About Worster-Underwood
WU is an ensemble college football ranking system that combines two fundamentally different models — one based on how teams perform, one based on what teams have accomplished. The tension between them is the point.
Underwood
Power ratingA weighted least-squares regression on margin of victory. Every game is a data point; the model finds the set of team ratings that best explains all observed margins simultaneously.
Two adjustments are applied: a home-field advantage correction (typically 2–4 points), and a recency weight so that recent games matter more than early-season results.
Underwood is forward-looking. A dominant blowout tells it more than a squeaker. A team playing great football with a so-so schedule will still show up highly here.
Worster
Résumé rankingA pure win/loss résumé ranking. Teams are sorted first by wins, then recursively by the quality of opponents beaten and lost to — at increasing depth.
Worster is backward-looking. Margin of victory is completely ignored — a 1-point win and a 40-point blowout are identical. A team that has beaten everyone in front of them ranks highly, full stop.
WU Rating
The WU Rating is a simple average of the Underwood and Worster adjusted ratings, both of which are scaled to a [−30, +30] range across FBS teams each week. A team at +30 is the top of that model; 0 is median; −30 is the bottom.
The Leancolumn in the rankings table shows which model rates a team higher and by how much — positive means Underwood favors them, negative means Worster does. Teams where the models strongly disagree are worth a closer look: they're either performing far better than their résumé suggests, or they've beaten everyone in front of them without looking particularly impressive doing it.
Game data is sourced from the College Football Data API (CFBD). Rankings are computed daily during the season. Only FBS teams are included in the ratings, though FCS opponents are accounted for where applicable.