Ranked picks with source labels.

Lambert's fallback data (1.0 IP, 9.00 ERA) provides no real signal; likely an emergency starter or recent callup. Model is high on CLE (62.9%) vs m...

Rodriguez's 5.0 IP over 10 starts is absurdly low (~0.5 IP/start); stat line is untrustworthy. Kay at 4.2 IP over 9 starts is similarly broken. Wit...

Model has CIN at 58.6% vs market's 45.4%—a 13.2pp gap triggering the disagreement flag. CIN's 16-8 record and 5.5 RPG offense support the model's c...

Model favors BAL at 58.6% but market has KCR at 54.5% — a 4pp gap. Bassitt has no resolved stats (null ERA/WHIP/K9), triggering structural uncertai...

Model has PHI at 58.6% but market implies CHC 60.9%. Both pitchers are disasters: Boyd 14.73 ERA in 3.2 IP, Backhus 20.25 ERA in 1.1 IP. CHC offens...

Market SDP 59.5% vs model 58.6% - close call. Buehler at Coors is concerning but SDP offense (4.4 RPG, .547 OPS) vs COL (3.4 RPG, .531 OPS) tilts a...

Model picks BOS at 57.5% vs market 43.6%—a 14pp gap flagged correctly. Fried's 0.00 ERA in 6.1 IP is obviously incomplete data; BOS offense at 3.4...

Model favors MIN at 57% vs market's 41%, a 16-point gap. Away pitcher has fallback data (no ERA/K9/WHIP), home pitcher Holmes lacks stats despite 1...

Model has WSN at 58% vs market's 43%, a massive 15-point gap. ATL has superior record (16-8 vs 11-13) and better RPG (5.8 vs 5.5). Model picking th...

Ohtani data is fallback (only 6.0 IP logged), making this a thin-information spot. Market has LAD at 66.5% implied, model at 55.2%—an 11.3pp gap. S...

Model has SEA 53.6%, market 62.8%—a 9pp gap flags structural miss. T-Mobile PF 0.855 is extreme pitcher park; model total 9.0 vs market 7.5 ignores...

Model has LAA at 53.4%, market implies 58.3% — lean with the market. Soriano's 0.00 ERA in 6 IP is extreme small sample. Lauer's 15.19 K/9 in 5.1 I...