A fair-value estimate of the chance a hitter homers in a given game. Pick a date and a hitter — the model finds their game, pulls live data, and prices the prop as American odds. You type in the book's price to see if there's an edge. Nothing is bet automatically.
It estimates expected home runs λ from the inputs below, then
P(≥1 HR) = 1 − e−λ. Inputs folded into every price:
Blend of career, season and recent HR/PA. Recent is noisy — keep its weight low.
Lineup slot sets expected PA (defaults are manual — confirmed lineups post ~2–4h pregame).
Odds-ratio: batter rate scaled by each arm's HR/PA allowed vs league. Platoon auto-fills from splits.
Handedness HR park factor (1.00 neutral). Auto-set from the home venue & hitter's side.
~1% HR per °F off 70°F; wind out/in to center adds or removes carry. Closed roof neutralizes both.
Manual nudges for travel fatigue, back-to-backs, and injuries. Rested & healthy = 1.00.
Baseline per-PA rate bent by each factor, then compounded across expected trips into λ. Bars center at 1.00.