‘PERFECTLY IMPERFECT’
MLB players come to defense of umpires
As Major League Baseball workshops automated strike zones in the minor leagues and moves toward perfecting technology that can track balls and strikes, those who play, litigate and consume the game find themselves at a pivot point where reality remains disconnected from the idyllic dreams of a mistake-free universe.
On the field, players learn from a young age how to navigate imperfect outcomes, desiring a more just ruling of balls and strikes even as they accept that it won’t always be so. Far beyond the playing surface, fans fueled by pinpoint technology and an itchy grievance finger point to ostensibly unimpeachable graphics that indicate what’s a ball, what’s a strike and what’s worth railing about. (Short answer: Anything a centimeter outside an arbitrary box.)
And that disconnect seems destined to widen well before so-called “robot umps” fulfill their optimized fantasies.
Yet for players, the cry for an immediate, computerized strike zone hits different – particularly when the median major league umpire calls balls and strikes at a 93.5% accuracy rate, according to one respected tracking service.
“Baseball’s perfectly imperfect,” says Pete Alonso, the two-time Home Run Derby champion on his way to a third All-Star Game appearance for the New York Mets. “People mess up all the time. To expect people, human beings, to be perfect, is ridiculous. It’s a perfectly imperfect game, and that’s why I love baseball.”
Alonso’s sentiments are largely echoed by a group of a dozen players queried on the future of umpiring by USA TODAY Sports. They see a game litigated at a level higher than possibly ever before, a concept buttressed by anecdotal and statistical evidence. Above all, there is a consensus often lost on observers who view balls and strikes as absolute:
The strike zone is not an end zone ringed by pylons, or a hoop uniformly hanging 10 feet above the ground. Rather, it is an idea, with a rulebook definition but no solid ground to rest upon, not with hundreds of batters ranging from Altuve-compact to Judge-expansive, some crouching, others upright, all tracking pitches increasingly approaching 100 mph with diabolical movement not seen in prior eras.
Tough to hit a spot. Tougher to hit. And toughest, maybe, to call.
“Being an umpire is a hard-ass job,” says Astros veteran reliever Ryne Stanek, whose four-seam fastball ranges between 97 and 100 mph. “Your job is based on centimeters, and everything is happening in four-tenths of a second or quicker. Everybody’s throwing harder. Time becomes shorter.
“It’s a mutual level of respect. We know their jobs are hard. They know our jobs are hard. Their jobs are equally difficult to read, react, determine in split seconds.”
So why not just take the job away from them?
‘The box that we are given’
MLB’s rulebook hasn’t changed much over the decades when it comes to the most important bit of real estate in the game, terming the strike zone “the area over home plate from the midpoint between a batter’s shoulders and the top of the uniform pants – when the batter is in his stance and prepared to swing at a pitched ball – and a point just below the kneecap.”
What has changed is the means of interpretation.
As the ball smacks the catcher’s mitt, TV viewers immediately see whether modern pitch-tracking technology deems it a ball or strike, though the placement and presentation of the determining box could vary if the game is broadcast on ESPN and its “K Zone” or a regional sports provider. Millions of people who have downloaded MLB’s app or are watching on a tablet or desktop can learn if the pitch was “truly” a ball or a strike, thanks to Statcast data.
And should the call go against the home team, righteous indignation is just a few clicks away.
“It’s just that people love to complain,” acknowledges Mets starter Chris Bassitt. “I think players and fans have become more prone to get angry, faster, just because it’s instant, ‘We know if you’re right or you’re wrong.’ And back in the day that just wasn’t the case.”
Yet the interpretation process is only just beginning.
One to two minutes after the public accesses a ball/strike call and responds in kind, the pitch data lands on league-approved iPads distributed throughout dugouts and bullpens. On this preloaded program, the box players see mirrors the online Gameday data but may differ from what’s on TV, and definitely from what the league will use to evaluate the umpire.
Yet when it comes to nightly questions confronted by hitters and pitchers alike – How wide will he go? Was I robbed, or was he right? – it is the only box that matters.
“A lot of times,” says Yankees righthander and union representative Jameson Taillon, “hitters will go on there just to understand where tonight’s strike zone is. At least if it’s consistent they can say, he’s giving a little bit on this side of the plate, he’s tighter over here. So it’s more about that, as opposed to looking at it and yelling at the guy immediately.”
And even if that’s the goal, reality often delivers a different verdict.
“I run into, maybe, two calls a game that are questionable,” says Mariners ace Logan Gilbert. “I go back and look, and half the time, I’m the one that’s wrong.”
Perhaps therein lies the disconnect between extremely online fans and a more sanguine player population that scarcely has the time nor emotional bandwidth to dwell on a potentially missed call that may amount to a coin flip.
“I feel like there’s a trap, to an extent, where some players will say, ‘Hey, the box says it’s a ball,’ ” says Bassitt. “And then two at-bats later, they’re like, ‘This box is wrong. They said it was a strike.’ You can’t play both sides of the fence. The box that we are given I take as the truth.”
The enemy they know
Ump haters, er, raters, set their mental timers for mid- to late mornings. That’s when the popular Twitter account Umpire Scorecards drops its analysis of ball-strike performances from the night before.
The account has more than 250,000 followers, and the site’s co-founder, Boston University computer science and statistics major Ethan Singer, says there are three scenarios in which its Twitter handle is guaranteed disproportionately large engagement:
• Any game involving the Yankees.
• Games umpired by top-rated umpiring cult hero Pat Hoberg.
• And every game in which popular whipping boy Angel Hernandez is calling balls and strikes.
When Hernandez missed an apparent ball call on a full-count pitch to Kyle Schwarber in a nationally televised game in late April, the Phillies hitter exploded on Hernandez, earning a quick ejection from the umpire and immediate martyrdom from many corners of baseball fandom.
“I’m not here to bury anyone,” Schwarber said, as diplomatic as possible, after the 1-0 loss, “but it wasn’t very good. I don’t know how to really say it.”
MLB’s Statcast was first to weigh in: Hernandez called 16 strikes that were out of the zone, nine against the Phillies, seven against the Brewers. The next morning, Umpire Scorecards’ rating painted an uglier picture: Hernandez’s 88% marks on accuracy and consistency were well off seasonal averages, and some of the “misses” under Umpire Scorecard’s deeper microscope looked closer to Upper Darby than south Philly.
The online verdict drew more than 6,000 retweets of fans going in for the dunk and 25,000 likes. It didn’t particularly matter that Brewers closer Josh Hader had not – and until June 7 would not – yielded an earned run, nor that the Phillies’ win expectancy was just 7% when Schwarber stepped in.
Umpire Scorecards’ algorithm uses a dizzying array of calculations – Monte Carlo simulations and collision geometry among them – to create what it calls “interpretability, validity, practicality, and fairness” to determine a call’s accuracy. The data is rooted in MLB’s Hawkeye system and strike zones set before each at-bat by Statcast. Singer, a rising junior at BU, runs the site with Penn undergrad Ethan Schwartz and says their ratings produce greater specificity. The error distribution of missed calls are publicly available data, and Singer created a normal distribution model that simulates what every pitch’s true location would be.
And runs the simulation 500 times, for every borderline pitch.
“Instead of saying, yes or no it was in the zone, we say, of the 500 possible locations it could have been in, we have a probability, instead of a binary,” says Singer. “I won’t say it’s more lenient (than Statcast and Hawkeye), but it’s more fair.”
A glance at the group’s work beyond a single-game result reveals a few data points that might surprise drive-by viewers: The median umpire renders a correct call on 93.5% of pitches – with an elite group of eight umpires averaging at least 95% in accuracy and 94% in consistency.
And Hernandez, scuffling along at 92.7% accuracy in nine appearances behind the plate this season, is far from the worst umpire in the league.
Through June 1, Hernandez was in a three-way tie for 73rd among 85 umpires with at least four games behind the plate, well above last-place Andy Fletcher’s 89.9% mark. It puts Hernandez less than a percentage point behind median man Todd Tichenor’s 93.5% and less than 4% worse than the gold standard – Hoberg’s 96.2% accuracy.
That’s 144 of 150 correctly called balls and strikes over a typical nine-inning game – and a difference of just five correct calls separating Hoberg with Hernandez.
How about Jeremie Rehak, Charlie Ramos, Shane Livensparger, Tripp Gibson, Alfonso Marquez, Adam Hamari? We’re guessing beyond 22-year veteran Marquez, you haven’t heard of these guys, either, but they join Hoberg in the elite 95% accuracy/94% consistency club.
Meanwhile, an informal poll of players revealed exactly none had heard of Ryan Blakney or Stu Scheurwater.
Both called near-perfect games of 98% accuracy recently, with Blakney’s masterpiece coming in the only solo nohitter thrown this year, by the Angels’ Reid Detmers. That’s going 147 for 150 under Umpire Scorecards’ lens, with little notice – as it should be.
Perhaps the robot umps are already here.
“I believe they’re working on it to get better. I believe also that (umpires) are better since I’ve come up,” says Nationals DH Nelson Cruz, who has hit 454 home runs since his 2005 debut. “They are more exposed now, because they show the strike zone and everybody’s aware. More consistent. More precise.
“You know what to expect. You know, ‘He likes to call a pitch outside, in the corner, away.’ So you’re more aware you have to swing at those pitches, protect the strike zone.”
‘Offense would die with the robo ump’
The drumbeat of the robo ump has only grown louder in recent years, as technology improved and MLB tested it in the usual workshops – the Arizona Fall League and the independent Atlantic League. MLB’s adoption of the more accurate Hawkeye tracking system and its May 19 implementation of the automated zone in the Pacific Coast League – one stop from the majors – means the robots are lurking.
It is potentially close, but not imminent. Guardians manager Terry Francona has called the auto zone “not ready” and the strike zone “not good.”
A switch to Hawkeye technology – a rousing success in its ability to call lines in tennis – has upside beyond a potential new strike zone. Hawkeye’s more perfect data has significantly boosted the effectiveness of MLB’s umpire evaluations, according to a person familiar with the details. The person spoke to USA TODAY Sports on the condition of anonymity because they weren’t authorized to speak publicly.
That evaluation process includes use of a more sophisticated strike zone, which undergoes an overnight auditing process that confirms the placement of the strike zone lines.
And the use of automated balls and strikes is not necessarily paving a through line to a binary future of robots or humans. More accurately, the person said, the minor league use of the Automate Ball/Strike System (ABS) is aimed at learning what may prove feasible on an expanded basis.
The future, in short, could look more nuanced than the opposing poles in each camp might envision.
Players are largely leery of a system that may upend what they believe to know are balls and strikes, particularly pitches that land in the dirt but perhaps dot the bottom of the box on their way by. There are aesthetic concerns as well, such as a pitcher missing his spot by a foot, the catcher diving to reel it in, yet the pitcher’s failure rewarded with a rule-of-the-law strike.
“I still love the idea of, ‘You hit your spot, you get rewarded for it,’ although that happens less and less now,” says Taillon. “Whereas if you have a robo situation, I think pitchers will also get curveballs at the top of the zone, stuff that I don’t expect to ever get as a strike. The game might change a bit.”
Call it concern or paranoia, but players remain wary when significant rules changes that could even benefit them – such as a ban on defensive shifts – create the potential for deeper, unintended consequences.
“I think whoever thinks robo umps are a good idea don’t know what they’re talking about. I think offense would die with the robo umpire,” says Stanek. “If the strike zone was explicitly called the way a robo umpire would call it, you get the whole zone – top to bottom, both sides. The really bad off-speed pitches that hit the top of the zone – they get balled all the time. And you don’t expect a strike because you made a bad pitch and you’re like, ‘That’s a ball.’
“But the robo umpire doesn’t discriminate. It just goes, ‘That’s a strike,’ and the hitter’s gonna be like, ‘Well, I have to swing at the really bad breaking ball at the top, but also the really, really good breaking ball at the bottom.’ It’s going to affect hitting, probably in a negative way.
“I don’t think the negative consequences are considered enough. Are they going to change the strike zone to make it harder to pitch? I’m sure the league doesn’t want offense to disappear.”
‘Human error needs to be a thing’
Singer, the Umpire Scorecards creator, says he thinks those problematic ball calls could be averted through multiple means, including improved technology when it comes to drawing an imaginary strike zone.
Failing that, he says, is another, more sacrosanct target.
“We might have to change the rules,” he says, “to align with the automation. (Bad calls) are less a robo ump’s concern
and more a concern with rules.”
Until the technical kinks are smoothed out and stakeholders agree the time has come for automation, MLB will be wedded to its “blues.” That includes the Cuba-born Hernandez, who unsuccessfully sued MLB for discrimination, claiming he was denied crew chief and postseason assignments based on his race and not performance.
In 2023, Hernandez will mark his 30th anniversary as an MLB umpire To the players?
Mention his name in private and the reaction is largely ambivalent. “Just another guy,” says one pitcher. “If he’s working a game in Seattle and I’m here, I’m not thinking about him,” offers another.
A hitter opts to cut the conversation short rather than opine. “The game’s hard,” says one player, letting his forthcoming silence tell plenty about his feelings for Hernandez. “We’ll leave it at that.”
Factoring temporary call-ups and other substitutes, players will deal with roughly 100 umps behind the plate. The greater enemy, always, is the grind, and within that the hope that over six months and up to 700 plate appearances, the good and bad calls will be a wash come October.
“If your team is winning and you’re playing well, it’s like all right, whatever, they’ll even out,” says Rockies third baseman Ryan McMahon. “And if it’s going bad, it’s like ah, you’re getting screwed all the time. But they do even out more than people realize.”
“In sports, human error needs to be a thing,” says Bassitt. “I think that’s what makes the sports in general beautiful is that there is human error, not just from the athletes but the people umpiring and officiating, managers calling the game, general managers ... everybody makes mistakes.
“We’re all human. We’re all trying to do our best. And I think that’s the beauty in it all.”