2025 College Football Week 3 Model Backtest
Quick answer: In this retrospective 2025
Week 3 FBS-vs-FBS backtest, the frozen model went 32/47 (68.1%) with mean Brier
0.1809. Full 2025: 550/762 (72.2%). These were not live picks.
Ratings use final 2023-2024 FBS-vs-FBS games only. They were frozen before any
2025 result was graded. Historical pregame spreads may be blended, but their probability
mapping was calibrated on 2024 outcomes—not on the 2025 test results.
Week 3 scorecard
| Metric | Value |
|---|---|
| Games | 47 |
| Decided | 47 |
| Correct picks | 32/47 (68.1%) |
| Mean Brier | 0.1809 |
| Full-season accuracy | 550/762 (72.2%) |
Full-season method comparison
The headline is the blended ensemble. This table separates the ratings model from
the historical spread so market information is not misreported as pure model performance.
| Method | Games | Correct | Accuracy | Mean Brier |
|---|---|---|---|---|
| Elo model only | 762 | 495 | 65.0% | 0.221 |
| Market spread only | 762 | 564 | 74.0% | 0.1733 |
| 40/60 ensemble | 762 | 550 | 72.2% | 0.1804 |
Every Week 3 FBS-vs-FBS game: pick vs actual
| Date | Away | Home | Home WP | Pick | Actual | Result | Source |
|---|---|---|---|---|---|---|---|
| 2025-09-11 | NC State 34 | Wake Forest 24 | 31.8% | NC State | NC State | HIT | ensemble |
| 2025-09-12 | Colorado 20 | Houston 36 | 44.9% | Colorado | Houston | MISS | ensemble |
| 2025-09-13 | Kansas State 17 | Arizona 23 | 39.7% | Kansas State | Arizona | MISS | ensemble |
| 2025-09-13 | New Mexico 35 | UCLA 10 | 79.9% | UCLA | New Mexico | MISS | ensemble |
| 2025-09-13 | Wisconsin 14 | Alabama 38 | 88.8% | Alabama | Alabama | HIT | ensemble |
| 2025-09-13 | Clemson 21 | Georgia Tech 24 | 40.0% | Clemson | Georgia Tech | MISS | ensemble |
| 2025-09-13 | Buffalo 31 | Kent State 28 | 9.8% | Buffalo | Buffalo | HIT | ensemble |
| 2025-09-13 | Central Michigan 3 | Michigan 63 | 94.9% | Michigan | Michigan | HIT | ensemble |
| 2025-09-13 | Oregon 34 | Northwestern 14 | 7.4% | Oregon | Oregon | HIT | ensemble |
| 2025-09-13 | Oklahoma 42 | Temple 3 | 10.3% | Oklahoma | Oklahoma | HIT | ensemble |
| 2025-09-13 | Memphis 28 | Troy 7 | 37.1% | Memphis | Memphis | HIT | ensemble |
| 2025-09-13 | South Alabama 15 | Auburn 31 | 78.2% | Auburn | Auburn | HIT | ensemble |
| 2025-09-13 | Louisiana 10 | Missouri 52 | 84.0% | Missouri | Missouri | HIT | ensemble |
| 2025-09-13 | UConn 41 | Delaware 44 | 33.6% | UConn | Delaware | MISS | ensemble |
| 2025-09-13 | SMU 28 | Missouri State 10 | 7.2% | SMU | SMU | HIT | ensemble |
| 2025-09-13 | Washington State 10 | North Texas 59 | 56.3% | North Texas | North Texas | HIT | ensemble |
| 2025-09-13 | USC 33 | Purdue 17 | 14.0% | USC | USC | HIT | ensemble |
| 2025-09-13 | Georgia 44 | Tennessee 41 | 39.5% | Georgia | Georgia | HIT | ensemble |
| 2025-09-13 | Oregon State 14 | Texas Tech 45 | 84.7% | Texas Tech | Texas Tech | HIT | ensemble |
| 2025-09-13 | Pittsburgh 24 | West Virginia 31 | 46.3% | Pittsburgh | West Virginia | MISS | ensemble |
| 2025-09-13 | Iowa State 24 | Arkansas State 16 | 19.2% | Iowa State | Iowa State | HIT | ensemble |
| 2025-09-13 | UTEP 10 | Texas 27 | 98.0% | Texas | Texas | HIT | ensemble |
| 2025-09-13 | South Florida 12 | Miami 49 | 87.1% | Miami | Miami | HIT | ensemble |
| 2025-09-13 | Liberty 13 | Bowling Green 23 | 33.6% | Liberty | Bowling Green | MISS | ensemble |
| 2025-09-13 | Middle Tennessee 14 | Nevada 13 | 66.4% | Nevada | Middle Tennessee | MISS | ensemble |
| 2025-09-13 | Florida Atlantic 28 | Florida International 38 | 61.8% | Florida International | Florida International | HIT | ensemble |
| 2025-09-13 | Jacksonville State 34 | Georgia Southern 41 | 52.9% | Georgia Southern | Georgia Southern | HIT | ensemble |
| 2025-09-13 | Western Michigan 0 | Illinois 38 | 89.7% | Illinois | Illinois | HIT | ensemble |
| 2025-09-13 | Ohio 9 | Ohio State 37 | 88.9% | Ohio State | Ohio State | HIT | ensemble |
| 2025-09-13 | Arkansas 35 | Ole Miss 41 | 72.1% | Ole Miss | Ole Miss | HIT | ensemble |
| 2025-09-13 | App State 22 | Southern Miss 38 | 33.5% | App State | Southern Miss | MISS | ensemble |
| 2025-09-13 | Old Dominion 45 | Virginia Tech 26 | 66.1% | Virginia Tech | Old Dominion | MISS | ensemble |
| 2025-09-13 | East Carolina 38 | Coastal Carolina 0 | 44.4% | East Carolina | East Carolina | HIT | ensemble |
| 2025-09-13 | Massachusetts 7 | Iowa 47 | 97.4% | Iowa | Iowa | HIT | ensemble |
| 2025-09-13 | Eastern Michigan 23 | Kentucky 48 | 83.8% | Kentucky | Kentucky | HIT | ensemble |
| 2025-09-13 | Florida 10 | LSU 20 | 64.7% | LSU | LSU | HIT | ensemble |
| 2025-09-13 | New Mexico State 14 | Louisiana Tech 49 | 69.0% | Louisiana Tech | Louisiana Tech | HIT | ensemble |
| 2025-09-13 | Texas A&M 41 | Notre Dame 40 | 75.3% | Notre Dame | Texas A&M | MISS | ensemble |
| 2025-09-13 | Vanderbilt 31 | South Carolina 7 | 69.4% | South Carolina | Vanderbilt | MISS | ensemble |
| 2025-09-14 | Duke 27 | Tulane 34 | 51.6% | Tulane | Tulane | HIT | ensemble |
| 2025-09-14 | Navy 42 | Tulsa 23 | 14.8% | Navy | Navy | HIT | ensemble |
| 2025-09-14 | Utah 31 | Wyoming 6 | 21.6% | Utah | Utah | HIT | ensemble |
| 2025-09-14 | Akron 28 | UAB 31 | 71.9% | UAB | UAB | HIT | ensemble |
| 2025-09-14 | Air Force 30 | Utah State 49 | 44.7% | Air Force | Utah State | MISS | ensemble |
| 2025-09-14 | Texas State 15 | Arizona State 34 | 83.5% | Arizona State | Arizona State | HIT | ensemble |
| 2025-09-14 | Minnesota 14 | California 27 | 47.7% | Minnesota | California | MISS | ensemble |
| 2025-09-14 | Boston College 20 | Stanford 30 | 22.7% | Boston College | Stanford | MISS | ensemble |
Highest-confidence misses
Limited to the 25 most confident misses so the table
stays readable; the complete game-level record is preserved in the JSON artifact.
| Game | Pick | Confidence | Actual | Margin |
|---|---|---|---|---|
| New Mexico at UCLA | UCLA | 79.9% | New Mexico | 25 |
| Boston College at Stanford | Boston College | 77.3% | Stanford | 10 |
| Texas A&M at Notre Dame | Notre Dame | 75.3% | Texas A&M | 1 |
| Vanderbilt at South Carolina | South Carolina | 69.4% | Vanderbilt | 24 |
| App State at Southern Miss | App State | 66.5% | Southern Miss | 16 |
| UConn at Delaware | UConn | 66.4% | Delaware | 3 |
| Middle Tennessee at Nevada | Nevada | 66.4% | Middle Tennessee | 1 |
| Liberty at Bowling Green | Liberty | 66.4% | Bowling Green | 10 |
| Old Dominion at Virginia Tech | Virginia Tech | 66.1% | Old Dominion | 19 |
| Kansas State at Arizona | Kansas State | 60.3% | Arizona | 6 |
| Clemson at Georgia Tech | Clemson | 60.0% | Georgia Tech | 3 |
| Air Force at Utah State | Air Force | 55.3% | Utah State | 19 |
| Colorado at Houston | Colorado | 55.1% | Houston | 16 |
| Pittsburgh at West Virginia | Pittsburgh | 53.7% | West Virginia | 7 |
| Minnesota at California | Minnesota | 52.3% | California | 13 |
Methodology and interpretation
- Train: 2023 and 2024 final regular-season FBS-vs-FBS games.
- Test: 2025 final regular-season FBS-vs-FBS games, with ratings frozen.
- Model: opponent-adjusted Elo with score-margin adjustment and empirical home field.
- Ensemble rows: 40% Elo plus 60% historical spread-derived win probability.
- Brier measures probability error; lower is better and 0.25 is the coin-flip reference.
FAQ
Did this model exist in the 2025 season?
No. This is a retrospective reconstruction, clearly labeled as such. It does not claim
these picks were published before kickoff.
Why only FBS-vs-FBS?
It gives one comparable
opponent pool and avoids pretending schedules across NCAA divisions are interchangeable.
Is this betting advice?
No. It is a model audit and historical scorecard.
Related: 2026 CFB predictions
and the NFL Facts Stats Hub.
Sources
Retrospective backtest rebuilt 2026-09-06.
No mock data and no 2025 result used to fit the frozen ratings.