Imagine you run a restaurant and replace your sommelier with a wine-dispensing robot that pours exactly 150ml every time. The old sommelier gave regulars a generous pour and shorted tourists — but most glasses were already within a few milliliters of the target. You've killed favoritism, but total wine served barely changes. That's what happened when Korea's KBO league swapped human umpires for the Automated Ball-Strike System (ABS) in 2024 — the first top-tier professional baseball league to do so. The committed claim: automating the strike zone eliminates human discretion (count-dependent calls, catcher framing value, star-batter advantages), but because the machine's new zone geometry differs from the old human zone, the net effect on plate-appearance outcomes is precisely estimated and close to zero. The "rents" destroyed — the value captured by skilled framers, favored batters, and pitchers who exploited umpire tendencies — amount to only a few runs per player-season. The paper decomposes the automation into two separable channels: removal of human bias and imposition of a different geometric zone that scales with batter height. The ladder here is unusual because there's essentially no prior empirical benchmark to beat. Minor-league robot-ump experiments (MiLB's ABS Challenge in the Atlantic League and other affiliates) generated descriptive reports, but nobody had a million-pitch before/after dataset from a top-tier league spanning 2021–2026. Cho and Lee own this field by default — which makes their honest admission that effects are "close to zero" more credible than if they were trying to outperform a rival estimate. Methodologically, this is difference-in-differences econometrics applied to pitch-level microdata. The architecture is classical causal inference: they exploit the sharp 2024 ABS adoption as a treatment, separate the discretion-removal channel from the geometry channel using pitch-location data, and estimate effects on called strikes, plate-appearance outcomes (walks, strikeouts, hit-by-pitch), and pitcher/batter behavioral responses. The key identification assumption is that pitch quality and player composition didn't shift discontinuously at the ABS adoption date for reasons unrelated to the system. Integrity is solid for an observational study. They use the full population of KBO pitches — over one million — not a sample. They report precisely estimated near-zero effects rather than hunting for significance. They explicitly decompose channels rather than lumping everything into a single treatment effect. The main vulnerability is the standard one for any single-league adoption study: there's no control league running in parallel, so any KBO-wide trend coinciding with 2024 could confound. The authors acknowledge this and use within-season variation to mitigate it. The most interesting behavioral finding: pitchers adjusted quickly, throwing higher to exploit the machine's consistent upper-zone geometry, with left-handers accelerating an existing trend. Batters showed no detectable adjustment. This asymmetry — supply-side re-optimization with demand-side inertia — is a pattern that generalizes well beyond baseball to any market where one side faces a rule change and the other doesn't. The paper's deeper contribution is conceptual: when you replace a human decision-maker with a rule-based system, who benefits depends on the geometry of the new rule and on how many decisions are actually exposed to the change. In the KBO case, most pitches were already called correctly by humans, so the "treatment dose" is small. The lesson for other automation domains — automated hiring screens, algorithmic sentencing, robo-advising — is that the incidence of replacing human judgment depends on how different the machine's boundary actually is from the human's, and on how much traffic sits near that boundary.
