How the 2026-08-29 predictions scored against actual results
Cumulative Δlog-loss vs always-pick-home: -0.0230 ± 0.0088 per game over 199 games (paired; negative = beating the baseline - currently ahead).
| Yesterday | Correct | Log-loss |
|---|---|---|
| model | 12/17 | 0.630 |
| always-pick-home | 10/17 | 0.683 |
Daily detail: 71% hit rate · Brier 0.219 · avg margin error 3.3 · avg total-runs error 2.9. Running: 120/199 (60.3%) · cumulative log-loss 0.667 · cumulative Brier 0.237
| Game | Pick | Predicted | Actual | |
|---|---|---|---|---|
| ✅ | Orioles @ Athletics | Orioles (59%) | 5.3–4.1 | 3–5 (home first) |
| ✅ | Rockies @ Braves | Braves (67%) | 5.3–3.8 | 2–1 (home first) |
| ✅ | Reds @ Cubs | Cubs (62%) | 5.2–4.1 | 17–5 (home first) |
| ❌ | Royals @ Guardians | Guardians (59%) | 4.7–3.9 | 3–8 (home first) |
| ❌ | Dodgers @ Tigers | Dodgers (55%) | 4.7–3.8 | 2–1 (home first) |
| ✅ | Phillies @ Angels | Phillies (62%) | 5.1–3.5 | 2–4 (home first) |
| ✅ | Rangers @ Brewers | Brewers (64%) | 5–3.6 | 5–3 (home first) |
| ✅ | White Sox @ Twins | White Sox (51%) | 4.8–4.5 | 2–3 (home first) |
| ✅ | Astros @ Mets | Mets (55%) | 4.5–4.3 | 6–2 (home first) |
| ❌ | Red Sox @ Yankees | Yankees (52%) | 4.2–4.1 | 0–6 (home first) |
| ✅ | Diamondbacks @ Giants | Diamondbacks (52%) | 4.7–4.2 | 1–7 (home first) |
| ❌ | Pirates @ Cardinals | Cardinals (54%) | 4.5–4.3 | 2–6 (home first) |
| ✅ | Padres @ Rays | Rays (54%) | 4.4–4.2 | 7–6 (home first) |
| ✅ | Mariners @ Blue Jays | Blue Jays (54%) | 4.5–4.3 | 4–3 (home first) |
| ✅ | Marlins @ Nationals | Nationals (52%) | 4.6–4.5 | 5–4 (home first) |
| ✅ | Red Sox @ Yankees | Yankees (54%) | 4.2–4.1 | 9–2 (home first) |
| ❌ | Diamondbacks @ Giants | Diamondbacks (51%) | 4.7–4.3 | 7–2 (home first) |
The headline is the paired per-game log-loss difference against a fixed always-pick-home forecast (p=0.534) on the same games, ± one standard error; daily hit rate is noisy and demoted on purpose. Reference points: 0.693 is a coin flip; the model's walk-forward backtest runs 0.680 (research/SP-BACKTEST.md).
Betting-blind Elo model (K=3, +24 home, MOV-weighted, 60% season carryover) · Can Tre Beat Vegas · whosyurgoat.app/vegas/mlb