CFB · Player Props

College Football player prop projections

Projects college football props with the sport's own realities simulated: early starter pulls, backup snaps, and wide differences in team tempo. Live on real lines for the season.

A free account opens the board and the graded record. Choose College Football in the sport selector.

Today's free sample

Every game on the 2026-11-15 CFB slate has started, so its props are closed. The free sample returns when the next slate's board is built.

College Football markets and the graded record

The College Football standard-line record over the last 30 days is 52.1% across 3,910 decided lines. 12 of 14 markets have a graded record in that window.

  • Passing Yards
    44.3% to 55.7% at 95% · n=294
    50.0%
  • Passing TDs
    49.9% to 75.2% at 95% · n=52
    63.5%
  • Interceptions
    28.5% to 57.8% at 95% · n=40
    42.5%
  • Completions
    44.4% to 60.3% at 95% · n=147
    52.4%
  • Pass Attempts
    41.4% to 57.3% at 95% · n=148
    49.3%
  • Rushing Yards
    51.3% to 58.5% at 95% · n=736
    54.9%
  • Rush Attempts
    48.3% to 60.0% at 95% · n=275
    54.2%
  • Receiving Yards
    46.9% to 52.3% at 95% · n=1,362
    49.6%
  • Receptions
    50.0% to 59.5% at 95% · n=416
    54.8%
  • Rush + Rec Yards
    52.1% to 65.6% at 95% · n=200
    59.0%
  • Pass + Rush Yards
    39.3% to 54.2% at 95% · n=169
    46.7%
  • Rush + Rec Touchdowns
    graded nightly
    —
  • Field Goals Made
    graded nightly
    —
  • Kicking Points
    38.0% to 60.7% at 95% · n=71
    49.3%

Rates are the standard-line record over the last 30 days: every decided standard line, recommended or not, wins and losses. The recommended cohort is reported as its own block in the Model Room and is not this number. Goblin and demon lines are over-only products and are not inside these numbers. Intervals are Wilson at the raw sample size, and because rows share slates and games they are a floor on the uncertainty.

What the College Football engine models

Plays out each game it projects drive by drive, thousands of times, with the college game's own rules and realities

  • Simulates each game drive by drive with the college game's realities built in: starters pulled in blowouts, backups taking the late snaps, and overtime under college rules
  • Accounts for the wide spread in team tempo and for lopsided matchups
  • Treats roster churn as uncertainty: transfers, freshmen and committee backfields lower confidence in the player's role and can lower the grade
  • Eligible settled projections are checked against results, and the record shows the selected window and the engine version where one was recorded
Documented engine version

Serving engine version ncaaf-sim-0.2

What to know before you read it

Real DFS lines priced against measured college data: player usage and efficiency from last season's play, opponent adjusted team ratings, returning production, and transfer moves. Rosters are current, player availability is not accounted for, and every player builds healthy, stated on the board. No sportsbook consensus enters a grade or an edge, so edges read model versus line, and every grade is capped at B+ until the backtest earns more. When measured on 2 September 2026, graded volume was thin (1,371 settled rows) and the probabilities ran overconfident at the extremes. A holdout evaluation on 14 October 2026 tested a probability correction market by market, and seven markets were then cleared to carry it while three stay uncorrected. Live slates have not yet confirmed the extremes, so treat the probabilities as directional until the calibration curve tightens.

No sportsbook consensus enters a college football grade, edge or Market %. Edges read model versus line. Where sportsbooks quote the same prop, their prices may be shown beside the line for comparison only, and they never set an edge or a grade.

  • Whether a starter is pulled early is the model's own estimate, and it is published as one
  • Unproven model, so every grade is capped at B+ until the backtest earns more
  • Player availability is not accounted for, so every player is projected as healthy and a suspension or injury does not appear on the board. The board says so
  • Anytime touchdown props publish as labelled no plays until graded weeks show how far the touchdown projections run high or low
  • Fantasy weights follow the college football charts Sleeper and Betr publish. PrizePicks publishes none we could read, so its weights and Underdog's follow their NFL conventions. The engine does not yet cover every published scoring item (two point conversions, fumble recovery touchdowns) and projects fumbles lost as zero, so fantasy legs never settle in grading
  • On 2 September 2026, across 1,371 graded rows, probabilities ran overconfident at the extremes: props the model called 94% landed 55% of the time and props it called 5% landed 56% of the time. A holdout evaluation on 14 October 2026 tested a probability correction market by market. It narrowed the gap between forecast and hit rate for seven markets, which are now cleared to carry it, and failed for completions, pass attempts and kicking points, which stay uncorrected. That evaluation did not re-measure the extreme bands, so whether the extremes are fixed on live slates is not yet confirmed. The near certain block declines the top band outright, which is why almost every posted line here publishes as no play
  • The strength of schedule adjustment depends on the data tier: when opponent adjusted team ratings are unavailable the board falls back to unadjusted season stats, demotes every grade one tier, and flags each signal

Read the full College Football board

A free account opens the Signal Board, the Prop Grid and the Model Room. Research and analysis only, never a bet placed for you.

The Slate

One weekly note: what published, what the graded record shows, what changed. No picks, and nothing is sent until you confirm your address.