How MLB, FIFA and Wimbledon decide who gets to overrule the camera

It was Sept. 30, the 10th inning, with the Phillies leading 4-3 and two Braves on base. The umpire called an 0-2 pitch a ball, and catcher J.T. Realmuto responded by tapping his helmet for a challenge. Camera tracking showed the pitch had clipped the strike zone. Strike three, and the game was over.

The same kind of optical tracking calls the lines at Wimbledon and supports offside decisions at the World Cup. Yet the three sports gave the machine very different powers, and a Nov. 5 analysis in Global Leaders Outlook argues the difference sits in the rulebook, not the hardware. Baseball rations appeals. Soccer pairs tracking with video officials. Tennis handed line calls to an automated system almost entirely.

That matters well beyond the stadium. Any operations or IT leader putting an automated system in charge of approvals, inspections or exceptions is writing a rulebook too, whether they call it one or not.

A sensor reads the touch, a referee reads intent

The World Cup offered a clean example of where measurement ends. Croatia's Igor Matanović got a slight touch on the ball, and the sensor inside it picked that up. The touch left a teammate offside and wiped out a late equalizer against Portugal. That part was data. The next part was not: the ball then deflected off a Portuguese defender, and officials had to judge whether his touch was deliberate, since only a deliberate play resets offside. They decided it was not.

One decision therefore contained two questions. Where was the ball, and when did it get touched? A sensor can settle that. What was the defender trying to do? That still needs a person. The Global Leaders Outlook piece argues that once measurement is routine, the real design question shifts to authority: who is allowed to ask for a check, who rules in the end, and how much of the process viewers get to watch. Every answer places a cost on someone, which explains why similar technology yields such different contests.

For anyone scoping an automation project, the useful move is to split the decision the same way before buying anything. The parts that are pure measurement can go to the system. The parts that turn on intent or context need a named human owner and a written test for when they step in.

Baseball's two-challenge budget

Major League Baseball kept its umpires and layered a limited appeal on top. Any batter, pitcher or catcher can dispute a ball-or-strike ruling. Each team begins with two challenges and holds on to one every time it wins. That scarcity makes the challenge a resource players must manage, and the timing data cited by Forbes reflects it.

The overturn rate is the number most likely to be misread. Because players choose which pitches to contest, 52.2% tells you how good they are at spotting bad calls. It doesn't tell you how often umpires miss. Any business that measures an automated or human process through its appeals has the same problem: the appeal rate and the win rate describe who complained and how well they picked, while the error rate across all decisions stays out of view.

MLB could have gone further. Its minor-league tests of full automation produced more walks and longer games, and the league said full automation would have ended pitch framing, the catcher's craft of presenting a borderline pitch as a strike. It chose the challenge model even though 69% of fans surveyed at spring training favored full automation. That is the strongest case against the league's design, and it comes from its own customers. The other side of the ledger: 72% of fans in the same survey said challenges improved their experience, though MLB ran that survey itself.

Soccer's speed and tennis's off switch

Soccer rarely pauses, so FIFA built for speed. Clear offsides are sent directly to the on-pitch officials with no video referee review, humans retain the final call, and 3D avatars help viewers follow replays. Showing the work also exposes the gaps. Before a penalty in Switzerland's match against Qatar, an outage prevented the offside graphic from being produced. FIFA said the lines drawn by its video officials showed no offside, but ESPN reported that the images came out only about 4½ hours later. The decision may well have been correct. The explanation took most of an evening to show up.

Wimbledon went the furthest. At its 2025 Championships, Hawk-Eye Live followed the ball with 18 cameras and handled all line calling. During a fourth-round match, an operator mistakenly turned the system off for one game. Three calls were never registered, and the chair umpire called for the point to be replayed. Wimbledon apologized, attributed the lapse to human error and took away operators' ability to turn tracking off.

Wimbledon swapped about 300 line judges for roughly 80 match assistants, and human error moved with them, to whoever held the off switch.

Review systems also change how the people being reviewed behave. A 2024 proof-of-concept study in BMC Sports Science, Medicine and Rehabilitation compared 1,838 Turkish Super League matches played with and without video review. Fouls fell for both home and away teams after it arrived, a medium-sized effect. Home advantage and the card patterns against away teams did not change significantly. The study covers one league and can't prove why fouls fell. Still, it suggests the most measurable effect of oversight may show up in conduct rather than in the fairness numbers the system was bought to fix.

Writing the fallback before the outage

Every one of these systems will eventually fail mid-game, and the sports that handle it best decided in advance what happens. MLB's published rules cover the untracked pitch: the original call stands and the team keeps its challenge. Nobody improvises, and nobody loses a scarce resource because the cameras blinked. Global Leaders Outlook's conclusion is that technology moves risk to the people who operate and govern the system rather than eliminating it.

The NFL showed the human version of the same principle. The league said it adjusted its Week 10 officiating assignments because of the nationwide impact on air travel after the Federal Aviation Administration's ground stop on Friday, moving officials to games within driving distance of their hometowns, NFL.com reported. Because of the changes, three backup referees will work as referees on Sunday. Side judge Don Willard was set to work two games in one day, Dolphins-Colts in Indianapolis and Steelers-Bengals in Cincinnati, roughly 110 miles apart. According to the league, each of its 17 crews has a named backup referee who trains all year for exactly this. The fallback was a specific, trained person, chosen before anyone needed one.

For teams putting automated decisions into credit, quality, scheduling or customer service workflows, the sports give a usable spec in four questions. Who can request a review, and is that right rationed or open? Who has the final say when the system and a person disagree? How much of the reasoning can the affected customer see, and how fast? And what happens to a decision, and to the reviewer's appeal rights, when the system goes down? These matter most for operators whose decisions mix pure measurement with judgment calls, the Matanović problem. A system that only checks whether a value falls inside a limit can be run more like Wimbledon's.

Realmuto's challenge counted as a decision because baseball requires someone to choose to ask. The camera only measured the pitch. The rule decided when that measurement would end a game.