How MMX-F rates college football

What MMX-F is

MMX-F is MarchMetrix's opponent-adjusted team rating. It estimates how a team would perform against an average FBS opponent on a neutral field, in points. Zero is average. A team at +21 would be expected to beat an average opponent by three touchdowns, and a team at -7 would be expected to lose to one by a touchdown. We update the rating as results arrive, then use it to power matchup probabilities, schedule grades, and the playoff simulations.

What goes into it

Three things: who a team played, where the game was played, and what happened on the scoreboard.

Every finished game between FBS teams counts. All of the ratings are worked out together, so a 10-point win over a strong opponent moves a team's rating more than a 10-point win over a weak one. Home field is worth a few points, and that value comes from the results themselves rather than from an assumption. Margins count with diminishing returns: a comfortable win and a blowout look similar to the model, so late scores in a decided game change very little.

The rating ignores polls, preseason reputation, recruiting rankings, and television narratives. The committee model, described in its own article, is the one that has to care about reputation, because the committee does.

The first few weeks

With only a handful of games played, the rating starts from where each team finished the previous season and leans on that starting point less with every game. By the middle of the season, current results carry the rating on their own. Until then, the "projection" label on our pages means exactly that: a projection, not a verdict.

How well it works

The model will miss games, because college football is noisy. The useful question is whether it performs consistently on games it has not seen.

To answer that, we fit the model on the 2025 season through the middle of November and then predicted the margin of every game that followed. On those unseen games, the average miss was about 12.7 points. That number needs context. A single college football game can swing by two touchdowns on a few bounces, and no rating system predicts that. A large share of the 12.7 points is the noise in the sport, and the rest is where we keep working. What we watch is whether the miss stays consistent on games the model never saw.

From a rating to a win probability

The gap between two ratings, adjusted for home field, becomes a win probability using the model's own error distribution. The larger the gap, the more often the favorite wins across many replays of the same matchup. A touchdown favorite is a clear favorite and nowhere near a certain one. The model never says a team will win. It says how often it would.

Quadrants: the schedule, graded

Basketball fans settle resume arguments with quadrant records. Football never had that vocabulary, so we defined it. Every game lands in a quadrant based on where it is played and how good the opponent is right now, using the opponent's current MMX-F rank among the 138 FBS teams.

Every FCS opponent is Q4. Conference championship games use the neutral band. The cutoffs are proportionally tighter than basketball's because a 12-game season makes Q1 chances rare, and a Q1 win in football should mean something.

Quadrants re-rate every week using current ranks. A September win over a team that collapses slides from Q1 to Q2, and a win over a sleeper that turns out to be good gets promoted. The movement is intentional. It is how the schedule grade stays honest as the season reveals who was actually good.

Strength of schedule and strength of record

Strength of schedule (SOS) is the average MMX-F rating of a team's opponents. It measures how hard the road is.

Strength of record (SOR) measures how well a team has traveled it. It compares the team's actual wins with the wins a top-10-caliber team would be expected to collect against the exact same schedule, at the same sites. A positive SOR means the team has done more with its slate than a top-10 team typically would. SOR needs played games, so it appears once the season is under way.

The playoff odds

Every night the model plays the rest of the season 10,000 times with a fixed random seed, so the same inputs always produce the same odds. Each replay resolves every remaining game with the win probability above, fills the conference title games from the simulated standings (with simplified tiebreakers, labeled as such), plays those games on a neutral field, ranks the finished season, and selects the field under the 2026 rules: the ACC, Big 12, Big Ten, and SEC champions get in regardless of ranking, the highest-ranked team from the other six conferences gets in, Notre Dame gets in with a finish of 12th or better, and the remaining spots go to the highest-ranked teams left. Seeds follow rank order, and seeds 1 through 4 get byes.

Until the committee model activates in the early weeks of the season, the simulation ranks teams by MMX-F. Once it is live, each simulated season is ranked the way the committee model expects the committee to rank it.

Probabilities are shown as whole percents. Anything between zero and one percent shows as "under 1 percent" rather than "0%", because a scenario that came up a few times in 10,000 replays is unlikely, not impossible.

#	Team	Playoff	Bye	Avg wins	Likely seed
1	IndianaBig Ten	93%	67%	11.1	1
2	Notre DameFBS Independents	93%	44%	10.4	3
3	Ohio StateBig Ten	90%	64%	9.9	1
4	Texas TechBig 12	90%	11%	11.6	6
5	MiamiACC	83%	38%	10.9	4
6	OregonBig Ten	70%	39%	9.7	1
7	UtahBig 12	65%	43%	10.2	2
8	GeorgiaSEC	62%	37%	9.2	2
9	Ole MissSEC	41%	19%	8.7	3
10	Texas A&MSEC	40%	13%	9.0	4
11	VanderbiltSEC	40%	3%	8.9	7
12	South FloridaAmerican	39%	0%	10.1	12

What it does not do

The rating only knows what has reached the scoreboard. Injuries, weather, a quarterback change, or a coaching change show up only through results. It does not predict single games with certainty, and it is not a betting tool. When the model changes, the version number changes, and this page is updated with a new date.