MLB Grade — 2026-08-17

How the 2026-08-17 predictions scored against actual results

Cumulative Δlog-loss vs always-pick-home: -0.0178 ± 0.0180 per game over 36 games (paired; negative = beating the baseline - currently ahead).

YesterdayCorrectLog-loss
model8/110.592
always-pick-home8/110.664

Daily detail: 73% hit rate · Brier 0.200 · avg margin error 2.4 · avg total-runs error 4.0. Running: 17/36 (47.2%) · cumulative log-loss 0.693 · cumulative Brier 0.250

GamePickPredictedActual
✅Diamondbacks @ Red SoxRed Sox (57%)4.7–4.111–1 (home first)
✅White Sox @ CubsCubs (61%)5.2–4.27–5 (home first)
❌Cardinals @ RedsReds (50%)4.5–4.41–2 (home first)
✅Dodgers @ RockiesDodgers (64%)5.6–3.85–11 (home first)
✅Athletics @ RoyalsRoyals (61%)5.2–4.29–5 (home first)
❌Braves @ TwinsBraves (53%)4.8–4.24–2 (home first)
✅Padres @ MetsMets (53%)4.5–4.42–1 (home first)
✅Marlins @ PhilliesPhillies (61%)4.9–3.96–5 (home first)
✅Tigers @ PiratesTigers (50%)4.7–4.45–8 (home first)
✅Orioles @ RaysRays (58%)4.7–4.17–6 (home first)
❌Cardinals @ RedsCardinals (51%)4.5–4.26–5 (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