By Zach Helfand

August 23, 2021

 

https://www.newyorker.com/magazine/2021/08/30/invasion-of-the-robot-umpires

 

Grown men wearing tights like to yell terrible things at Fred DeJesus. DeJesus is an umpire in the outer constellations of professional baseball, where he’s been spat on and, once, challenged to a postgame fight in a parking lot. He was born in Bushwick, Brooklyn, to Puerto Rican parents, stands five feet three, and is shaped, in his chest protector, like a fire hydrant; he once ejected a player for saying that he suffered from “little-man syndrome.” Two years ago, DeJesus became the first umpire in a regular-season game anywhere to use something called the Automated Ball-Strike System. Most players refer to it as the “robo-umpire.” Major League Baseball had designed the system and was testing it in the Atlantic League, where DeJesus works. The term “robo-umpire” conjures a little R2-D2 positioned behind the plate, beeping for strikes and booping for balls. But, for aesthetic and practical reasons, M.L.B. wanted human umpires to announce the calls, as if playacting their former roles. So DeJesus had his calls fed to him through an earpiece, connected to a modified missile-tracking system. The contraption looked like a large black pizza box with one glowing green eye; it was mounted above the press box. When the first pitch came in, a recorded voice told DeJesus it was a strike. He announced it, and no one in the ballpark said anything.

 

The eeriest thing about the robo-umpire is the silence. This summer, I attended some games in Central Islip, New York, home of the Long Island Ducks, to check it out. The pizza-box device is made by a company called TrackMan, founded by two Danish brothers, Klaus and Morten Eldrup-Jørgensen, who created it to train golfers. It is easy to miss. At one of DeJesus’s games, I observed a kind of Turing test. Starting in the fifth inning, a lanky middle-aged guy behind home plate started heckling. “Move the fucking game along!” he said, after DeJesus announced a ball. A few minutes later, after a call he disliked, he yelled, “Look at him! How can he even see over the catcher?” A man in a Mets cap nearby pointed up at the device, explaining that the calls were automated. The heckler appeared confused: “Can he overrule it?”

 

Mets Hat shook his head. The heckler, looking embarrassed, replied, “He’s called a good game, I gotta say!”

 

Baseball is a game of waiting and talking. For a hundred and fifty years or so, the strike zone—the imaginary box over home plate, seventeen inches wide, and stretching from the batter’s knees to the middle of his chest—has been the game’s animating force. The argument between manager and umpire is where the important disputes over its boundaries are litigated. The first umpires were volunteers who wore top hats, at whom spectators “hurled curses, bottles and all manner of organic and inorganic debris,” according to a paper by the Society for American Baseball Research. “Organic debris” wasn’t defined, but one wonders. A handful of early umpires were killed.

 

Rules of engagement evolved in fits and starts. Today, everyone knows that an aggrieved party can kick dirt, but not over the plate, which the umpire maintains with his special brush. You may scream in an umpire’s face, but you must never touch him. Kevin Costner’s character in “Bull Durham” doesn’t get ejected when he says that the ump made “a cocksucking call,” but he does when he calls the man himself “a cocksucker.” That’s a no-no. Lip-readers or hot mikes sometimes reveal these arguments to be admirable examples of candor and of dispute resolution—two stressed-out guys trying their best, with fans or bosses breathing down their necks. More often, arguments are like stock-car wrecks: grotesque, morally indefensible, and the thing a lot of people secretly root for. In 1980, the umpire Bill Haller wore a wire during a dispute with Earl Weaver, the Baltimore Orioles manager at the time:

 

Weaver: You’re here and this crew is here just to fuck us! (Haller ejects Weaver.) That’s good! That’s great! And you suck!

 

Haller: Bah, you shit! (Haller points his finger at Weaver.)

 

Weaver: Get your finger off of me! (Weaver slaps Haller’s finger away.)

 

Haller: I didn’t touch you!

 

Weaver: You pushed your finger into me!

 

Haller: I did not! Now you’re lying!

 

Weaver: No you are!

 

Haller: You are lying!

 

Weaver: You are a big liar!

 

Haller: You are a liar, Earl!

 

Weaver: You are!

 

This continued for nearly three minutes.

 

When the robots came, the arguments basically stopped. After the Ducks game, I met DeJesus outside the ballpark. “There were six calls that I disagreed with,” he said, referring to the words that came through his earpiece from the robot. “One pitch was right down the middle. I went to call strike three, and it said, ‘Ball,’ and I went, ‘Ball!’ And I looked at both dugouts.” No one had come out to argue. He continued, “I miss the battles.” In his day job, DeJesus works as a special-education teacher on Staten Island. His commute to Islip can be three hours. The Atlantic League pays him a hundred and sixty dollars a game. His dream is to umpire the College World Series. He trains himself using a virtual-reality headset, and he rewatches footage after every game. He has worked more than six thousand games and called upward of half a million pitches. “When I first heard about A.B.S., I was very angry,” he said. Rick White, the Atlantic League’s president, told me, “We had some umpires go rogue. A very small percentage of them.” They refused to call the pitches that the system called. One unhappy umpire called a game from ten feet or so behind his usual position, as a protest. But the system won DeJesus’s respect. It was, he admitted, better than him.

 

During the first robo-ump season, players complained about some strange calls. M.L.B. tweaked the dimensions of the zone, and this year the consensus has been that A.B.S. is profoundly consistent—and bound for the major leagues. The Ducks manager, the former Mets second baseman Wally Backman, is known for being an enthusiastic arguer; he once threw dozens of bats onto the infield after an ejection. (“Pick that shit up, you dumb motherfuckers!”) But he loves the machines. Smoking Marlboro Reds in the grandstand one day, he told me, “It’s gonna be in the major leagues in a lot shorter time than people think.” M.L.B. has already concluded that the device is near-perfect, precise to within fractions of an inch. “It’s going to be more accurate, it’ll reduce controversy in the game, and be good for the game,” the M.L.B. commissioner, Rob Manfred, has said. But the question is whether controversy is worth reducing, or whether, like the scratches and grooves on a vinyl LP, it is the sign of a human hand. Joe Torre, the former Yankees manager, who now works in the commissioner’s office, has argued publicly against the robots. “It’s an imperfect game and has always felt perfect to me,” he said.

 

A human, at least, yells back. When I spoke with Frank Viola, the pitching coach for the High Point Rockers, an Atlantic League team in North Carolina, he said that A.B.S. worked as designed, but that it was also unforgiving and pedantic, almost legalistic. “Manfred is a lawyer,” Viola noted. Some pitchers have complained that, compared with a human’s, the robot’s zone seemed small. Viola was once an excellent big-leaguer himself. When he was pitching, he said, umpires rewarded skill. Throw it where you aimed, and it would be a strike, even if it was an inch or two outside. There was a dialogue between pitcher and umpire. During the first inning of the Rockers’ first game using A.B.S., Viola said, “my guy on the mound threw three pitches right there. And all the pitches were strikes!” A.B.S. said otherwise. This got Viola frustrated. Which is how he became the first person to get ejected for arguing with the robot.

 

Machines replaced the film projectionist and the subway attendant, and, chances are, they will eventually replace us all. The umpire can already seem a man out of time, like a milkman or a doctor who makes house calls. Maybe it’s the uniforms. The average umpire is male, white, and conservative. (No women have worked the majors outside of spring training; until last year, there were no Black crew chiefs.) Perhaps he smokes Winston Lights. His backup career may have been in law enforcement. A visitor to an umpire-training academy twenty years ago discovered that everyone there was obsessed with “NYPD Blue.” Umpires are talented, diligent, and seem to be ethically unimpeachable—there’s been only one case of umpire corruption, ever, and that was in 1882. But accuracy fluctuates by era. There are compelling claims that the nineties were anarchy. (Ted Barrett, a Christian minister, and an umpire since 1994, once recalled that, when he started out, the profession was full of boozing and carousing. “How can I put this delicately?” he said. “It was a devil’s playground.”) In response, in 2001, M.L.B. instituted video evaluations to enforce uniformity. The league says that umpires now call an astounding ninety-seven per cent of pitches correctly.

 

The evaluations began a season before Michael Lewis started working on his book “Moneyball.” Soon, teams, in their thirst for data, began using tracking systems to measure such things as a ball’s velocity off the bat and a pitch’s spin rate. Fans could access the data online. It was suddenly possible to know every time an umpire erred. In a typical season, one study showed, this happened about thirty-five thousand times—enough to decide a game’s winner and loser regularly. Calls for automation grew insistent.

 

The executive tasked with running the experiment for M.L.B. is Morgan Sword, who’s in charge of baseball operations. He’s red-headed, thirty-six, and amiable, a boyhood fan of the Mike Piazza Mets. In late spring, I joined him at the baseball headquarters, in midtown, along with Reed MacPhail, who oversees the system’s testing and validation. MacPhail played ball, briefly, at Claremont McKenna College. His batting average was .833. Four of his five hits, he noted, came against CalTech, which hadn’t won a game in twenty years.

 

According to Sword, A.B.S. was part of a larger project to make baseball more exciting. Executives are terrified of losing younger fans and worry that the sport is at risk of becoming the next horse racing or boxing. “We started this process by asking ourselves and our fans, ‘What version of baseball do you love the most?’ ” he said. Everyone wanted more action: more hits, more defense, more baserunning. This style of baseball essentially hasn’t existed since the eighties. The “Moneyball” era and the hundred-mile-an-hour fastball, difficult to hit and to control, have flattened the game into strikeouts, walks, and home runs—actions lacking much action.

 

Sword’s team brainstormed potential fixes. “Any rule that we have, we’ve talked about changing: change the bats, change the balls, change the bases, change the geometry of the field, change the number of players on the field, change the batting order, change the number of innings, the number of balls and strikes,” Sword said. “We talked about regulating the height of grass on the infield to speed up ground balls and create more hits. We’ve never talked about this in any serious way, but we talked about allowing fans to throw home-run balls back and keep them in play. That’s one that I don’t even like.”

 

Sword views A.B.S. “not as a change in itself but as a vehicle. Once you get the technology right, you can load any strike zone you want into that system.” A strike zone exists that could create a perfect version of baseball, but it might be a triangle, or a blob, or something shaped like Texas. Sword and MacPhail toyed with ovals and slanted rectangles. “A lot just didn’t pass the test of ‘If you’re playing Wiffle ball in the back yard, could you enforce that strike zone?’ ” MacPhail said.

 

Over time, as baseball evolves, A.B.S. can allow the zone to change with it, functioning like an engine’s governor. “The human umpires are remarkably accurate, and they’re the best in the world at what they do,” Sword said. But learning and calling a new strike zone could take years. “On A.B.S., it’s literally a matter of, like, changing a setting.” M.L.B., in its labor deal with the umpires’ union, which declined to comment, agreed to include the union in any plans to use A.B.S. in the major leagues. Such a move would likely meet with resistance from the rank and file. “It is the umpire’s decision to make whether it’s a ball or strike,” Joe West, who earlier this year broke the record for most major-league games umpired (fifty-three hundred and seventy-six), and who formerly served as the union’s president, told me. He argued that a disaster scenario would be a pitch in, say, the World Series failing to register on the machine, leading to chaos. (M.L.B. says such a scenario is highly unlikely, and that, in any case, the human umpire could step in to make the call.)

 

M.L.B. has already concluded that the technology works. Now the organization is measuring outcomes. This year, it rolled out the experiment to a class-A league in Florida. (That league uses a device made by a company called Hawk-Eye, instead of the one from TrackMan; M.L.B. is likely to use Hawk-Eye if the system reaches the major leagues.)

 

Sword invited me to watch a Ducks game wearing an umpire’s TrackMan headset. It was a pleasant summer night. A few kids blew duck whistles. The TrackMan’s green eye glowed. The “strike!” call in my ear was peppy, congratulatory. The “ball” sounded faintly disappointed. I followed each pitch on an app, which displayed the ball’s location as it crossed the plate. I tried to guess each call. Even from my seat directly behind home plate, I barely had a sense of whether a ball was a foot outside or right down the middle. It was pointed out to me that, were I to switch places with the umpire, almost no one would notice.

 

Before another Ducks game, I visited the umpires’ locker room. DeJesus wore a T-shirt that said “ring em up.” John Dooley, the Atlantic League’s supervisor of umpires, was sitting nearby. The umpiring crew was talking about a robo-umped Atlantic League game the previous evening in Lancaster, Pennsylvania.

 

“Wanna know how long it took?” Dooley said. He had a Queens accent so thick it sounded Bostonian. “Five hou-ahs! Sixteen to fourteen. Nine innings.”

 

“Thirty-five walks!” DeJesus said—a horrific amount.

 

I asked DeJesus if he’d ever called a game with thirty-five walks.

 

“With TrackMan or without?” he said. “Without, it’s called ‘pitch management.’ A lot of guys call it ‘cheating.’ If I start to feel that the game is dragging and we’re not getting a flow, you’re gonna have more strikes called. Not anymore. It used to be, if you have two borderline pitches in a row, one gets called a strike, one gets called a ball. Everybody is equally upset, and everybody’s equally happy. For me, it’s ‘Can we get through this today without everybody killing each other?’ ”

 

In the past twenty years, sports have moved away from these kinds of judgment calls, seeking more precision. This may be owing to technological improvement, or to corporatized gambling. (“People are betting a lot of money,” Joe Maddon, the Angels’ manager, who in 2016 led the Cubs to their first World Series in a hundred and eight years, explained before a recent game. “They truly want the accurate outcome.”) Soccer has Video Assistant Referees. Tennis has Hawk-Eye. For almost a decade, baseball has used instant replay on the base paths. This is widely liked, even if the precision can cause its own problems; one umpire told me he had to overturn a call when the video showed a loose string on a fielder’s glove grazing a runner’s back—technically, this counted as a tag. But these applications deal with something physical: bases, lines, goals. The boundaries of action are precise, delineated like the keys of a piano. The strike zone is a fretless bass. Historically, a certain discretion has been appreciated.

 

For many years, an umpire’s strike zone was like an extension of his personality. Some umpires were literalists, uncompromising. Some preferred expediency; their boundaries were enormous. No matter who was working, when it rained suddenly everything was a strike. West, the record-holding umpire, is a burly man with a Carolina drawl who moonlights as a country singer and used to pal around with Merle Haggard. He told me one umpire described the old standard for learning the strike zone as “You call them strikes until someone goes, ‘Hey!’ ” Another of his friends liked to say, “The strike zone is like a television set, and every now and then you need Earl Weaver or Billy Martin”—the Yankees’ volatile manager in the seventies and eighties—“to come out and adjust the knob.” Martin once sent an umpire a Christmas card that read “I hope you and your family have a wonderful holiday season.” On the inside, he wrote, “Because you sure had a horseshit summer.” Video evaluation has reined in some quirks, but the strike zone still changes measurably depending on the score, the team batting, and the pitcher’s race. When a pitcher is struggling, the zone becomes as much as fifty per cent bigger. This is known as the “compassionate-umpire effect.”

 

Of course, compassion toward the pitcher is cruelty toward the hitter. “I don’t know of any other sport in which the umpires would even talk about making up their own rulebook,” Bill James, a writer and a former Red Sox executive, widely considered the godfather of advanced statistics, told me. Joe Sheehan, a sportswriter and one of the earliest and most fervent proponents of the automatic zone, told me, “I get literally angry when I see a pitch three inches off the plate called a strike. Like, No way. The hitter did his job, and this middle manager behind the plate basically reversed what should have happened.” The Supreme Court Justice Brett Kavanaugh, a textualist, once delivered a treatise called “The Judge as Umpire.” “We do not design our own strike zones,” he said.

 

In 2019, the robo-ump followed the rulebook strike zone to the letter. Players complained: it was too high, too narrow, and things got weird around the edges. M.L.B. adjusted the parameters in 2021. A three-dimensional zone was jettisoned for a two-dimensional one. The zone was shortened and widened. Now a ball skimming an inch and a half off home plate counts as a strike.

 

This has always been a gray area. Context determined the call. In 1956, the Yankees’ Don Larsen delivered baseball’s most famous pitching performance, a perfect game in the World Series. The game ended on a called strikeout. The umpire was Babe Pinelli—a newsboy at ten, a steelworker at twelve, he’d called thirty-four hundred games in a row without sitting one out. The last pitch of the last at-bat of his last game behind the plate is the only one anybody remembers. Most observers swear that it was noticeably outside. Stephen Jay Gould, in “Triumph and Tragedy in Mudville,” argued that the pitch did miss the zone—and that Pinelli was correct to call strike three. It was close enough for history. “Truth is a circumstance, not a spot,” Gould wrote.

 

In 2010, Armando Galarraga, a pitcher for the Detroit Tigers, was one out away from his own perfect game. With two outs in the ninth inning, he induced a weak ground ball. The throw beat the runner by a step: an easy call. Inexplicably, the umpire in the field, Jim Joyce, called the runner safe. The perfect game was ruined. On seeing the replay after the game, Joyce was distraught. He cursed in the clubhouse. He spent the night at his mother’s (she lived nearby), chain-smoking. He received death threats. The next day, on the field, Galarraga embraced him and released him from guilt. Joyce cried. They later became friends.

 

Within a few seasons, M.L.B. instituted instant replay for plays on the base paths. “I lovingly say that I’m the poster boy,” Joyce told me recently. I asked if he felt the same way about A.B.S. What if there had been robo-umps when he was breaking in to the minors? “I would have continued driving a truck for U.P.S.,” he said. “Anybody can go back there and have somebody else tell you what to do. You might as well just watch a video game.” I asked if he thought it best that his bad call couldn’t be overturned, given how it all worked out. “Ask Armando,” he said.

 

When I did, Galarraga told me, “The story is so beautiful because of what happened in the end.” He wouldn’t want to change it. He argued that mistakes are part of the game—accepting them with composure or exploiting them to your advantage are skills, just as throwing a good fastball is. He noted that, before he lost his perfect game, he’d had success with his sinkerball, a pitch he threw a couple of inches off the outside edge of the plate. He’d realized that the umpire was consistently calling it a strike. Using A.B.S., he said, would be like driving a car with a navigation system: “When you don’t have the G.P.S., you have to pay way more attention to the directions. You have to be more in the moment. This is the beauty of the game.”

 

Bill James suggested that enough significant mistakes from umpires could, from a fan’s perspective, make a game seem almost arbitrary. He offered a middle ground: A.B.S. could rule on obvious balls or strikes. But a couple of inches around the border would be a “zone of discretion”—up to a living, breathing umpire to decide.

 

Alva Noë, a professor of philosophy at Berkeley, completed “Infinite Baseball: Notes from a Philosopher at the Ballpark” just before M.L.B. announced the robo-umps. I e-mailed him, wondering if he’d given the new development any thought. “Hardly a day goes by that I don’t wake up and run through the reasons that this is such a terrible idea,” he replied. He later told me, “This is part of a movement to use algorithms to take the hard choices of living out of life.” A pitcher has a strange job. He wants to throw a pitch that is hittable but that the batter, nevertheless, can’t hit. Noë argued that what the umpire adjudicates is whether the pitcher succeeded. It’s inherently a judgment call. “What we’re seeing in baseball is something that is kind of a core dispute in Western civilization. It really is about ‘What is objectivity?’ Is objectivity something that is physical? Is it mathematical? Is it knowable?”

 

But people want answers. Subjectivity, generally, is on the run. We prefer Yelp stars and Big Data dating apps. May a thousand Tomatometers bloom. Recent decades have birthed baseball’s own theory-of-everything statistic, war (wins above replacement), which does a decent job of ending the barstool argument about which player is the best. The big-money jobs these days are in data analytics; hard numbers make vaccines and launch rockets. If you’re trading baseball players, you’d better know their value. But what you’re measuring matters. Accuracy is not the same thing as enjoyment. We watch baseball to kill time, not to maximize it.

 

TrackMan was created to quantify and optimize. The company broke into baseball, as a player-analysis tool, in 2008. Soon, it was everywhere. “There was a while there when we were tracking and facilitating a sharing of data from just about every pitch of every play of every professional game in the world,” John Olshan, who runs the company’s baseball business, told me. For the robo-umpires, professional baseball was only a starting point. Olshan predicted that the system would reach all levels of the game, down to Little League. TrackMan sells a portable version of the hardware, intended to help players train. But it can also tell you if a pitch is in the strike zone. He invited me to try it out in a beer league.

 

The Passaic Bulls, according to their player-manager, Joe Moran, are perhaps the best men’s-league team in New Jersey. “You say our name, people know us,” he told me. Moran is twenty-six, a former construction worker—Local 754. During the pandemic, he lost his job, so he became a covid disinfector. “One of those guys who comes into a place and fogs it up, with the hazmat suits and all that,” he said. “I honestly think I put more hours into baseball.”

 

Moran happily agreed to let me use the Bulls as robo-umpire guinea pigs. “There’s a big umpire shortage, so some of the umpires they get us are really bad,” he told me. I’d heard this often. Gil Imber, a former recreational umpire, who runs something called the Umpire Ejection Fantasy League (“Joe West is always a top draftee”), told me that the supply of umpires generally runs opposite to the economy. Strangely, the current unemployment spike hasn’t created new umpires—likely, Imber says, because the pay has stagnated, and the abuse has not.

 

The game was in Passaic at Third Ward Park: old stone wall in the outfield, commuter trains rumbling fifty feet behind the backstop. The infield grass was prodigious, practically a prairie; Sword and MacPhail would’ve brought in the Weedwackers. Ten feet behind home plate, Will Gilbert, a former minor-league pitcher who works for TrackMan, erected a tripod, atop which he mounted the device, a miniature version of the stadium machine. With its little eye, the setup looked like the Pixar lamp. For the experiment, the umpire, a rookie named Mirquis Erazo, was making the calls, but could appeal close ones to TrackMan during the first half of the game. As a control, the second half would be old-school.

 

The Bulls walked the first four batters: no close calls. Tensions soon rose. When the Bulls, now trailing, came up to bat, Erazo called a ball inside, and the opposing team hollered. “That was a strike?” Erazo asked Gilbert.

 

“No, about five inches,” Gilbert said. Not close.

 

“Thank you!” Erazo said. “The shortstop is already looking at me.”

 

Erazo settled in. After four innings, the machine was making him look astonishingly good. TrackMan disagreed with him only four times, and two were really close. After one walk, the entire Bulls infield shouted in protest. Erazo gestured toward the machine. The shouting stopped. At one point, Moran ran up to the fence and yelled, “Yo, this guy’s the best umpire in the league, no question!” TrackMan was a hit. The portable system cost about twenty thousand dollars. By the middle of the game, Moran was talking with Gilbert about buying one.

 

The tight zone, however, made for a long evening. Sometime past 10 p.m., in the eighth inning, the Bulls first baseman, a twenty-three-year-old pest-control technician named Joe Russo, took off his spikes, packed up his gear, and sat down. He took a dissenting stance toward the robots. “With technology, people just want everything to be perfect,” he told me. “That’s not reality. I think perfect would be weird. Your teams are always winning, work is always just great, there’s always money in your pocket, your car never breaks down. What is there to talk about?”

 

Then he realized he was on deck and scampered off, retying his spikes. When his turn to bat came, the Bulls were up eight runs. The game was in its fourth hour—TrackMan had long ago been turned off. Russo took a fastball, quite high. Erazo called it a strike. The second pitch was a strike, too. On the third, he watched as another fastball came in near his shins. Erazo bellowed, “Strike three! ” The Bulls dugout hooted. Moran yelled, “We want the robot back!” Russo looked sheepish. “That was a ball!” he said to Erazo. “But I’m not mad at you!”

 

After the game, I caught up with Erazo in the parking lot. He laughed when I asked about the pitch to Russo. “It was time to go, that’s what it was,” he said. “It was a ball.” I told him I agreed. But who at the park could say for sure? 

 

Published in the print edition of the August 30, 2021, issue, with the headline “Kill the Umpire.”

Zach Helfand is a member of The New Yorker’s editorial staff.