The Dressing Room Has No Data Columns

On my desk in Shenzhen, a forty-page document sits on top of an old train ticket. It has everything a professional analysis is supposed to have: a preamble, a risk matrix, a resource comparison table, an industry transmission diagram, a glossary, and at the very bottom a small line stating that the report must not be used for any decision whatsoever. I read it twice in one June morning. On the second pass I finally understood what bothered me: it was complete in form and empty in everything else. Every cell in the risk matrix carried the same sentence — insufficient information to assess. Sixty-seven times. I counted, because that is the habit of a man who once sat in stadiums noting down every pass. And what troubled me more than the number was another question: if a document like this were handed to a dressing room, would anyone inside be able to read it?

My job is to follow a football team. I arrive at the training ground at seven in the morning, stand by the touchline, and note who is limping, who is laughing, who has drifted off alone under a tree. For forty-nine years I have done exactly that. The new generation of reporters chases the ball; I chase what the ball has already rolled past.

But the world around us has changed. Football today runs on reports. Every match, before the ball moves, already has a document longer than the team sheet; after the whistle, three more follow. European clubs employ entire departments of twenty to forty people just to read and write this kind of text. In China, where I live, youth academies have begun teaching fifteen-year-olds how to read a heat map. I myself have grown used to receiving files titled opponent risk profile, even when the contents were simply the ten passes I had described to someone over the phone the night before.

This summer, that machinery will run faster than ever. The 2026 World Cup in the United States, Canada and Mexico expands to 48 teams and 104 matches, under a format announced by FIFA in 2026 — sixteen more teams and forty more matches than the 2026 edition in Qatar. Forty extra matches mean forty pre-match squad analyses, forty opponent dossiers, forty statistical tables. The machine will need a great deal of words to run at full capacity. And the machine will produce them, even when there is nothing to say.

The structure of an empty report is easy to recognise but hard to name, because it defends itself with its own caution. It states nothing false. It simply states nothing at all. Each section is divided into criteria, each criterion is assigned a confidence level, and when no source exists, the confidence level is marked indeterminate. The result is a text shaped like knowledge: it reads as though someone inspected twenty things and found nothing in all of them. In many cases, that is exactly what happened. In other cases — and I believe these are the majority — the writer inspected nothing at all. They simply built the frame, filled it with hedging sentences, and sent it out.

The danger lies in the fact that an empty report can still be cited. It still has a cover, an author's name, an issue date. An editor in a hurry can open it, find three lines, and build a headline. Nobody checks the risk matrix to see what is inside. And so a void gets packaged as an event.

I learned this lesson in 2026, when technology first left me behind. That March, at fifty-six, I followed a club in China's second tier through a financial crisis. The club was behind on wages, had to sell two key players mid-season, and I wrote a five-thousand-word series on the city's supporter culture, based on three hundred surveys I handed out and collected myself. Three days later, a digital platform published a training-ground story twenty minutes after the session ended. Two hundred words. It never mentioned who the club had sold, to whom, or why. My piece came back with a note saying it was too long and not engaging enough for digital.

I do not tell this story to complain. I tell it because it explains what I saw in that forty-page document. When speed becomes the only measure, people begin writing texts that look thorough in order to compensate for not having the time to be thorough. The frame becomes the only thing left.

A dressing room has no data columns. I have sat in them more times than I can recall — after wins, after defeats, after draws nobody wants to mention again. Inside, people do not talk about metrics. They talk about who forgot his ankle tape, whose child has just been admitted to hospital, which young player could not bring himself to call his mother after being substituted in the sixtieth minute. This is the kind of information that decides the next match, and it never appears in any spreadsheet, because it exists only in the humid air and menthol smell of a cramped room.

An analysis department two thousand kilometres away can model everything except that. And when a model has no data, it does not stay silent. It writes: cannot be assessed. Those three words, multiplied sixty-seven times, produce something that looks like academic honesty. I think it is closer to a framed form of silence.

In June 2026 I was in Sochi for Portugal against Spain. I had been invited to commentate for a television station after the analytical series I wrote the previous year. In the first half I mispronounced the name of the Spanish midfielder Isco three times. Three simple syllables, and I ruined all three. That night, Chinese social media reminded me of it thousands of times. The match finished three-all, Cristiano Ronaldo scored all three goals, and I sat remembering the shame in my own voice.

Isco taught me that a name is a whole person.

The following week I rewatched the qualifying footage of all thirty-two teams and noted the pronunciation of seven hundred and thirty-six players in their own local languages. I built a small database by hand. It had only one column: the name, and the way people say that name at home. But from then on I understood something no statistical table had taught me. Accuracy, in the end, is a way of showing respect. A dataset that gets a name wrong is not technically wrong. It is wrong on a human level.

In September 2026, stadiums had been empty for seven months. I went into the dressing-room area of a club in southern China after a training session and found a twenty-one-year-old midfielder sitting alone on a bench, his captain's armband still on his sleeve. He said something I have kept verbatim in my notebook: Uncle, with no crowd, I don't know who I'm playing for.

I did not write the line down straight away. I sat with him for a long while, then suggested he write an open letter to the supporters himself. It took him four days. When the piece ran, it was shared fifty thousand times. No data model predicted that, because the input did not exist until someone handed him a pen.

The silent dressing room that day still carried the smell of grass and someone's tears.

In December 2026, while the world was fixed on a World Cup held in winter, that player's agent told me his client would be loaned to a Dutch club in the winter window. I held the information for three weeks, waited for the contract to be signed, and only then published. The article reached two million reads.

A group of supporters immediately criticised me for staying silent so long. They had a point in their own terms: fans pay to know, and three weeks is three weeks spent on the outside. I hosted a two-hour online Q&A and prepared no answers in advance. What I learned that night was this: an article has not only readers, it has people affected by it.

Since then, whenever I sit down with a file of information, I ask myself the same four questions. How will this player feel reading the first line? What about his family? What about the supporters? And what happens to trust if I am wrong? Those four questions appear in no risk matrix I have ever seen, though together they matter more than every cell inside one.

I have also seen the reverse of the problem, in another market. A Gulf league spent hundreds of millions of dollars over several years bringing in players past their peak. On paper, that is progress for a football nation. In practice, I watched those matches and saw something quite different: the player arrives, poses for photographs, signs commercial deals, and plays roughly fifteen matches a season. What gets bought there is image, not football rhythm. And this is where I want to stop, because it connects directly to the story about reports.

When a thing is valued by its name rather than by what it produces on the pitch, the reports learn to follow. They become beautiful in structure, correct in terminology, and empty in substance. A report with nothing to say is like a contract with no one to play.

The counter-intuitive point here is this: an empty report does less harm than a confident one. An empty report, at least, admits its limits on the final line. It says it does not know. In an industry where everyone is paid to sound certain, saying you do not know carries a value of its own. The problem is that the system does not allow that honesty to exist in that form. It forces the writer to package not-knowing into a matrix, to file not-knowing into cells, to name not-knowing with terms that sound scientific. The result is a document that appears more trustworthy than it deserves, and that is the real danger.

A document willing to write we do not yet know, and here are three things we will go and watch over the next two weeks, is more useful than any matrix. The problem is also not that data analysts have entered football. The problem is that dressing rooms are required to answer in their language, when that language has no word for a twenty-one-year-old sitting alone at nine in the evening.

When new media knocks loudly at the door, I still hear the old drums from the stands.

In the coming weeks, as forty-eight national teams converge on North American pitches, thousands of pages of analysis will be produced every day. I will read some of it, but I will still be at the training ground at seven in the morning. I go to the ground not to score, but to keep the rhythm for the stories. The question I carry into this summer is simple: among all those pages, is there one line willing to say that its author does not yet know what happens next?