What Abstap measures
Abstap tracks every Automated Ball-Strike (ABS) challenge in Major League Baseball: a batter, catcher, or pitcher asks the system to review a ball-strike call, and ABS either overturns the umpire or confirms the call. We record each challenge with its full game context — inning, count, score, baserunners, the pitch location, and the umpire behind the plate — then rank them.
If you're new to the system itself, start with our plain-English guide to ABS challenges. This page is about the numbers.
The Robbery Score
Not every overturned call matters equally. A phantom strike three to end a tie game in the ninth is a bigger deal than a missed call in a blowout. The Robbery Score is how we rank them. It is the product of two independent measures:
Call Magnitude — how badly the call missed
Call magnitude is how far off the call was, in inches, against the 17-inch rulebook zone — sized per batter from the pitch's reported top and bottom. ABS counts a pitch as a strike if any part of the ball touches the zone, so for a pitch called a strike but really a ball we measure the gap to the ball's nearest edge (accounting for its ~2.9-inch diameter) — which is why those numbers track what a broadcast strike-zone graphic shows, not a larger center-of-ball figure.
- A strike overturned to a ball: the ball was outside the zone. Magnitude is the gap from its nearest edge to the zone — how far it was from being a strike.
- A ball overturned to a strike: the ball was inside the zone. Magnitude is how far inside the edge the pitch was — how clearly it should have been a strike.
Either way, a larger number means a more obvious miss. Pitches with no reported location can't be measured and are left unranked.
Leverage Index — how much the moment mattered
Leverage Index (LI) is a standard sabermetric idea: how much a single play can swing a game's outcome. An average situation is about 1.0; a blowout collapses toward 0; a tie game with the bases loaded in the ninth climbs to 5 or higher. We compute LI from the inning, half-inning, outs, baserunners, and score differential at the moment of the pitch.
Our LI is a parametric approximation anchored to well-known game states, not the original (unpublished) Tango leverage table. It is built for ranking — sorting the day's challenges by how much was at stake — and it does that reliably. It is not intended as a precise win-probability figure.
Stakes weighting — why October scores differently
The base formula treats a swing in win probability the same in April as in October. That understates the postseason, where a close call in a tight spot can end a season rather than a Tuesday. For postseason games only, two adjustments apply:
- Leverage is squared. In the regular season a leveraged moment is one of 2,400 games. In October it is one of a few dozen. Squaring lets a small miss in a huge spot outrank a big miss in an ordinary one.
- A round multiplier is applied on top: Wild Card ×1.0, Division Series ×1.1, League Championship Series ×1.2, World Series ×1.35.
- An elimination multiplier for a game a team can be knocked out of: ×1.25 when one team faces elimination, ×1.5 when both do and the winner takes the series.
Elimination is worked out from the series record going into the game, derived from the results of the earlier games in that series. We do not use MLB's own series status field for it: that reports the record after the game, and its win total follows whichever team is ahead rather than a fixed team, so in a series that turns around it silently starts describing the other side.
Elimination is a series-level stake and leverage is a within-game one, and they turn out to be close to independent: across the 2026 Wild Card and Division Series, average leverage was 1.27 in games with nothing at stake and 1.39 in elimination games. A blowout in a win-or-go-home game has low leverage and enormous consequence, which is exactly the gap this multiplier fills.
The exponent is 2, and that ceiling is deliberate. At 2, a 0.79-inch miss at leverage 3.04 correctly outranks a 3.16-inch miss at leverage 1.32. Past roughly 2.2, a 0.22-inch call would start outranking a 1.81-inch blown one, which no one watching the game would accept. We would rather the ranking lean toward the moment than produce a result a fan can immediately see is wrong.
The elimination values were chosen by modelling them against the 2026 Wild Card and Division Series rather than picked for feel. At ×1.25 and ×1.5, exactly one call moves in the top eight of that sample: a 0.79-inch miss in an elimination game passes a 2.14-inch miss in a game one. Gentler values changed no ordering at all, which would have made the multiplier decoration.
Three things this does not change. Regular-season and spring scores are untouched, so the historical ranking is exactly what it always was. Leverage and call magnitude are always reported raw, so a postseason call and a June call can still be compared directly on their parts. And the weighting affects ranking only, never the counts: overturn rates, challenge totals and the postseason comparison use unweighted figures throughout.
Zone tendency: who should be ready to challenge
The umpire card on tonight's games reads whether a plate umpire's zone runs tight (the misses are strikes called balls, so the pitcher and catcher have the better challenge case) or wide (balls called strikes, so the batter does). It is the share of that umpire's overturned regular-season calls that went ball-to-strike, ranked against every umpire with at least 30 overturned calls this season, labelled by thirds. No label is given until at least 10 umpires qualify, since a percentile among three is an artifact of there being three.
Early in a season almost no umpire has 30 overturned calls yet. Until one does, the card uses that umpire's label from last season's full record, ranked against last season's umpires, and says so on the card. Once the umpire qualifies this season, the card switches to this season's numbers. The two seasons are never mixed into one share.
We publish this one and not a per-umpire miss map because we tested both. Splitting each umpire's season into alternating halves, zone tightness correlates +0.40 between halves (a full-season reliability of about 0.57), and umpires in the tightest third of one half still average eight points tighter than the widest third in the other. Where an umpire misses, by contrast, correlates −0.01 between halves against a label-shuffled null: indistinguishable from noise. A card that said "this umpire misses low and inside" would be presenting sampling error as scouting.
One limit worth knowing. This is built from challenged calls that were overturned, which is where challenges succeed, not a full record of where umpires err. A location signal could exist in every called pitch and be invisible in this slice.
What counts
- Regular-season ABS pitch challenges are the default everywhere a leaderboard ranks calls. Spring training is shown separately and labeled — spring uses looser rules and far more games, so the two should never be mixed.
- Overturned vs. confirmed. A challenge is "overturned" when ABS reverses the umpire and "confirmed" when it agrees. Success rate is overturned ÷ total challenges.
- The robbery leaderboards rank by Robbery Score — overturned calls only, sorted by leverage × magnitude.
Data source & update cadence
All data comes from the public MLB Stats API — the same feed that powers official box scores. We sync the previous day's games every morning at roughly 6 AM ET, so leaderboards reflect completed games. Pitch locations, umpire assignments, and challenge outcomes all come from that feed; we do not hand-enter or estimate calls.
Limitations & honest caveats
- Leverage Index is an approximation. Use it to compare the relative stakes of two challenges, not as an exact win-probability swing.
- Baserunner state is coarse. The MLB feed reports baserunners as a category (empty, men on, scoring position, loaded) rather than exact bases, so LI uses the most common configuration for each. It barely affects ranking but isn't exact.
- Counts are ABS-only. Every total and leaderboard here counts only ABS pitch challenges (the MLB feed's
MJreview type), matching how the official ABS leaderboard is defined. A small tail of non-ABS reviews (manager, fair/foul) exists in the feed but is excluded from these numbers.
Using & citing this data
Journalists, broadcasters, and researchers are welcome to cite Abstap. Please attribute figures to Abstap (abstap.com) and link to the relevant page. For citation guidance, story angles, and data requests, see the press & data page.
- The Robberies leaderboard. Every overturned call, ranked by Robbery Score.
- Umpires by overturn rate. Whose called zone diverges most from the rulebook.
- Team challenge stats. Which teams challenge most, and which win most.