ABS Took the High Strike — and That’s Roughly 40–50% of the Walk Spike. Pitchers Own the Rest.

Through Wednesday’s games, MLB’s walk rate is 9.77% — the highest since 1950. The Associated Press ran the story that morning. Players are blaming the new ABS-era strike zone. ESPN ran a follow-up. Diamondbacks closer Paul Sewald said pitchers are getting squeezed. Cubs infielder Nico Hoerner said hitters are laying off pitches at the top of the zone. Yankees center fielder Cody Bellinger said the sample is too small to know yet.

Two weeks ago, we wrote a piece arguing the walk spike was pitchers, not umpires. The called-strike rate on borderline pitches near the zone edge had actually risen 4 points year over year — our conclusion was that pitchers were nibbling more, and ABS wasn’t squeezing the zone. With 17 more days of data and a more rigorous analytical pass, we’re here to update that piece. We were partly right and partly wrong. The honest answer:

The 2026 ABS-era zone really did move — the top edge shrank, the bottom expanded. Roughly 40–50% of the +0.82 percentage-point walk spike traces to that zone change. The remaining 50–60% is pitcher behavior, pitch-mix, and adaptations our analysis can’t capture (more on that below). We were wrong to dismiss the zone effect at zero.

The walk spike is real, and it’s not the cold weather

Walks always run hot in March and April — pitchers are still ramping up, batters work the count, north-city games are cold. The ESPN article notes the caveat in passing. We tested it directly. Across 2018–2025, the league’s walk rate over the same Mar 27–Apr 22 window has been remarkably stable: a mean of 9.02% with a standard deviation of just 0.17 percentage points. The range across all eight prior years is 8.69% to 9.17%.

2026, at 9.77%, sits 4.4 standard deviations above that historical mean — far outside what any prior cold April could explain. It’s 0.60 percentage points above the prior 8-year April maximum (the 2018 reading of 9.17%). Whatever is driving this, it’s not seasonality.

Walk rate (incl. intentional walks) over the Mar 27–Apr 22 window, 2018–2026. The 2018–2025 mean is 9.02% with SD 0.17pp. 2026 sits 4.4σ above that. 2020 is absent: no MLB games during COVID.

The strike zone moved — up

Hoerner was the most precise of the players quoted. He said hitters are laying off pitches at the top of the zone. The data agrees with the geometry, even if not exactly with his behavioral framing.

Comparing the matched Mar 27–Apr 14 windows, called-strike rate by plate location shifted in two directions at once:

• The top edge (around z = 3.2–3.9 ft) lost roughly 7–8 percentage points of called-strike rate, with single cells dropping more than 14 points. This is the biggest contiguous region of change.
• The bottom edge (around z = 1.0–2.0 ft) expanded by roughly 5–6 percentage points, on a smaller area. Where umpires used to call ankle pitches as balls, the new ABS rule book gave them a precise floor (27% of standing height, 1.62 ft for a 6-footer) and they’re calling the bottom more confidently.

The new rule definition matches the geometry exactly. ABS’s vertical zone is 27% to 53.5% of standing height. For a 6-foot batter, the top is at 3.21 ft — about an inch below the typical 3.29 ft top umpires were calling in 2025. The bottom is at 1.62 ft — right where umpires already called it in 2025. The top got a hard ceiling. The bottom got a hard floor. Both showed up in the calls.

Called-strike rate change, 2026 minus 2025, from a 2D bin map (0.1 ft cells) of taken pitches in absolute plate coordinates. Blue = lower 2026 rate (zone shrank). Red = higher 2026 rate (zone expanded). Dashed box: ABS rule zone for a 6-ft batter (x ±0.71 ft, z 1.62–3.21 ft). Outlined cells: changes of at least 3 percentage points that are statistically significant at 95% confidence. Hover a cell for details.

To make sure we weren’t fooled by quirks in how Statcast labels pitch height, we ran the analysis a second way using a different machine-learning approach and a different vertical-coordinate system (relative to each batter’s size instead of absolute feet). That second pass initially disagreed with the first. The disagreement turned out to be a real and important data caveat — explained in the methodology section — and once resolved using absolute plate coordinates, both approaches agreed: top down, bottom up.

The walk damage isn’t at 3-2

Our prior CalledThird work has shown that 3-2 is the most leveraged count in the ABS challenge system — just 2.8% of called pitches but 24% of all challenge run value. The natural assumption when walks rise is that the worst damage concentrates there. It doesn’t.

Walk-rate change by count, 2026 minus 2025 (rate of PAs that pass through each count and end in a walk):

• All walks (the league rate): +0.82pp
• 1-0: +1.27pp
• 3-1: +1.22pp
• 3-0: -1.42pp (yes, fewer walks once you’re ahead 3-0; pitchers throw more pitches there now)
• 3-2: -0.11pp (could be slightly negative or slightly positive — within margin of error)

Statistically, no count stands out from the average pattern (p = 0.67 across all twelve). The walk increase is broadly distributed. Pitchers get to 3-2 about as often as before; once they’re there, the walk rate is essentially flat. The leverage concentration we documented in the challenge system doesn’t map to a leverage concentration in the underlying walks.

Eventual walk rate among PAs that passed through each count, 2025 vs 2026 same Mar 27–Apr 14 window. 3-2 in clay highlights the flat YoY delta (−0.11pp, p = 0.93). Heterogeneity across counts: Cochran’s Q p = 0.67 — no per-count concentration.

About 40–50% of the spike is the zone

The biggest question, and the one our Apr 9 piece got wrong: how much of the walk spike is the zone change actually responsible for?

To answer it, we built a model from 2025 data that predicts how an umpire would call any given pitch (strike or ball, based on its location). Then we applied that 2025 model to every 2026 plate appearance, replaying each one pitch by pitch under the old zone, and compared the resulting walk rate to what actually happened. (Swings, fouls, balls in play, and ABS challenges stayed exactly as they actually played out; only the called-strike-vs-ball outcomes were resampled.) We did this twice independently — once with the zone model from one of our analysts, once with a fresh from-scratch implementation as a tiebreaker.

Actual 2025 walk rate, actual 2026 walk rate, and two independent counterfactual replays of 2026 plate appearances under the 2025 zone (averaging 32 random draws per PA). Both replays land below the 2026 actual rate — meaning the new zone is calling more strikes net, which would otherwise have been balls. The gap between each counterfactual bar and the actual 2026 bar is the "zone-only" share of the walk spike: ~40–50%. The remaining 50–60% is unmodeled pitcher behavior. We report the range across implementations rather than a single point estimate.

Both implementations agree on direction and converge on a range. We report the range, not a point estimate, because two careful builds differ by about ten percentage points and we haven’t formally decomposed where the residual variance comes from. Roughly half the walk spike is the zone change. The other half is something else. The most likely candidate is pitcher behavior — throwing more pitches off the plate to compensate for the new geometry — but our counterfactual holds pitcher behavior fixed by design, so we can’t prove that part directly.

One result we owe the reader as an open question: when we restrict the analysis to just 0-0 called pitches, the sign flips. The 2026 zone is actually more strike-friendly on first pitches than the 2025 zone was — but less strike-friendly on late-count pitches at the top edge. The overall effect across all pitches is positive (zone change adds walks), but the first-pitch flip is genuine and we don’t have a clean mechanistic story for it yet. That’s the biggest skeptic gap in this analysis. Round 2 will break the model down by count to chase it down.

Updating the Apr 9 piece

This is the part of the article that’s harder to write. Two weeks ago, on 10 days of data, we published a piece concluding the walk spike was driven by pitchers nibbling, not umpires squeezing. The shadow-zone called-strike rate had risen, not fallen, year over year. We wrote: “The ABS zone squeeze hypothesis is rejected.”

That conclusion was too strong. With 17 more days of data and a substantially harder methodological pass, the picture is:

• The pitcher-nibbling story is still real. Roughly 50–60% of the +0.82pp walk spike traces to behavior our zone-only model can’t capture — where pitchers are throwing the ball, what pitch types they’re mixing in, how they’re sequencing. First-pitch called-strike rate did fall (by 1.77 percentage points), and PAs reaching 3-0 rose by more than the entire walk-rate increase. Pitchers are throwing first pitches further off the plate, and that ripples through to later counts.
• But we under-weighted the zone effect at zero. The zone change is a real and substantial contributor — not the dominant cause, but on the order of 40–50% of the walk spike. The shadow-zone aggregate we used in the Apr 9 analysis flattened a directional story (top down, bottom up) into a single number that read as “net up.” Looking at the 2D geometry, the top edge clearly moved.

Hoerner’s top-of-the-zone observation is the closest match to what the data show, even though we’re measuring the call (mechanical) rather than the swing decision (behavioral) he was describing. Sewald and McCann’s “smaller zone” framing is partly right on the net effect on walks but misses that the bottom edge actually expanded. Bellinger’s sample-size caveat is well taken: 27 days is a short window, and we’ll re-run this analysis with mid-May data before declaring victory.

What it means for tonight’s game

If you’re watching a game and a pitcher misses a high strike that looked good, you’re probably right and the umpire is probably calling the new rule. If you’re watching a low pitch get called a strike that would have been a ball last year, that’s the bottom-edge expansion. The walk that comes from a tight 2-2 fastball at the letters is half-rule-change, half-pitcher-being-careful. The 3-2 walk that ends an at-bat is, statistically, no more or less likely than it was last year — even though every other outcome around it is shifting.

And if you’re trying to read the league-wide pattern, the answer is closer to a yes-and than a yes-or-no. Players are right that the zone changed. They’re right that pitchers are getting squeezed at the top. They’re wrong that the zone is smaller everywhere — the bottom edge actually got bigger. They’re right that walks are up; they’re partly right about why. The cleanest single-number version: about half the walk spike is the new zone, about half is pitcher behavior. Half and half.