How the 2026-08-20 predictions scored against actual results
Cumulative Δlog-loss vs always-pick-home: -0.0196 ± 0.0137 per game over 75 games (paired; negative = beating the baseline - currently ahead).
| Yesterday | Correct | Log-loss |
|---|---|---|
| model | 5/9 | 0.678 |
| always-pick-home | 4/9 | 0.703 |
Daily detail: 56% hit rate · Brier 0.242 · avg margin error 4.2 · avg total-runs error 5.2. Running: 40/75 (53.3%) · cumulative log-loss 0.680 · cumulative Brier 0.244
| Game | Pick | Predicted | Actual | |
|---|---|---|---|---|
| ✅ | Yankees @ Orioles | Yankees (52%) | 4.4–3.9 | 1–6 (home first) |
| ❌ | Braves @ White Sox | White Sox (51%) | 4.6–4.5 | 0–2 (home first) |
| ❌ | Cardinals @ Reds | Reds (51%) | 4.4–4.3 | 9–10 (home first) |
| ✅ | Giants @ Guardians | Guardians (58%) | 4.5–4 | 5–2 (home first) |
| ❌ | Angels @ Astros | Astros (63%) | 4.8–3.6 | 3–18 (home first) |
| ✅ | Athletics @ Royals | Royals (57%) | 5–4.6 | 6–2 (home first) |
| ✅ | Mariners @ Brewers | Brewers (62%) | 4.9–3.9 | 7–4 (home first) |
| ❌ | Blue Jays @ Rays | Rays (57%) | 4.6–4.2 | 1–5 (home first) |
| ✅ | Nationals @ Rangers | Rangers (55%) | 4.7–4.4 | 2–0 (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