ABS Changed Everything: How the Automated Strike Zone Fixed Umpire Bias by Count
Every baseball fan has seen it: a pitcher throws a nasty breaking ball on 0-2, the pitch paints the corner, and the umpire rings it up as ball one. Or a 3-0 fastball drifts just off the plate and gets called a strike anyway. It always felt like the count influenced the call. Now, for the first time, we have data that proves it — and shows exactly what changed when MLB introduced the Automated Ball-Strike system in 2026.
The Methodology
Using pitch-by-pitch Statcast data from 2024, 2025, and 2026, we identified every pitch that resulted in either a called strike or a called ball. We then classified each pitch as in-zone or out-of-zone using continuous plate coordinates: a pitch is in the strike zone if |plate_x| ≤ 0.83 feet and sz_bot ≤ plate_z ≤ sz_top (the sz values are batter-specific, set by Statcast for each at-bat).
From there, we computed two error rates for each ball-strike count:
- Missed Strike % — in-zone pitches incorrectly called ball (favors the batter)
- Phantom Strike % — out-of-zone pitches incorrectly called strike (favors the pitcher)
The full dataset covers roughly 350,000 called pitches per full season in 2024 and 2025, and approximately 45,000 so far in 2026.
The Two-Strike Effect: Umpires Were Protecting Batters
The most striking pattern in the pre-ABS era is what happens at two-strike counts. When a batter is one strike away from a strikeout, umpires were dramatically more likely to call an in-zone pitch a ball — effectively giving the batter a second chance.
2025 (Human Ump)
2026 (ABS)
2025 (Human Ump)
2026 (ABS)
At 0-2 in 2025, nearly 1 in 4 pitches that crossed the strike zone was called a ball. In 2026, that dropped to roughly 1 in 9. The effect holds at every two-strike count:
| Count | 2024 | 2025 | 2026 (ABS) | Change ('25→'26) |
|---|---|---|---|---|
| 0-2 | 22.1% | 24.2% | 11.8% | ▼ 12.4 pts |
| 1-2 | 18.8% | 19.8% | 8.3% | ▼ 11.5 pts |
| 2-2 | 15.5% | 17.7% | 7.8% | ▼ 9.9 pts |
| 3-2 | 11.7% | 12.8% | 6.5% | ▼ 6.3 pts |
The pattern is consistent: the further into a two-strike hole the batter was, the more the umpire compensated — and the larger the ABS correction. The 0-2 count, where batters faced the most danger, saw the biggest swing.
The Three-Ball Effect: A Mixed Picture
The flip side of the two-strike bias should be a three-ball bias — umpires calling borderline pitches strikes to avoid walking batters on close calls. The data partially confirms this, though the signal is noisier.
At 3-1 and 3-2, phantom strike rates are modestly lower in 2026 than in either prior season, consistent with ABS reducing pitcher-favorable errors. The 3-0 count is the outlier — phantom strikes appear slightly elevated in 2026, but with fewer than 1,000 called pitches at 3-0 so far this season versus roughly 6,200 in a full season, this reading is almost certainly noise.
| Count | 2024 | 2025 | 2026 (ABS) | Change ('25→'26) |
|---|---|---|---|---|
| 3-0 | 20.4% | 17.3% | 22.6% | ▲ 5.3 pts* |
| 3-1 | 13.2% | 10.3% | 11.1% | ▲ 0.8 pts |
| 3-2 | 7.4% | 6.2% | 5.4% | ▼ 0.8 pts |
* 2026 3-0 sample: ~950 called pitches. Small sample — treat with caution.
The Bigger Picture: Count-Based Bias Was Systematic
The human umpire data from 2024 and 2025 tells a coherent story: umpires were unconsciously (or consciously) adjusting their strike zones based on the count. The further a count favored the pitcher — 0-2, 1-2, 2-2 — the more umpires expanded the effective zone in the batter's favor. The further a count favored the batter — 3-0, 3-1 — the more umpires helped the pitcher.
This isn't a new observation. Researchers have documented count-based umpire bias for years. What's new is the magnitude of the correction in 2026. ABS didn't just reduce errors at the margins — it cut the two-strike missed-strike rate nearly in half across every count.
Conclusion
The 2026 ABS data confirms what many suspected: human umpires weren't just missing pitches randomly. Their errors were systematic, count-dependent, and consistently tilted toward keeping at-bats alive. At two-strike counts, they missed in-zone pitches at nearly twice the rate they do now under the automated system. At full counts, phantom strikes were common.
ABS hasn't eliminated all error — the 3-0 phantom strike rate remains elevated, and even ABS has calibration variance — but the count-based bias that shaped at-bat outcomes for decades has been dramatically reduced. Baseball's strike zone in 2026 is, for the first time, genuinely consistent from pitch one to pitch three.
Data source: MLB Statcast pitch-by-pitch data. 2024–2025: full seasons. 2026: through approximately April 15, 2026. Strike zone defined as |plate_x| ≤ 0.83 ft, sz_bot ≤ plate_z ≤ sz_top (batter-specific).