Ranked picks with source labels.

Model (64.1%) and market (64.1%) perfectly aligned on NYY moneyline. Missing pitcher stats means total confidence is low despite park factor 0.93 s...

Model has WSN at 62.9% but market implies only 44.2% — an 18.7pp gap flags structural misread. Christian Scott's null stats suggest data_quality='f...

Model likes ARI at 60.2% vs market's 56.4%, but both pitchers show tiny IP samples (Gallen 4.0 IP, Ray 5.1 IP) making season ERA/WHIP unreliable. G...

Model has LAD at 58.6%, market at 56.6% — essentially aligned. King's 0.00 ERA in 5.0 IP is microscopic sample noise, but Yamamoto's 3.00 ERA over...

Model picks DET at 57.5% but CLE is the superior team (26-22 vs 20-27, 5.6 RPG vs 2.3 RPG, .600 OPS vs .582 OPS). Cecconi's 12.46 ERA in 4.1 IP per...

Pitcher stats are catastrophically broken: Imanaga shows 7.2 ERA in 5.0 IP (10 starts impossible), Sproat 21.0 ERA in 3.0 IP. Data quality marked '...

Model 57.5% vs market 60.3% on SEA is modest disagreement. Schultz missing all rate stats is a data gap. T-Mobile 0.855 park factor plus extreme_pa...

Model picks BAL at 55% but market has TBR -149 (59% implied). TBR is 30-15 vs BAL 21-26; BAL offense is anemic at 3.4 RPG/.662 OPS. Rogers' 0.00 ER...