Every umpire's zone is a little different, and a zone that is an inch wider at the knees turns walks into strikeouts. But the effect on a single game's score is easy to overstate, because every umpire also works games in every kind of park, weather and matchup.
How the model measures it
For each game the umpire has worked, the model compares the total runs with what was expected given the league's scoring level and the park. It averages those differences over up to seven seasons (Retrosheet game logs extend the history) and then regresses heavily — the equivalent of 300 games of 'average umpire' is added before dividing. An umpire has to work hundreds of games before his number moves much.
What it finds
Among the 70 umpires with at least 150 plate assignments since 2019, the spread of regressed effects is about 0.11 runs per game (one standard deviation). The most extreme are around a quarter of a run in either direction — Andy Fletcher and Mark Wegner on the high side, Brennan Miller and Cory Blaser on the low side in 2026. A quarter of a run is real but modest: less than the difference between a 65°F and a 75°F night.
When it is known
Umpire crews rotate on a predictable schedule and home-plate assignments are posted on game day. Until then the umpire effect is left at zero, and each game page shows it once it is known.