves (or vote totals and arrays in the case of polls) more accurately describe the relative merits of the teams within that system. Then again, can you imagine a championship game touted not as No. 1 versus No. 2, but rather as No. 0.9757 versus No. 0.9479?

Though it is largely overlooked by everyone other than the team or teams left out of the BCS by mere decimal points, ranking versus rating is an important distinction. The distribution of team power across college football isn't linear, but ranking systems classify teams as though it were. The voted polls have shown some flexibility in recent years, but mostly voters still anchor teams to certain poll positions until a loss forfeits a given team's place in line. The sequence of games played and the timeliness of losses impact the final pecking order as much as any other factor. Florida and Oklahoma have been playing some of their best football of the season down the stretch and they also lost less recently than most of the other contenders. Shake up the order of the Big 12 South triple-header this season and Texas might very well have come out on top in the eyes of writers and coaches.

Ranking numbers are authoritative and rating numbers are descriptive. The team with the No. 1 ranking is the best team. The team with the 0.333 rating played like the best team more often (or more consistently, or more dynamically, or against the best opponents, etc...) than anyone else.

This distinction is why I am not content with simply presenting the final pre-bowl Week 15 FEI Ratings as a validation of the BCS selection of Oklahoma and Florida for the championship. I'd rather try to illustrate the complexity of arriving at that conclusion.

The arguments over team resumes boil down to the following: Who did you play, and how did you play against them? FEI answers the first question and Game Efficiency answers the second. Figure 1 charts the Game Efficiency recorded by every team in 2008 (y-axis) against the power (FEI) of the opponent faced in that game (x-axis). A third variable, the size of the bubble, is plotted according to the consistency of the team's performance -- the smaller the bubble, the less representative the result was relative to the team's other performances. All 1,366 performances of 120 FBS teams in 683 games to date form "The Cloud."

If it helps, you can picture me gesturing wildly in front of a green screen as I attempt to talk you through this storm. The very best performances against the very best teams appear in the upper-right portion of your screen. Moving to the left are performances against weaker competition. Towards the bottom of the cloud are weaker game efficiency performances. We would expect the best teams to populate the upper right perimeter of the cloud, and the seven one-loss contenders are right there in the vicinity. USC's dominant victories over terrible Washington and Washington State, and Oklahoma's beatdown of Washington, stand out in the upper-left end of the spectrum. Texas Tech's obliteration at the hands of Oklahoma is buried in the lower-right. Nearly every other performance falls somewhere in between.

Now, doesn't this clear things up? The difference between Florida's season and Texas Tech's season is somewhat apparent, but for the most part, the elite teams in college football didn't exactly distance themselves from each other or even the rest of the pack. But should we really expect anything more of our top teams? The No. 1 and No. 2 teams in the country don't perform a certain way against a certain level of competition week in and week out. Instead, they perform more consistently like the best team than others do, they never perform terribly, they appropriately stomp on weak opposition, and they are at least sometimes very successful against good and very good opponents.

The FEI ratings are a kind of one-dimensional representation and synthesization of the three-dimensional cloud. Florida and Oklahoma measure up to the expectations we have of a No. 1 versus No. 2 match-up, and the game has the potential to be an outstanding clash of the titans. Since 2003, FEI top-10 teams have met only 48 times, FEI top-5 teams have met only ten times, and No. 1 and No. 2 have met only once. If Florida and Oklahoma can deliver some of the magic of the 2005 Texas/USC finale, we'll have an extraordinary champion, albeit an imperfect one.

An Imperfect Solution

I am neither pro-BCS nor anti-BCS, and I do not have a playoff system proposal to share. I don't dispute the financial windfall that a college football playoff would garner. I have just illustrated the challenge of distinguishing two and only two teams to play for a national championship, and I don't dispute that a great way to help sort out similar team profiles would be to actually have the teams play against one another. I don't dispute that a national college football playoff of four, eight, or 16 teams would be one of the most entertaining and engrossing series of events in all of sports.

I do dispute, however, that a single playoff solution, any playoff solution, is necessarily better for college football. The best solution for one season is rarely the best solution for another season. An eight-team playoff with six automatic bids would leave two at-large spots this season for Alabama, Texas, Texas Tech, Utah and Boise State to fill. A "Best Eight" playoff might be a great fit for 2008, but would have left out Hawaii in 2007 -- the only undefeated team heading into the bowls -- and several major conference champions every year. If the conferences would agree to it, maybe that's a price worth paying, but I'm still not sure such a system would be best for college football on the whole.

Consider that one-loss Texas Tech would make the 2008 "Best Eight Playoff" field and two-loss Ohio State would be on the outside looking in. Then consider that Ohio State played USC out of conference and lost, and Texas Tech won each of its non-conference games against Eastern Washington, Nevada, Massachusetts, and SMU. Might Ohio State (and other BCS contenders) reconsider scheduling even one strong out-of-conference opponent in the future in order to more easily qualify for such a playoff? When anti-playoff folks fear for the preservation of the regular season, they aren't necessarily talking about protecting the importance of the Red River Rivalry.

College football's intersectional games are worth protecting. I'll go a step further. The number of college football's intersectional regular season games needs to grow, and the energy spent barking about playoffs and inventing postseason solutions would be better spent first on creating more top matchups in college football's regular season. Let's fight for Texas Tech and Georgia Tech to skip one of their two FCS games in 2008 and play each other instead. Let's insist that the top conferences meet some minimum scheduling expectations with other conferences. Let's reward that type of change with lucrative TV dollars and a playoff-like atmosphere, but let's do it in September first before we worry about fixing the championship in January.

We are college football and we are imperfect. But that doesn't mean we can't do better, and it certainly doesn't mean we can't improve on the greatest regular season in sports.

Perfection On the Line

Two teams remain perfect in 2008, and they have been scarcely mentioned by me all season long. Mountain West Conference champion Utah is headed to its second BCS bowl game appearance in five years and Boise State completed its undefeated regular season and claimed the WAC championship. Neither team is considered by FEI to be on par with the elite teams, however. What gives? Does FEI have some kind of mid-major bias?

Not exactly. FO readers might recall that FEI correctly forecasted Boise State to defeat Oklahoma in the 2006 Fiesta Bowl [13]. And four non-BCS teams currently appear in the FEI top 25, more than any other year in the FEI era. The following table lists every season-ending FEI top-25 finish for non-BCS conference teams since 2003.

Note the absence of the 2007 Hawaii Warriors from this list, a team rated No. 53 by FEI and totally overmatched in last year's Sugar Bowl. Utah will face SEC runner-up Alabama in the Sugar Bowl this season, and does figure to be more competitive. But not only are the Utes rated as the third-best non-BCS conference team in 2008, they are only the third-best Utah team on this list since 2003.

The Urban Meyer-led 2004 Utah team broke through to a BCS game victory and are the only "mid-major" to complete a season ranked among the FEI elite. What was their secret? They thoroughly dominated their schedule, recording a season Game Efficiency average of 0.398 (only Meyer's Florida Gators have a better GE in 2008). Their narrowest non-garbage margin of victory was 17 points, and they outscored their opponents in non-garbage possessions by 335 points, nearly 28 per game.

This isn't to suggest that Utah hasn't had a successful season in 2008, but they do not have the profile of an elite team. If plotted in The Cloud, the narrow escapes over Michigan and New Mexico would appear more lower-left than any of the performances of the seven one-loss teams, and Utah never notched a victory in the upper-right region. Against their best opponents, TCU and Oregon State, the Utes won each game by three points at home. Alabama will be a formidably greater, though not impossible, challenge.