College football betting research, built on the roster.
College football is where Edge Labs is built. The research runs on proprietary recruiting data, the talent a roster actually carries, and it ends in a number we hang on every priced game. Real numbers, the sample attached, the side stated.
Years of recruiting capital mapped to the roster on the field, then priced against the market line. A three-seat analyst panel states the pick, the tier and the vote on every priced game, and every pick grades in public at the number we posted. One free CFB pick on Tuesday, Thursday and Saturday, stamped when it drops. Members see every one.
From the desk
Football insights, recent posts, and researchReturning almost everyone can still leave you near the bottom. Akron entered our 2026 preseason board with returning production of 0.918 and a -7.96 rating. Continuity added points; the talent and prior-performance inputs kept the rating low.
Friday's college football board: every pick cashed. 7 for 7 on spreads and totals. Saturday's full board is live for members.
Research articleWhat our model refuses to claim. Our NFL model beats its own predecessor on margin accuracy, loses to the closing spread in every held out season, and we print both numbers instead of picking the one that flatters us.
Research articleThe preseason prior is worth eight games. Every team in our college football rating starts the season at its preseason number, then updates one game at a time on the live ratings board; expected wins on that board still come from the preseason walk. The...
The edge
What the books overlookRecruiting capital
The talent a roster actually carries, mapped across recruiting cycles and the portal. It moves our number on a game before it moves the line.
Talent vs record
The gap between the talent a team signed and what it did on the field. Who is overrated, who is underrated, surfaced from five seasons of data.
Regression signal
Recruiting talent is a leading indicator and records drift back toward the roster. We name the likely risers and fallers, and we bet them before the market catches up.
The rest of the desk
4 in season · 3 preseason reads, frozen at the Aug 13 build · 1 closed testGames
LiveThis week’s board: every FBS game, the market spread and total, our number beside it, and the analyst panel pick, tier and vote. One free CFB pick on Tuesday, Thursday and Saturday mornings and stamped when they drop; members see every one.
Open Games →Talent vs Record
Preseason · locked Aug 13The gap between the talent a team signed and what it did on the field. Who outran the roster, who got robbed, the sample on every team.
Open Talent vs Record →Power Ratings
LiveThe EL Rating v1: every team starts at its preseason composite and moves one game at a time on opponent adjusted margin, refit after each week’s finals. Expected wins beside every team.
Open Power Ratings →Win Total Standings
LiveWhere every team stands against the total we locked before week 1: record so far, the remaining schedule re-walked with the in-season rating, and the gap. Updates every week.
Open Win Total Standings →Regression Watch
Preseason · locked Aug 13Talent is a leading indicator and records drift back toward the roster. This names the likely risers and fallers, and it is one input to the season-long positions.
Open Regression Watch →Portal
Preseason · locked Aug 13Transfer-portal influx and roster churn by program: who is reloading, who is bleeding talent, and which win totals have not caught up yet.
Open Portal →CFB Data
LiveSeason ratings, recent results, pregame ELO, and the excitement index. The base layer the talent work sits on top of.
Open CFB Data →Early-Season Totals
Closed · graded Sep 20Closed and failed. FBS weeks 1-3 totals of 47+ went under 54.75% in the backtest (n=884, 2018 to 2025), then 47.3% in our pre-registered 2026 forward test (n=131, gate 52.38%). The rule, the history and the final count stay public.
Open Early-Season Totals →Records regress toward the roster
Across five seasons (n over 150 per side, 2020 to 2025, point-in-time), recruiting talent acts as a leading indicator. Teams that outran the talent they signed gave back about 15 win-percentage points the next season. The underrated rose about as much. That finding is the spine under the number we publish on every game.
Get the weekly board
The tools are live and the season is underway. A three-seat analyst panel states the pick, the tier and the vote on every priced game. One free CFB pick on Tuesday, Thursday and Saturday, stamped when it drops; members see every pick first.
Our leans and the work behind them, straight to your inbox. No spam. 21+.