Counterfactual Optimization of Baseball Pitch Sequences and Estimation of Its Impact on Season-Level Statistics
arXiv:2606.17345v1 Announce Type: cross Abstract: Although pitch sequencing is a central topic in baseball analytics, previous studies have primarily focused on optimizing the final pitch within a single plate appearance, leaving the role of preceding setup pitches and their impact on long-term season-level performance insufficiently examined. To address these issues, this study conducted counterfactual analyses using MLB Statcast data. A Transformer-based machine-learning model was trained to predict whether a target pitch would result in an in-play outcome or swing-out. Counterfactual pitch
The increasing availability of detailed sports data (e.g., MLB Statcast) and advancements in AI models, specifically Transformers, enable more sophisticated analyses of complex sports strategies such as pitch sequencing.
This study demonstrates how advanced AI can optimize nuanced strategic decisions in complex, high-stakes environments, extending beyond sports into other domains requiring sequential decision-making under uncertainty.
The ability to counterfactually optimize multi-pitch sequences and estimate long-term impact on season-level statistics provides teams with a significantly more powerful tool for player development, game planning, and tactical adjustments.
- · MLB teams with advanced analytics departments
- · Sports analytics firms
- · AI/ML researchers in sports applications
- · Teams solely relying on traditional scouting
- · Players unable to adapt to optimized pitching strategies
Baseball teams will adopt or enhance AI-driven models to optimize pitch sequencing for improved on-field performance.
This optimization will lead to subtle but measurable shifts in player performance evaluation, training methodologies, and potentially player valuations.
The success in baseball could inspire similar AI-driven counterfactual optimization in other sports or analogous complex, sequential decision-making scenarios, further permeating AI into strategic planning.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI