Ranked picks with source labels.

Away starter is TBD and market data is missing. Joe Ryan's 5.1 IP across 10 starts averages 0.5 IP per start which is clearly corrupted data. MIL's...

Abbott's 0.0 ERA across 60 IP (10 starts at 6 IP each) is transparently impossible—likely a data error. Without reliable away pitcher stats, moneyl...

LAD strong favorite at both model (60.2%) and market (69.3% implied). LAA offense anemic at 2.6 RPG vs LAD 4.2 RPG. Kochanowicz 11.25 ERA in 4.0 IP...

Madden's MLB debut (1 GS, no seasonal stats) makes DET side unreadable. TOR offense (.786 OPS vs .582) is clearly superior but model only gives the...

Model has NYY at 56% vs market 60%, reasonable convergence. Model's 9.04 total vs market 7.0 is a 2-run gap—flag fires correctly. Citi Field slight...

Model total 9.72 vs market 11.5 is a 1.8-run gap at Coors (PF 1.47). Market knows this park inflates; model underweights it. Kelly's missing stats...

Model has MIA 52.2%, market has TBR 53.8% — very tight. TBR 28-14 at home vs MIA 20-24, .802 OPS vs .774. Pitcher stats null for both sides limits...

Market disagrees strongly: Pinnacle has BAL -132 (57% implied), model has WSN 53.5%. Model is wrong-weighting away_pitcher_last_short_era despite B...

Both pitchers show 6.0 IP and 0.00 ERA over 10 starts—physically impossible. Data quality flags 'full' but these are clearly placeholder/corrupted...

Model picks BOS at 53% but market implies ATL 57.5%. ATL is 30-14 vs BOS 17-26; model offense_defense_mismatch flag confirms it backs the worse tea...

Dustin May's 13.50 ERA over 4.0 IP is extreme small-sample noise; Wacha's 0.00 ERA over 6.0 IP equally so. Both starters have ~10 total IP, making...

Market favors PIT at 55.8% vs model's PHI slight comparison at 52.2%. Ashcraft missing all season stats (ERA/IP/WHIP null) creates data uncertainty...

Model picks SFG 52% but market has OAK -127 (55%). OAK's .512 OPS vs SFG's .388 OPS is a 124-point gap; model underweights this recent form. Mahle...