Rankings
Automatically calculated from completed games. COMPARE Ranking remains the independent default ordering.
Tap a heading to sort: descending → ascending → COMPARE order. The official COMPARE rank stays visible.
| COMPARE | Team | Conf | W | L | Win % | Move | Résumé | Performance | Talent | SOV | SOR | PD |
|---|
Games & Predictions
Pregame predictions use prior-week data. Score forecasts are historically calibrated; rankings, picks, and win probabilities are unchanged.
Matchup Lab
Pick any two FBS teams and use COMPARE's most recent completed-week data. These are hypothetical matchups, not scheduled games.
Any Two Teams
Using latest available snapshot…
Favorability Rankings
Every FBS team against every other FBS team on a neutral field, using the original COMPARE Game Predictor. A team's score is the number of opponents it would be favored to beat. Teams favored against the same number of opponents share a rank. Early-week teams with limited FBS data are marked *. This is separate from the official COMPARE Ranking.
Season Projections
Simulation-based projected conference standings and playoff after weekly finals, with the single-outcome COMPARE forecast still available for What-If.
Projected Conference Standings
Projected Conference Championship Games
Projected Final COMPARE Rankings
A representative simulated finish after remaining games. These are hypothetical COMPARE rankings for one coherent simulated season, not a change to live COMPARE rankings; choose single-outcome to see the original forecast. CFP seed and COMPARE rank are separate.
COMPARE — Projected 12-Team CFP
Comparison
Select a What-If game to see how it changes projected playoff seeds and conference champions.
Projection notes
Season Simulation · Team Probabilities
Uses the existing COMPARE predictions and winner probabilities, then recomputes COMPARE's final ranks, conference championships and playoff field in each simulated season. These are COMPARE-based playoff odds, NOT a modeled CFP committee. No additional CFBD requests.
The site automatically calculates a standard 1,000-run projection after a completed week is confirmed in the scheduled score sync. Choose a different depth for additional analysis or a What-If scenario. Probabilities vary with draws; live COMPARE rankings are unaffected.
What-If Mode · Remaining Games
Flip a projected winner. Changes stay on this page and never alter the site's rankings or predictions. What-If scenario scores continue to use the original model, because hypothetical score margins feed the projected rankings.
Upset Watch
True upset candidates use COMPARE rank versus projected spread. Close games, model disagreements, and talent differences are separate information.
Upset picks: lower-ranked team favored by COMPARE's game predictor. High alert: ranking gap 20+ with projected spread ≤3.5. Potential: gap 20+ with spread ≤7.5. Talent alone never triggers an upset.
Weekly Results
Scoring performance relative to COMPARE’s pregame forecast, split into offense, defense, and overall. Previously completed weeks use clearly labeled retrospective reconstructions where available; actual saved pregame forecasts take priority.
Team scoring performance
Compared with the forecast before the game: offense = points scored minus projected; defense = projected points allowed minus actual points allowed; overall = offense + defense (margin surprise). Green is better than forecast on all three.
Biggest upsets
Most accurate predictions
All evaluated games
Team
One place for résumé, performance, talent, movement and schedule.
Team
COMPARE Ranking by week
Season schedule
Historical Benchmark
COMPARE versus BCS and CFP releases. Useful as a diagnostic, not a target.
Official releases
| Official week | Model through | Spearman | Top 4 | Top 10 | Top 25 | Official #1 | Model #1 |
|---|
Final top tier
Model & System
The spreadsheet mathematics, separated into distinct jobs and run automatically.
Ranking layers
Automation
Score forecasts
Displayed game scores and spreads receive a post-prediction calibration fitted on 2014–2021 seasons and checked against 2022–2025. Original model scores remain available in game details. The original winner pick, win probabilities, résumé, rankings, and What-If projected rankings remain unchanged. Upset Watch now groups games using the displayed calibrated spread and the original COMPARE ranking.
Rules that stay fixed
FCS results count in W/L but provide no positive strength credit. SOV is pure defeated-opponent quality. SOR uses the verified 2:1 adjusted OWP/OOWP construction. Rank is WP → wins → losses → SOV → SOR → point differential. FullRank keeps shared ties. Talent never awards résumé points and cannot independently reverse the base prediction.