MLB Predictions

MLB Engine Overview

Baseball simulation is uniquely suited to Monte Carlo methods because of the sport's discrete, sequential structure. Each at-bat follows a count-state Markov chain from first pitch through outcome. The engine models pitcher fatigue across innings, park-specific run environment adjustments, and platoon advantages (lefty/righty splits) for every batter-pitcher matchup.

Bullpen modeling tracks available relievers, pitch counts from recent outings, and closer usage patterns. This is critical for total projections and late-game moneyline value.