How the ratings work

The Index is predictive, not a resume. It answers one question: how many points better than an average FBS team would this team be on a neutral field today? That number is turned into the 0–99 scale shown everywhere, the same scale used for players, coaches and programs.

One scale, anchored to history. Every rating measures how far a season stood above its own contemporaries — a team against that year's teams, a running back against that year's regular running backs — and places that on a single all-time scale, so different eras compare fairly:

A season doesn't automatically produce a 99. Early in a season, ratings start from what we expected of a player or team and move as the games add up, so a hot first month reads as “very good so far,” not “best ever.” Hover any rating for its tier.

It comes from the model, not from a poll. The model rates every team from every play, adjusted for opponents, starting each season from a preseason projection built on returning and incoming production, recruiting, coaching changes and the quarterback room. Early in the season the projection carries more weight; game results take over as they accumulate. Games against FCS opponents count less.

No betting lines go in. The public model never uses market lines, win totals or polls as inputs, so when it disagrees with the market, that disagreement is real. Every prediction will be logged publicly with a timestamp.

Resume and Gap. Resume is what a team has earned: wins compared with what a top-25 team would expect against the same schedule. Gap is the Index minus the Resume: blue means a team is better than its record shows, red means the record is ahead of how good it is.

Margin of error and “Sure?” Ratings early in the season rest on fewer games, so the margin of error is wider. Three dots means the model is very confident in a rank.