This study develops a Monte Carlo simulation framework for estimating baseball offensive production using empirically derived event probabilities from play-by-play data. The model simulates complete innings and games to predict run distributions, player contributions, and lineup efficiencies. The proposed framework incorporates situational baserunning probabilities stratified by sprint speed categories, stolen-base attempts, double-play tendencies, and pitch-level outcomes, providing a comprehensive stochastic representation of offensive events. This approach uniquely integrates sprint-speed-conditioned baserunning transitions and lineup-level Monte Carlo optimization, achieving state-level fidelity not captured in prior Markov or aggregate simulators. Applications to multiple Major League Baseball (MLB) teams demonstrate the model's interpretability and potential for lineup optimization. Validation against multi-season empirical data confirms strong alignment between simulated and observed run distributions.
© Copyright 2026 Journal of Sports Analytics. IOS Press. All rights reserved.