Can computers call a better game than human umpires? Major League Baseball wants to find out

For 13 years, Tim Rosso has been calling balls and strikes as an umpire in NCAA DIV 1 baseball and the Atlantic League of Professional Baseball. But in the blink of a 0-1 letter-high fastball (the first pitch of the game was fouled off), his career changed.

Rosso casually righted himself from his crouched position, shuffled his left foot and loosely pumped out his right hand. Not too fiercely like it was an inning-ending strike, but a strike all the same.

But while Rosso signalled the call, he didn't actually make the decision.

"I've been an umpire for 13 years and this is the first time that I've had a machine, you know, tell me what to call," Rosso said after the game.

On that late summer night in New Britain, Conn., in a game between the New Britain Bees and the Long Island Ducks of the ALPB, technology had pierced the heart of America's pastime.

A computer was calling balls and strikes.

"What we're doing is making them better umpires by assisting what they do with just another tool to make them more efficient," says Rick White, commissioner of the Atlantic League of Professional Baseball.

High above the grandstand looking down at home plate sits what looks like a giant flat screen TV. It's actually a system made by TrackManthat uses radar to track pitches as they cross home plate. That data is combined with preprogrammed specifications for each batter in the league, because the height of every batter's strike zone is different. It all comes together on a laptop in the press box, and then the signal as to whether the pitch is a ball or strike is sent to the home plate umpire via wifi.

The ALPB is testing this system for Major League Baseball, with the notion that someday it might be used in the big leagues.

"We think the reception has been so strong, and the interest in this by people surrounding the game and on the field is so strong, that inevitably it's going to find its way to Major League Baseball," White says.

In fact, it's already there.

Just watch any MLB game on television and you can similar systems at work showing whether a pitch crosses the strike zone or not.

And MLB has been using it for more than a decade to evaluate umpires.

As it turns out, they could use some help.

"We found in 2018 there were 34,000 errors made behind home plate," says Mark Williams, a Boston University researcher who studied the decisions made by umpires in MLB games.

That's 14 bad calls a game.

Williams and his team asked for and received from MLB the data it uses to evaluate its umpires. He says he was shocked when he got it.

"I think they probably thought no one would go to the extent that this study did, that we would go over 4 million pitches over 11 seasons and we would show this disturbing trend," Williams says.

Imagine a spreadsheet with more than 4 million lines on it. Each line representing every single pitch. And each line having more than 50 columns of data including the type of pitch, speed, when it was thrown, who threw it, the spin rate, which of 13 zones it crossed in front of the batter, what the umpire called, who the umpire was and most importantly, whether or not it was a bad call.

"Well, in general, umpires get it wrong almost 10 per cent of the time, that's all pitches," says Williams. "But certain pitches they get wrong almost 30 per cent of the time."

That would be what Williams refers to as the two-strike bias pitch. When a batter has two strikes on him, the umpire is twice as likely to call the next pitch a strike even if it's a ball, than when the count is less than two strikes.

"I can't get into the head of the umpires," says Williams. "But what I can assume is that there's almost this bias that with two strikes, the probability that the next ball is going to be a strike probably increases in their mind."

In fairness to umpires, the rate of that particular bad call has gone down from 37 per cent in 2008 to 21 per cent in 2018.

"I think the argument to be made here is that umpires behind home plate make errors that influence the outcome on the field. And the question is how much human error should be allowed to interfere with the professional ballplayers on the field?," Williams says.

For New Britain Bees pitcher Devin Burke, he thinks there should be some.

"I made a living having the ball down in the corners. I don't throw hard enough to blow people away," said Burke before he took the mound for a game. "So I need to be able to expand the plate a little bit."

By expanding the plate, he means hitting a spot that just might be a bit out of the strike zone.

"If you can show an umpire that you can consistently hit a spot, they usually give it you, you know, nines times out of 10," Burke says.

Burke might call that the art of pitching. Tim Rosso calls it the art of umpiring.

"I think umpiring is an art," says Rosso. "When you put science into fixing art, I think that's gonna be fallible too."

Indeed. In the sixth inning, the app on his smartphone that connects to the pitch-calling computer via wifi glitched.

Rosso was able to practice his art most of that inning.

And he disagreed outright with a couple of calls made by the system.

As Rosso spoke to reporters after the game, you could sense he was conflicted.

"If this is here to stay, it's going to take a little while to get used to it," Rosso told them.

"I work at other levels of competitive baseball where the machine is not being used. And I just feel like I'm free to work. You know, I feel like there's restrictions on my job now when I'm calling balls and strikes. I'm not calling them anymore."

Meanwhile, pitcher Devin Burke adjusted pretty well to the new technology that he had feared would make his job more difficult. He gave up just two runs on six hits through seven innings as the Bees won 3-2.

Despite his strong performance, the significance of how the game was umpired was not lost on Burke.

"It makes for a different game. Whether it's better or not is something we're going to find out."

• WATCH | The National's story on how technology is being applied to baseball's officiating problems: