CFB Betting Edge

2025 · Week 1 · Big 12 · backtest — what the model predicted vs what happened, and vs the market
Week 1 backtest · Big 12 · 9 games — what the model predicted vs what happened
Straight up
6–367%
ATS vs close
3–633%
Model margin MAE
15.1
Market margin MAE
13.8
→ market was closer (model can't beat the close ATS)
Results — model read vs the final score + how the closing-line lean held up
Model
pred Kansas 15.2 · actual Kansas 24
winner Kansas
ATS vs close
leaned Kansas -14.0Cover ✓
margin err: model 8.8 · mkt 10.0 · closer
Hawai'i @ ArizonaFinal 640
Model
pred Arizona 9.6 · actual Arizona 34
winner Arizona
ATS vs close
leaned Hawai'i +15.5Miss ✗
margin err: model 24.4 · mkt 18.5 · mkt closer
Auburn @ BaylorFinal 3824
Model
pred Baylor 5.6 · actual Auburn 14
winner Baylor
ATS vs close
leaned Baylor +1.0Miss ✗
margin err: model 19.6 · mkt 13.0 · mkt closer
Model
pred UCF 5.1 · actual UCF 7
winner UCF
ATS vs close
leaned Jacksonville State +20.5Cover ✓
margin err: model 1.9 · mkt 13.5 · closer
Model
pred Kansas State 3.4 · actual Iowa State 3
winner Kansas State
ATS vs close
leaned Kansas State -3.0Miss ✗
margin err: model 6.4 · mkt 6.0 · mkt closer
Model
pred Colorado 2.9 · actual Georgia Tech 7
winner Colorado
ATS vs close
leaned Colorado +3.5Miss ✗
margin err: model 9.9 · mkt 3.5 · mkt closer
Utah @ UCLAFinal 4310
Model
pred Utah 0.5 · actual Utah 33
winner Utah
ATS vs close
leaned UCLA +6.5Miss ✗
margin err: model 32.5 · mkt 26.5 · mkt closer
Model
pred Nebraska 1.1 · actual Nebraska 3
winner Nebraska
ATS vs close
leaned Cincinnati +6.0Cover ✓
margin err: model 1.9 · mkt 3.0 · closer
Model
pred TCU 3.2 · actual TCU 34
winner TCU
ATS vs close
leaned North Carolina +3.5Miss ✗
margin err: model 30.8 · mkt 30.5 · mkt closer
The model beats SP+ but not the closing line (ATS ≈ break-even). “Model−Mkt” is the gap between our projected home margin and the market's implied margin — a read on disagreement, not a profitable edge. Upcoming games have no closing line yet, so the market number marked ~ is the consensus midpoint of the best posted spreads across books.