Automated ball-strike calls dent the edge of star hitters in Korea: after ABS went live, high-status batters saw declines in on-base and walk-related metrics compared with lesser-known players, whereas pitchers showed no measurable change.

Main Finding

The Korea Baseball Organization’s 2024 adoption of an Automated Ball-Strike System (ABS) reduced the on-field performance of high-status batters relative to low-status batters, while high-status pitchers showed no significant change. Key affected batter outcomes include on-base percentage (OBP), IsoD, walk rate (BB%), strikeout rate (SO%), and the BB/K ratio.

Key Points

• Natural experiment: ABS implementation in the KBO (2024) provides exogenous change in ball-strike adjudication.
• Method: difference-in-differences (DiD) regressions comparing 2023 (pre) vs 2024 (post) seasons.
• Sample: 148 batters and 112 pitchers across the two seasons.
• Main result: high-status batters experienced statistically meaningful declines in OBP, IsoD, BB%, BB/K and increases in SO% relative to low-status batters after ABS adoption.
• High-status pitchers showed no differential effect post-ABS.
• Interpretation: removing human umpire discretion appears to have eliminated some preexisting status-related advantage for batters.

Data & Methods

• Data: player-season statistics from KBO for 2023 (pre-ABS) and 2024 (post-ABS). (Sample sizes reported: n = 148 batters; n = 112 pitchers.)
• Identification strategy: difference-in-differences linear regressions estimating the interaction of post-ABS period with a high-status indicator to capture differential impacts.
• Outcomes analyzed: OBP, IsoD, BB%, SO%, BB/K (among other standard batting/pitching metrics).
• Notes/assumptions: results rely on the DiD parallel trends assumption and on the study’s operational definition of “high-status” (as defined by the authors). Robustness checks, heterogeneity analyses, and specific status definitions were not detailed here.

Implications for AI Economics

• Status rents and automation: Automated adjudication (an AI-enabled system) can materially reduce status-based advantages that previously arose from subjective human judgment. Where subjective evaluation confers rents to high-status workers, automation may compress performance gaps and measured returns to status.
• Distributional consequences: Technologies that standardize or objectify evaluation can harm incumbents who benefited from favorable discretionary treatment while potentially benefiting lower-status workers through fairer treatment — with implications for wages, contract renewals, and career trajectories.
• Complementarity and adaptation: The absence of immediate effect for pitchers suggests heterogeneous adaptation costs and complementarities by role. Economics of automation should consider role-specific responses and whether affected workers can adapt skills/strategies to the new objective standard.
• Policy and organizational design: Firms and leagues should anticipate unintended effects when deploying automated evaluation tools — e.g., contract renegotiation, incentive redesign, and communication strategies — and monitor for behavioral responses that could offset intended fairness gains.
• Further research directions: quantify long-run earnings and contract impacts, verify mechanisms with pitch-level call data (e.g., changes in strike-zone calls), test parallel-trend validity, and explore generalizability to other labor markets where subjective evaluations are being automated.

Limitations to keep in mind: one-league, short-run (one-year) post period; potential omitted variable or adaptation dynamics; exact status definition and granular mechanisms were not supplied in this brief.