A team's GordStats rating is how many points better than an average FBS team it would be on a neutral field: a +14 team would be a two-touchdown favorite over an average one.
How It Works
Using every FBS and FCS game since 2014, the model finds the one number per team that best fits all the final margins at once, plus a home-field edge (about 3 points lately). Because every team is rated together, beating a good team counts more than beating a bad one, and an FCS opponent counts as itself, not as a generic one.
Recent games count most: a game counts half as much after 150 days. A pull toward its conference's average keeps a team with few games from swinging wildly (ridge regression).
It's then blended with how well a team plays snap to snap, adjusted for opponents: expected points added per play (EPA), success rate, explosiveness and more (what these mean).
Alongside ours: the committee's CFP rankings (from November), the AP poll, ESPN's FPI (its own points-better-than-average rating), SP+ and Elo.
Keep in Mind
Injuries, transfers and coaching changes count only once they show up on the field. Early in the season it leans on last year. Bowl and playoff games are left out.
When It Updates
Four times a day, around 5:30 and 11:30 AM and PM Eastern.
For the Curious
Scores come from ESPN, the efficiency stats from CollegeFootballData. Tested on every 2020-25 FBS game, each predicted only from earlier games, the efficiency blend cut the typical miss (root-mean-square error, RMSE) from 16.18 to 16.07 points, rating FCS teams individually with the conference pull took it to 15.88, a stronger pull with a shorter memory to 15.84, and a September nudge toward teams from the stronger conferences to 15.82. Blowouts count in full (no cap on margin), and garbage time stays in: taking it out tested worse.