Sat, Aug 30, 12:00 AM
| 1 | 2 | 3 | 4 | T | |
|---|---|---|---|---|---|
| GT | 0 | 13 | 7 | 7 | 27 |
| COLO | 7 | 3 | 3 | 7 | 20 |
Our model blends Elo ratings (COLO Elo 1634, GT Elo 1536) plus home-field advantage. That projects COLO -6.3 (68% to win) — 9.8 points of value on COLO versus the market line of +3.5.
GT up 7 entering the 4th quarter. Across 1,299 historically comparable game states (within ±2 pts and ±3 min, from 3,056 games):
Georgia Tech 27, Colorado 20.
No — the model picked COLO, which didn't hit. We report the misses too.
Why trust the number?Gridpex's model is walk-forward validated on 2014–24, out-of-sample: +0.28 pts mean Closing Line Value and 53.5% beat the close. We report the record — wins and losses — never fake locks.
I had COLO pregame and it didn't hit. We report the misses — that was one.