College Football Conference Power Ratings
After an offseason of historic realignment, we examine the state of the new college football conference landscape.
Go back to 2010 and imagine me telling you that Dallas-based SMU is in the ACC, the Pac-12 is gone, and that USC would be traveling to Piscataway, NJ to face Rutgers for a conference game…in the Big Ten! You’d likely have me swiftly institutionalized. Well, folks, that is the world we are living in now. This season, conference standings will look quite a bit different, and with the different makeups of conferences we may need to re-evaluate the conference hierarchy that we have become accustomed to. Last week, I published my pre-season college football Elo ratings, and I thought a fun follow up to that would be to examine those ratings from a conference-level perspective. With all the shuffling around of teams, it is worthwhile to take a step back and assess how these additions and subtractions of programs have effected the conferences themselves.
2024 Pre-Season Conference Power Ratings
Using the average team Elo from my pre-season projections, I ranked each conference. Juxtaposed is the average team Elo from the end of the season last year via CollegeFootballData.com, prior to all the realignment action.
Unsurprisingly, the SEC and Big Ten rank the highest, with the other two “Power Four” conferences fairly close behind. What was surprising, though, was that even with all the additions to the Big Ten and SEC, their average Elo rating declined by 15 and 40 points (or 0.8% and 2.4%), respectively. In the Big Ten, this was a result of 100+ point Elo declines in Ohio State, Oregon, Washington, and Michigan. This makes some sense, as Ohio State lost Heisman finalist WR Marvin Harrison Jr. to the NFL and QB Kyle McCord to the transfer portal. Oregon sent eight players including star QB Bo Nix - also a Heisman finalist - to the NFL. Michigan and Washington bore the brunt of the losses, however, as each lost their coach - Michigan’s Harbaugh to the NFL’s Los Angeles Chargers and Washington’s DeBoer to Alabama - along with a slew of players to both the NFL and transfer portal. The SEC’s drop is a bit more puzzling, but in all likelihood is due to some regression to the mean that is inherent in nearly all in pre-season projection systems, including mine, which pulls down some of the more dominant of last years teams like Alabama, Texas, and Georgia - which actually had the highest end-of-season Elo per CollegeFootballData.com, even though they were not selected for the College Football Playoff. The former of the three, shall I remind you, will enter the season without legendary coach Nick Saban (the greatest football coach of all time, in my opinion) for the first time since 2006. Regardless of all this, these ranking shook out largely how I expected.
Talent Variability
Not all teams in the SEC are Georgia and not all teams in the Big Ten are Oregon. With most conferences fielding between 14 and 18 teams, there will surely be a few bottom-feeders. The SEC has Vandy, ACC has Stanford, the Big Ten has Rutgers. All of which are distinguished institutions and a degree from any is quite valuable, but unfortunately for them their football fans prefer wins over A’s. By taking the standard deviation of team Elo in each conference, we can assess which conferences have the widest ranges in the quality of their member programs.
Admittedly, I probably should have excluded the five independent teams from this, both from a statistical and practical standpoint, but for the sake of consistency in charts I opted to keep them in. The widest ranges and greatest standard deviations again went to the SEC and Big Ten. Being that they also produce 11 of the top 13 teams per projected Elo, we can consider them the most top-heavy conferences from a talent perspective. The variability in the Sun Belt is not surprising. The conference has traditionally been a landing spot for teams transitioning from the FCS/D1-AA level, including James Madison, Appalachian State, Georgia State, Georgia Southern, and Coastal Carolina just since 2013. It has also seen surprise breakout teams emerge in various seasons, like Appalachian State in 2019 or James Madison - which heard its name pop up in College Football Playoff discussions just this past season.
Overperforming and Underperforming Conferences
By comparing the expected share of Elo a conference would have, based its number of teams, against its actual share of the total Elo across college football, we can identify which conferences are truly the stronger or weaker ones.
Since Elo is a closed system - any point of Elo lost by a team is gained by another team - we can calculate an expected Elo total for each conference by taking the total Elo across college football and dividing it by the number of teams in the conference. We can then calculate the percent of total college football Elo a conference would be expected to represent by taking this expected conference Elo total and dividing by the overall college football total. By comparing the percent of total Elo a conference actually has (the sum of the total team Elo in the conference divided by the number of teams in the conference) against what we would expect based upon its number of members, we can see which conferences are in total better or worse than others. For those with a mathematical or probability background, this is essentially a quasi-expected frequency calculation. The results fell largely in line with the average conference Elo from the first table. However, this gives the added insight into which conferences are better or worse than average, which ones generally consist of above or below average teams, and the magnitude in which certain conferences are better/worse than others1.
While these results were fairly unsurprising, it did generate some interesting observations in some conferences and give some broad, albeit very early, insight into the effects of the realignment of this past offseason. I am very interested to see how conferences shake out once the dust settles from this massive shakeup in the sport.
Conferences with a positive “Difference from Expected” value represent above-average conferences by total team talent level, while negative values correspond to below-average conferences.