In the top of the 10th inning in Sunday night’s nationally televised contest between the Astros and Rangers — one that will most likely be remembered as the night a 44-year-old nearly no-hit the defending World Series champs — the visiting Rangers grabbed a 3-1 lead.
In the bottom of the frame, the home team’s hopes rested on Jake Marisnick, who, with runners at the corners, two outs, and his team still trailing by a pair of runs, worked a 3-1 count against Jake Diekman. A Marisnick walk would load the bases for the Astros, bringing reigning World Series MVP George Springer to the plate, a hit away from tying or winning the game.
On Diekman’s fifth pitch, it appeared that Marisnick had earned a walk. “This is not a strike, this is off the plate,” ESPN broadcaster Jessica Mendoza opined as the networks’ K-Zone showed the pitch a few inches outside.
Home plate umpire Adam Hamari disagreed, however, calling the pitch strike two. Marisnick struck out swinging on the following pitch to end the game, and the outfielder slammed his bat in disgust.
Umps miss balls and strikes all the time. But the strike two in that Marisnick at-bat is emblematic of a larger pattern of borderline calls, albeit one that umps probably produce unwittingly: In extra innings, umpires will vary ball and strike calls in ways that tend to end the game as quickly as possible.
To find this pattern, we looked at pitches thrown in the bottom of extra innings, when the game could quickly end. If the away team scored in the top half of an inning and held a lead, as was the case in Marisnick’s at-bat, an umpire hoping for a faster exit would call more strikes, making it more likely that the home team will be sent down quickly. Alternatively, if the home team got a runner aboard, umps would be more likely to favor them by calling fewer strikes, giving the team more chances to get the runner across the plate and send everyone home.
Here’s a chart showing how umps changed their behavior in these situations between 2008 and 2016, a sample of roughly 32,000 pitches. Each square shows the percentage increase or decrease in the likelihood that a pitch is called a strike in that part of the strike zone. The color of each square (green for more balls, pink for more strikes) corresponds with which side umps are favoring, while how darkly shaded the square is reflects the size of the change (in percentage points).
The left panel shows the comparative rate of strike calls when, in the bottom of an inning in extras, the batting team is positioned to win — defined as having a runner on base in a tie game — relative to those rates in situations when there’s no runner on base in a tie game. When the home team has a baserunner, umps call more balls, thus setting up more favorable counts for home-team hitters, creating more trouble for the pitcher, and giving the home team more chances to end the game.
The right-hand side of the chart shows squares at identical strike zone locations, but shaded according to changes in strike rates when the extra-inning scenario favors the away team. More specifically, any time the away team is trying to hold onto a lead in the bottom half of an inning after the ninth. Here, and as in the pitch to Marisnick, umps call more strikes, giving the batting team fewer chances to extend the game.
Altogether, teams that are in a position to win get up to a 27 percentage point increase in the rate of called balls, while teams that look like they’re about to lose see increased strike rates of up to 33 percentage points. Differences are largest in fringe areas of the strike zone, where the opportunity for umpire discretion is the highest: 62 percent of these squares in the left panel are green, while 72 percent of fringe squares on the right panel are pink. In both settings, umps are more likely to use whatever behavior gets the game over with the quickest. That may not necessarily be a bad thing. MLB games are already slow, and extra-innings play often comes late at night, which means smaller crowds and fewer television viewers.
MLB did not immediately respond to a request for comment, but the league has made no secret of its interest in shortening games. Even so, umpires may not be consciously deciding who should win. Humans are susceptible to various biases they may not be aware of, and even just a bit of fatigue could unintentionally push umpires in one direction or the other on borderline calls.
Moreover, according to sources within the umpire union, umps don’t get paid more when games go to extra innings. In other words, MLB asks them to take on extra work without providing any extra compensation. That’s one more reason they may want the game to end early — their paycheck’s the same regardless.
We’re defining the fringe area as squares within one square of the black line marking the edge of the strike zone.