The VAR Room and the Discipline of "Insufficient Data": When Staying Silent Is Harder Than a Verdict

On the monitor in the VAR room, the eleventh slow-motion frame appears, and the ball is still inside a grey zone. No line is clear enough to confirm handball or no handball. The referee stands in the middle of the pitch, one hand on his earpiece, waiting for a decisive answer from the closed room. But this time, the technology gives no answer at all. The data is there, yet insufficient to conclude. And still a decision must be issued, by people, by responsibility, and by an admission that the system has reached its own limit.

Over twenty years working in and around football data rooms, I have learned that the hardest moment is not when you are wrong. It is when you do not have enough information to be right or wrong, and are still required to say something. That pressure comes from four directions: the stands want an answer, the media want a headline, the organisers want a conclusion, and the analyst himself wants to feel useful. That feeling of usefulness is the most dangerous trap.

I once saw a report rejected simply because it reached the right conclusion. In 2026, I spent six weeks logging dozens of 50/50 duels at a data centre and found a clear bias trend. The report was dismissed on the grounds that a referee's instinct matters more than statistics. Months later, when the handball rule was adjusted on the basis of exactly that kind of data, the document was pulled out of the drawer. The lesson was not that I was right, but that a conclusion is only worth something when it dares to stand on its own data.

In football analysis, insufficient data is a valid conclusion, and often the most honest one. But it is also the most costly. No one wants to read a piece that ends by admitting it cannot conclude. No one wants to hear a referee say he is not sure. So the market always rewards certainty, even empty certainty.

I see this most clearly in the metrics fans hear too often and understand too little. Take xG, expected goals. It measures the quality of a chance based on the probability that chance becomes a goal. But xG is built on a specific data sample, in a specific league. When you apply it to a match in another league, you are borrowing a ruler that may not measure what you want to measure. Then PPDA, a pressing metric, is the same: a low figure does not automatically mean good pressing, only that the team defends earlier.

The problem with football data is not that it is scarce, but that it is far too easy to fill with numbers that sound plausible. A metric quoted without its collection method is not evidence; it is decoration. I have grown used to stating sources and methods at the end of every analysis, so readers know what they are trusting.

There was a period that made me think hard about this. When leagues returned after the pandemic with empty stands, the numbers showed home-win rates falling and cards also falling. The media wrote in unison about the death of home advantage. It sounded very plausible. But on closer comparison, the cause was not that home players lost their support, but that referees lost the crowd-noise signal they used to calibrate the foul threshold. Empty stadiums do not create ghost football; they create storytellers. When the crowd is no longer roaring, people are forced to look at themselves.

That is why I always add a section few want to write: alternative hypotheses. Not to appear objective, but to remind myself that the first conclusion is never the only one. A team can lose because of tactics, fitness, refereeing, luck, or all of them combined in a ratio no one measures precisely. If I pick just one cause to tell a tidy story, I have turned analysis into propaganda.

I do not watch the match; I read its rhythm frame by frame. And I always ask: on what data did the referee base the decision, and who checks that data? The line never lies, but the person drawing it can. An offside line drawn on a screen looks very decisive, until you know that a single small calibration error, such as the average 0.43-metre error I once detected before a major final, can flip an entire conclusion. Numbers do not speak. People speak for them.

The irony is that the more technology there is, the greater the pressure to have an answer. When only human eyes existed, people accepted that a referee could misjudge. With VAR and semi-automated offside, people assume the system must be right, and when it is not, they look for someone to blame, rather than accept that there are grey zones no technology can erase. Technology only shifts responsibility; it does not erase it. It all remains human.

So when an analysis comes back empty, the most honest conclusion is not to invent content to fill it, but to say plainly that there is nothing to analyse yet. In an industry where everyone wants to hear what they want to hear, the ability to say 'I do not know' may be the highest professional quality left. If the process fails at the input stage, returning the result to be re-run is not a failure; it is a sign that the system still knows itself. And if one day a referee dares to say on the pitch that he lacks the basis to conclude, perhaps we should reward him, not suspend his whistle.