About
What this is, and how it works
A lineal championship for college football: one title, passed hand to hand on the field since the first game ever played, and a site that computes every step of it.
The idea
Boxing has lineal champions: to be the champ, you beat the champ. College football never had one, so this site made one. The belt starts with Rutgers, who won the first college football game ever played, 6–4 over Princeton on November 6, 1869. From that day on, whoever beats the holder takes it. That is the entire rule. No polls, no committees, no rankings — 158 years and 328 reigns later, Notre Dame holds it.
Because the rule is so simple, the belt goes places the polls never would. It spends years in the Ivy League, wanders through programs that no longer field teams, disappears into the middle of a conference for a decade, and then a single September upset sends it somewhere new. The full history is all 1,634 games it was ever on the line for.
How it’s computed
Nothing on this site is researched by hand. A script walks every college football game on record, in date order, from the College Football Data API, applying the ruleset to each one: if the holder is playing and loses, the belt moves. Ties stay with the holder. A bye, a canceled season or a bowl opt-out just means the belt waits. Every reign, record, map and story on the site is regenerated from that walk several times a week during the season, usually within a few hours of the holder’s game going final.
That also means the site can be wrong exactly one way: if the underlying game record is wrong. When CFBD corrects a score, the next build corrects the lineage. If you spot something that looks off, email hello@collegefootballbelt.com with the game and we’ll trace it.
What’s written by software
Three kinds of text on the site are generated rather than written by a person, and each is labeled where it appears: the next-game preview and its prediction (“the lean”); the short recaps on game pages that have a full box score; and the brief historical notes on older games, which are built only from the matchup, date, score and what it meant for the belt. They are written by Claude, an AI model, strictly from the facts shown on the same page — a plain-English reading of the numbers, not reporting, and never an invented play, player or crowd. The prediction is a for-fun editorial call, never betting advice, and every pick is graded against the real result on the lean’s ledger so you can see how it’s doing.
The stories are a fourth thing: their structure is written by a person, but every name, date and number in them is filled in live from the lineage, so they can’t drift out of date as new reigns happen.
Who runs it
collegefootballbelt.com is an independent fan project. It first went online in 2018 and was rebuilt from the ground up in 2026 as the automated site you’re reading now. It has no affiliation with any school, conference, the NCAA or the College Football Data project. The site carries a small number of ads to cover hosting and data costs; the data itself is free — see the API if you want to build on it, and the badge if you want the current holder on your own page.
Follow along at @CollegeFBBelt on X and Instagram, or get an email only when the belt changes hands via the feed. Privacy details are here.