2026 Win-Loss Predictions

Predicted records for every FBS team, reported as full probability distributions rather than single guaranteed outcomes. Each game is given a win-percentage chance based on roster strength, coaching, home field, and the gauntlet penalty for tough multi-week stretches: a team favored 90% is nearly automatic; a team favored 52% is essentially a coin flip. Add those per-game chances together to get how the season is most likely to play out.

The key thing to remember: a team favored in 12 close games rarely wins all 12. Stringing wins together compounds, so even the strongest teams usually drop a game or two they were favored to win. Teams are ranked by expected wins, with strength of schedule breaking ties: when two teams project for the same record, the one that earned it against the tougher schedule ranks higher. RPI (a blend of expected win percentage, opponent talent, and opponents' opponent talent) is shown alongside as an additional measure.

How to read each column:

• SOS: opponent talent on a 0 to 1 scale. A 0.80 means the team faces opponents in the top fifth of FBS rosters. This is the tiebreaker between teams with equal expected wins.
• RPI: a blended rating of win percentage and schedule strength. Higher is better. Shown as a secondary measure.
• Stat W-L: average wins if you replayed the season many times. Captures fractional outcomes (a team favored in eight games by ~70% will average around 5.6 wins from those games).
• W-L: the single most likely actual record (the peak of the bell curve). This is usually the most realistic single answer for "what will this team finish at."
• P(undefeated): chance the team runs the table.
• P(1 loss): chance the team finishes with exactly one loss.

The probability conversion was tuned against 1,625 FBS games from the 2024 and 2025 seasons; the win-percentages here are calibrated to match how often favorites in real games actually win.