FEI Week 11 Ratings and Elite Win Margin
by Brian Fremeau
We need to talk more about strength of schedule. As the end of the regular season approaches, conference championships will be contested, the College Football Playoff committee will make its final decisions on the four participants in the national championship semifinal games, and debates around strength of schedule will escalate. How much value should be placed on schedule strength when debating the merits of teams? How is strength of schedule best calculated, and why do some measures produce drastically different results than others? And how can we better understand the terms of how the debate will be framed, and how perhaps it should be framed?
I have done a good deal of work on the subject, much of it focusing on the difficulty of settling on how best to measure schedule strength. I produce three different schedule strength outputs, in fact, because perspective matters. An elite team will only reasonably be challenged by teams in the top 30 or so, so every opponent worse than that may as well be considered equally difficult to defeat. A good team will be reasonably challenged by a much larger swath of teams. An average team will have difficulty competing with elite teams, will almost certainly dominate awful teams, and may be reasonably competitive against nearly everyone else. If we're measuring the difficulty of a given schedule, then, it matters which perspective is taken into account. The difficulty of winning games as an average team has more to do with how many games are played against the meat of the bell curve. The difficulty of winning games as an elite team has much more to do with how many games are played against other top teams.
There are other ways to rank schedules as well. ESPN produces something they call "strength of record" as part of their playoff picture ratings, defined as "the chance that an average Top-25 team would have team's record or better, given the schedule." These are only published as a ranking number, but they are designed to reflect some of the CFP committee's priorities. My colleague Bill Connelly has kicked the tires on strength of schedule arguments as well, resulting in his "Resume S&P+" ratings, designed to measure something akin to strength of record, but considering margins of victory or defeat against the schedule as well.
I have also attempted to produce a rating that attempts to measure the selection committee's priorities, a rating output of the FEI system I called "degree of difficulty" (DOD). The ratings were a reasonably good approximation of the committee's ultimate selections from 2014 through 2016, but I abandoned it as a concept last year, in part because I was torn between devising something that accurately predicts what the committee would do, versus devising something that speaks to what I think the committee should do as far as their priorities.
I'm leaning more toward the latter, lately. I know there are many good reasons why a selection committee was established to select the four participants in the College Football Playoff, and I don't believe a single computer output should replace that process. But I do wish the committee was more consistent in its process and leaned a bit more on the advanced analytics work that could help inform that process. In that spirit, I'm presenting a revised version of the "degree of difficulty" work I produced previously, but rooted in the same concepts. In my view, the committee should select the teams that performed the best over the course of the season, in terms of wins and losses, against the schedule they faced. This doesn't mean that the best four teams would necessarily be playoff selections, but rather that the four most accomplished teams would have that privilege.
My approach to calculating "accomplishment" is from the perspective of an elite team, not from a "typical top-25 team" like the ESPN strength of records do, or from the perspective of an "average" team, like many other schedule strength systems do. I think elite team perspective matters most when determining the four most accomplished teams in the nation at season's end. My calculation is relatively simple -- produce the average number of losses an elite team would be expected to accrue against a given schedule (EL, the strength of schedule rating produced in my FEI ratings output each week), and subtract the actual number of losses that team accrued. The result is the "Elite Win Margin" (EWM), how many wins above or below elite expectations were achieved by the given team.
|Rk||Team||EWM||CFP Rk||FBS Rec||EL||Rk|
This output has some similarity to the current CFP committee rankings. The most notable exception is how the committee is treating the non-Power 5 teams. Roughly speaking, and using this Elite Win Margin output as a benchmark, the committee is essentially saddling the Group of Five conference teams with an extra loss. Central Florida, undefeated currently, would be ranked in the EWM ratings at 11th if they had one loss instead of zero, which is precisely where the CFP committee has the Knights ranked. Other Group of Five teams would also fall further down the EWM ratings -- not necessarily in sync with the committee, but closer to how the committee has positioned them.
I'm going to continue to produce these Elite Win Margin ratings down the stretch, and also re-run these ratings over the last few years to see if similar patterns and similarities between the two rating sets emerge. As always, questions and feedback on this ratings approach and its applications is always appreciated.
FEI Week 11 Ratings
The Fremeau Efficiency Index (FEI) is a college football rating system based on opponent-adjusted possession efficiency. Adjusted Possession Advantage (APA) ratings represent the per-possession scoring advantage a team would be expected to have on a neutral field against an average opponent, calculated as a function of current FEI overall, offense, defense, and special teams ratings.
Strength of Schedule ratings (PSOS) represent the average number of losses an elite team (two standard deviations better than average) would have against the team's regular season schedule to date. Offensive FEI (OFEI) is scoring value generated per drive adjusted for starting field position and opponent defenses faced. Defensive FEI (DFEI) is scoring value generated per opponent drive adjusted for starting field position and opponent offenses faced. Special Teams FEI (SFEI) is scoring value generated per possession by a team's non-offensive and non-defensive units adjusted for opponent special teams units faced. The team's record to date against opponents ranked in the FEI top 10 (v10), top 20 (v20), top 30 (v30), top 40 (v40), and top 50 (v50) are also provided.
Ratings and supporting data are calculated from the results of non-garbage possessions in FBS vs. FBS games.