Football's Data Voids: Lessons from an Empty Column at the 2026 World Cup Quarter-Finals

In July 2026, in a small studio in Shanghai, the screen in front of me lit up with a result almost no one in the World Cup group stage had dared to bet on: South Korea beating Germany 2-0. My model, built on PPDA and back-line height, had been right, and within hours I became the most-cited name on betting forums. Thirty minutes later, when I typed the command to pull data for the quarter-final between Brazil and Belgium, the system returned an empty column. No xG. No shot count. Not a single number. Where a familiar value should have sat, around 1.87, say, there was only a silent blank, in the most literal sense.

That night I did what I had sworn fifteen years earlier I would never do: I kept writing on faith. Not on data, but on faith. I concluded Brazil would beat Belgium because of a more stable defensive base, and I said it live on air in the voice of a man who had just seen the future. Belgium won 2-1. Plenty of people who followed me lost money. I spent the next three weeks rewriting the code, but what I was really rewriting was a question: when the data does not arrive, why does a person still have to speak?

That is what I want to discuss here. Not Belgium or Brazil, but an entire football analytics industry, one where a data void is barely allowed to exist, because nobody pays for a silent spreadsheet.

The football-data industry runs on an unspoken assumption: every match can be measured. Every pass becomes a number. Every shot becomes a probability. Every pressing sequence becomes a PPDA figure. From the Premier League to the Chinese Super League, data providers sell clubs, journalists and bookmakers spreadsheets detailed to an almost unbelievable degree. A single English second-tier match can generate more than a thousand individual data points, from the coordinates of every touch to each player's running distance, minute by minute.

That system has a blind spot few are willing to face head-on: it is not designed to say I do not know. When a feed drops, when a source sits behind a paywall, when a match is captured on video rather than in text, the machine returns no warning. It returns silence. And people, standing before that silence, tend to fill it with intuition, with memory, with whatever sounds plausible.

I once watched this happen at scale. As a senior analyst for a sports platform, I published a preview before Shanghai SIPG met Shandong Luneng, using xG to predict a 3-1 scoreline while traditional pundits all picked a draw. The match finished exactly 3-1, and the piece reached fifty thousand views within twenty-four hours. But the thing I remember most is not that number. It was the feeling behind it: I immediately abandoned the series to jump into a basketball model, and my editor was furious. I did not abandon it out of boredom. I abandoned it out of fear, fear that if I stayed with that subject any longer, I would have to answer questions my data was not thick enough to answer.

There is a line I still give students whenever they ask for my secret to prediction: Every model is wrong, but a few are wrong usefully. That is not false modesty; it is a technical description. A model that is right because of good structure is useful. A model that is right by luck teaches nothing. In both cases, what decides its value is not the final result, but whether it knows what it does not know.

That is why I began treating data voids as assets rather than faults. When an empty column appears in a spreadsheet, it is not wasteland on which to build a house of imagination. It is a signal, a signal that something in the process has just broken, and that I need to find it before I say anything at all.

In 2026, when global football stopped and thousands of matches were postponed or cancelled, I understood what I call the principle of randomness: Football stopped rolling in 2026, but randomness never took a lunch break. What stopped was data. What did not stop was uncertainty. And as the data stopped flowing, people kept pouring money into models built on old data, as if a frozen spreadsheet could predict a world already turned upside down.

I was part of that mistake. In that period, some analysts, myself included, kept issuing predictions for competitions that might never kick off. The models kept running. The numbers kept jumping. But the foundation had evaporated. In essence, we were analysing the memory of a tournament whose future nobody knew.

There is a paradox that runs against the intuition of almost everyone in the trade. When you have a great deal of data, you are prone to error because you trust it. But when you have little data, or none, you are even more prone to error, because you have nothing to hold back your own intuition. Emptiness does not make you humbler. It makes you more confident, in a dangerous way, because no evidence steps forward to contradict you.

That is the blind spot the football analytics industry rarely admits. People debate the quality of xG models, the evolution of defensive metrics, the differences between one provider's data and another's. Yet almost no one discusses what happens when the data does not arrive. That void, in most newsrooms and analysis rooms, is handled with silence and a few meaningless sentences. A data-poor piece gets rewritten on intuition and phrases such as needs more time or too early to judge. That is how voids get quietly filled.

I believe in a different view: Data disappearing is not lost data, it is a kind of data. A data stream stopping is itself information. It tells you about the system, the source, the process, the limits of the method you are using. A good analyst is not only someone who can read a full spreadsheet. He must also be someone who can read an empty one.

Back to that quarter-final night in 2026. If I could do it over, I would not say Brazil would beat Belgium. I would say: the data for this match has not arrived, I have no basis for a prediction. Maybe the audience would be disappointed. Maybe the broadcaster would find someone else. But I would have kept what I lost that night: my own credibility.

Many people think silence is the mark of a weak expert. I think the opposite. Silence at the right moment, grounded in evidence, is the mark of someone who knows the limits of the tool in his hands. In an industry where everyone must speak, the one who dares not to is often the one who understands data most deeply.

Since 2026 I have noticed a new trend: systems hold ever more data, yet tolerate voids ever less. Artificial intelligence can generate thousands of lines of commentary from a match with three data points. It does so not because it understands football, but because it was trained to always answer, whether or not there is anything to say. What football analytics may need in the coming years is not more metrics, but more discipline about emptiness. A trustworthy model in the future will be one that can say my data is not enough. And if you are reading this out of curiosity about the next match, I wish you the thing I lacked at thirty-five: the calm to say I do not know yet. It will not sell many tickets, but it will keep your money longer than any model.