Ninety Minutes of "Correct" Decisions

A brief departure from the usual territory of cybersecurity and digital privacy. Football has been on my mind lately (as always), specifically what VAR and semi-automated officiating reveal about the same tensions we debate in AI governance every day. Who gets to define objectivity? What do we lose when we hand decisions to algorithms? These questions play out every weekend in La Liga, the Premier League, and across European competitions, in real time, in front of millions of people.

The Promise of Perfect Decisions

VAR arrived with genuine good intentions. After decades of controversy, the technology was supposed to deliver something football had always lacked: a kind of truth. Cameras would see what human eyes missed.

It has not worked out that cleanly. What VAR actually did was shift the terrain of disagreement. We still argue constantly, just about different things: whether the technology made the right call, whether “right” is even the correct frame, whether a pixel-level offside reflects the spirit of what the rule was ever meant to enforce.

When Systems Fail

The 2024-25 season offered a strange catalogue of failures. In August 2025, a La Liga match between Barcelona and Rayo Vallecano had to continue without VAR after the system crashed entirely. A potentially decisive penalty could not be reviewed. The match went on. The questions did not.

Semi-Automated Offside Technology, the system rolled out as VAR’s successor, tracked 29 body points per player during Euro 2024 with impressive accuracy. Still, questions about interference with active play required lengthy human deliberation anyway. Millimeter precision at one end, subjective interpretation at the other.

The Premier League’s Key Match Incidents panel identified 18 VAR errors across the 2024-25 season. Long review periods kept draining stadiums of atmosphere. Matches stalled. The tool designed to reduce uncertainty became its own source of it.

Flick’s High Line

One thing the new technology has genuinely changed: what is tactically possible.

When Hansi Flick took over Barcelona, he implemented a defensive line that pushed further forward than almost any top-flight team in recent memory. The Blaugrana averaged 34.7 meters from their own goal, highest in La Liga. That figure is not accidental. It is the direct consequence of trusting SAOT to catch infractions that no linesman’s eye could reliably flag.

The logic holds up. If millimeter-accurate offside detection is guaranteed, a coordinated defensive line can push to limits that would have been reckless a few years ago. In Flick’s first 50 league matches, Barcelona caught opponents offside 181 times, and even more across all competitions. Flick reportedly used giant-screen video drills in training, giving defenders real-time visual feedback on their synchronization. Iñigo Martínez, Pau Cubarsí, Alejandro Balde, Jules Koundé: four defenders using the technology’s certainty as an actual weapon.

This kind of tactical evolution was simply unavailable before SAOT. The high defensive line has always existed in football theory. What changed is that marginal offside calls are now predictable enough to build an entire season around them.

Something Has Been Lost

Watch a stadium when a goal goes in. There is a pause now. Fans look up at the screen before they let themselves celebrate. The spontaneous eruption has become conditional. We cheer provisional goals.

That hesitation is a relatively new thing, and its cumulative effect on what football feels like is real, even if it is hard to quantify. Decisions that were once absorbed as part of the game’s texture, part of the injustice that made eventual justice sweeter, are now subject to frame-by-frame review. The game becomes correct. Whether it remains compelling is a separate question.

VAR 3.0

The roadmap keeps extending. “VAR 3.0” in its current conception involves AI-assisted decisions delivered in seconds rather than minutes. Smart balls already feed data directly to officiating systems. Some technologists genuinely believe fully automated refereeing is achievable within a couple of decades.

The harder question is whether the technology can do what actually matters. Can an algorithm understand intent? A cynical foul versus a clumsy challenge? The unwritten negotiations between a referee and players that keep a heated match from escalating? The advantage played because the spirit of the game demanded it? These are not edge cases. They are the game.

What Precision Costs

There is a tension at the center of the VAR project that sharper technology will not resolve. Football has always carried a degree of controlled chaos. The referee’s error was part of the story, something supporters argued about for decades, something that shaped how a club’s history got written.

The dream of perfect officiating rests on the assumption that eliminating human error is straightforwardly good. Maybe some of those errors served a function we did not appreciate. Maybe the unpredictability of certain calls made wins feel earned and losses feel meaningful in ways that a technically correct outcome does not. I am genuinely not sure. I do not think anyone is.

What I do know is that the same debates happening in AI ethics circles, about accountability, interpretability, and what gets lost when judgment is automated, are being lived out by football fans every week. The technology is already here. The reckoning about what it means is still catching up.