Body-Language Experts, Mislabeled Data and Football's Disease of Reading Minds

Inside an 80-second clip, Elizabeth Holmes sits very upright, her chin slightly raised, her eyes open a little wider than usual. A body-language expert counted six blinks in a single minute, against the 15 to 20 that an ordinary person performs while talking. The expert concluded that the woman in the frame was pressuring whoever sat across from her, and that her sincerity was suspect. That is the opening of a documentary titled You Can See Everything, scheduled for release on the 16th of October, after being introduced at a prestigious film festival.

That night, an automatically generated data file rolled onto my screen under the label: Football.

I sat in Guangzhou, read the label three times, then read the entire passage below it — barely two hundred words — three more times. Not a team, not a player, not a coach, not a match, not a scoreline. Only Holmes, the two directors Nathan Fielder and Lance Oppenheim, three people introduced as behavioural and psychological analysts, and one blink-rate figure. I closed the laptop, brewed a pot of tea, and understood that a certain April night had returned to me in a different shape.

In 2026 I was seventeen, fresh out of an academy after a knee injury, curled up in front of a screen rewatching Monaco beat Dortmund 3-1 in the Champions League. I logged every touch of Mbappé's into a notebook: 34 touches, six maximum sprints, one goal, one assist. I cross-checked expected-goal figures, running distance, sprint rhythm, and then held the draft three days before publishing it. That eight-thousand-word piece began to circulate, and my career began there. Data does not lie, but crowds do. That is the line I tell myself every time a number gets bent to fit someone's eye.

But tonight, the thing that lied was the label.

I am used to excavating fragments buried under the dust of youth football. I am used to an academy article filed under the wrong section, a player profile tagged with the wrong position, an eighteen-year-old's contract recorded as nineteen. Those small distortions, added together, form a skewed picture that an entire industry leans on. When the pitch falls silent, memory begins to dig. And this dig led me to a question far larger than a single classification error.

The question is this: why is modern football addicted to reading human beings?

Let me start from the beginning, because this story begins far from any pitch, and only then comes home.

A football data file does not create itself. It flows through a long chain of small machines: a site's harvesting system, an automatic tagger, a content classifier, a human editor at the end of the line. Each link can fail, and most failures are never discovered because they drift past too quickly. The publication that owns the article once ran a great deal of football coverage, employed correspondents following European leagues, and maintained a steady sports section. When a tagger learns at the level of the site rather than the level of the article, it will label anything appearing near a football piece as "Football". I have no absolute proof of the mechanism, but the probability is high.

And so a story about Holmes, about a collapsed biotech company, about a fraud conviction, slipped into the database I use to analyse eighteen-year-old players.

The confusion sounds harmless. One bad record, delete it, done. But it points to a larger sickness: football's ecosystem runs on streams of data that very few people check at the source. We trust the number, trust the label, trust the headline, forgetting that behind every number is a person — or a machine — that decided how to name it.

I began digging into the very mechanism that produced this error. And what I found made me think more about football than about Holmes.

Look at the blink figure. Six times a minute, against a norm of 15 to 20. Three experts — a body-language analyst, a clinical psychologist, a behaviour analyst — read the same clip and delivered a judgement about sincerity. It sounds scientific. There is a number, there are experts, there is a conclusion. But when I peeled back each layer, the structure was so familiar it chilled me.

Because it is exactly the structure football uses every day.

I have spent years watching matches at academy and first-team level, and I have lost count of how often I have heard people talk about a player's "body language". A striker who does not celebrate a goal — "he is losing spirit". A defender who lowers his head after being substituted — "he has an attitude problem". A coach crossing his arms on the bench — "he has lost the dressing room". A young player who does not look straight into the camera — "he lacks character". Those lines get spoken on broadcasts, written into articles, shared thousands of times.

And like the blink figure, they all rest on a sample far too small, with no baseline, no cross-check, no verification.

This is the core point: football has turned the reading of human beings into an industry, while the only trustworthy thing lies in numbers nobody bothers to check. We believe a player's eyes more than his chances created over three seasons. We believe a coach's crossed arms more than his team's underlying metrics. We believe an 80-second clip more than a forty-page document.

I wonder: if Holmes appeared on a pitch, with that same blink, that same raised chin, how many experts would use it to draw conclusions about her career?

Many. I have seen it.

Last season I tracked a nineteen-year-old midfielder at a southern academy. He played well across four straight matches: 2.4 chances created per game on average, a passing accuracy above 88 per cent in the opposition half, three ball recoveries per game in midfield. But in one match he was substituted on 70 minutes and sat on the bench with a blank face. The next day an article appeared: "His attitude is worrying."

I called an acquaintance who works in data at the club. It turned out that the night before the match, the boy had stayed up until two in the morning caring for his grandmother in hospital. Nobody checked. Nobody asked. The label "attitude problem" had been applied, and it would follow him for a long time.

That is how a small sample becomes a judgement about a person.

And that is why the mislabelled "Football" tag in that data file was not harmless at all. It was merely the technical version of the same disease: a machine saw something, and named it without needing to know what it was looking at.

I spent three months of 2026, when the pandemic closed every academy, building a database of 1,200 young players from five major European leagues between 2026 and 2026. I compared youth-team minutes with first-team appearances after the age of twenty-one. The result: players who had suffered a break of more than six months were 27 per cent less likely to reach fifty professional appearances than the rest.

That figure is not a prophecy. It is a correlation. But it is a correlation cross-checked across thousands of hours, not a judgement made from a single clip.

The difference between the two is my entire profession.

I do not watch a match, I excavate it. I peel back each layer: contracts, trial sessions, childhood promises, physical curves, injury diaries. Every contract is a geological layer. To know why a player stands where he stands, I must read from the bottom up, not from the moment he celebrates or lowers his head.

Meanwhile, most of the public reads from the top down. They see the moment, and they rewrite history from the moment.

That is why the Holmes story matters to a football man like me. Not because of Holmes. But because of the way people handle her — and the way they handle a young player when he blinks out of rhythm.

Let me say plainly what many do not want to hear: most body-language analysis we consume every day, in football or outside it, carries no more scientific value than a rumour. It has a number, an expert, a screen — three things that manufacture a false sense of certainty. But it lacks the one decisive thing: a long-term baseline.

The blink figure of six against 15 to 20 means something only if we know how often that person normally blinks when not being watched. If Holmes was always a low-blinker from the age of ten, then six says nothing about sincerity. If a player has lowered his head while concentrating since childhood, then that lowered head says nothing about cowardice.

Without a baseline, every judgement is a guess disguised in the language of science.

And football is a gold mine for this kind of disguise.

Think of how we talk about players from places nobody watches. They played three peak seasons in the dark, scored goals nobody recorded, ran distances nobody measured. Then in one big match they explode. The crowd says: "Talent arrived out of nowhere." But talent does not arrive out of nowhere. It is simply excavated late. People see the moment and imagine the moment created the person. The truth is the person was there for years, under the dust of time.

Mbappé did not appear in a single night, he was dug up over many nights. The same logic applies to Holmes. She appeared to the public as a phenomenon, but the phenomenon was built long before the 80-second clip the experts are now examining.

The problem is that in both cases, people prefer the moment to the geological layer.

Let me tell another story. When I was in the 2026-born academy cohort, a boy joined a year below me. He did not blink much when speaking. He did not smile much. He did not look coaches in the eye. Many in the squad thought him arrogant, cold, disconnected. He was pushed off the main list. Three years later I happened to read an interview: he had a condition affecting his ability to make eye contact. Nobody at the academy had ever asked. They simply looked, and named.

That label followed him out of the academy, into a lower-division club, and then off the map.

I tell this not to appeal for sympathy. I tell it to show that the same mechanism that labelled a Holmes article "Football" is the mechanism that labels a sixteen-year-old "attitude problem". Both are consequences of naming too quickly, too confidently, without checking.

The crowd looks toward the light; I look down at the soil beneath.

Tonight, the soil told me that even the machines make the same mistake as the humans.

When I checked my entire football data stream that week — roughly four hundred records — I found eleven similar cases. An article about a former star's personal finances tagged "transfer". An esports piece tagged "national league". A film-festival piece tagged "youth player". Eleven out of four hundred, close to three per cent. It sounds small. But if an analytics system uses this data to compute, three per cent of noise is enough to skew the finest conclusions. And in football, the finest conclusions are the ones that decide a contract.

I once wrote that I do not fear bad numbers. I fear bad numbers presented as good ones.

That is why I held my draft three days before publishing, back when I started writing. I re-checked every figure, every source, every date. That habit made me slow. It also made me right. And in an industry that treats speed as a virtue, slowness is an act of resistance.

Back to Holmes. You Can See Everything is not an independent investigation. It is a media product, launched precisely when a film festival needs attention, using the crowd's emotion as fuel. The experts in the film offer opinions, not evidence. An 80-second clip is not enough to conclude anything about anyone's sincerity. But it is enough to create a frenzy.

And that frenzy has a structure identical to a transfer frenzy.

I have tracked hundreds of transfer rumours. They share one formula: a vague source, a shocking moment, a spreading crowd, and a conclusion that arrives before the truth does. When the truth arrives, it is slower, duller, and less shared. The false always runs faster than the true. That is the rule.

But one thing I have learned after years of going against the crowd: going against does not mean going wrong. Sometimes going against simply means going slower, and thereby seeing more.

I do not deny that intuition has a place in football. A good scout can feel something about a player before the number appears. But that intuition must be cross-checked, recorded, verified, not presented as truth. The difference between a humble intuition and an arrogant judgement lies here: the humble intuitive always leaves open the possibility of being wrong.

The experts in that clip did not leave it open. They did not say "perhaps". They spoke as if they knew.

That is the moment an analysis becomes a verdict.

And football has handed down many such verdicts to young people, on the strength of a single look.

I think of my 1,200-player database. In it are names I have seen undervalued, called "no quality", "lacking hunger", "not a cultural fit". Those labels were usually applied from a few short observations, like an 80-second clip. A few years later, some of them exploded, and people called it "talent arriving out of nowhere". But my data shows they were there from the start. It is just that the readers did not read carefully.

This is my contrarian take on the whole story: the guilty party is not the machine that mislabelled. The machine only learned from us. It learned that speed matters more than accuracy, that naming matters more than checking, that sharing matters more than understanding. We taught it that, and then were surprised when it did exactly as taught.

If the crowd stopped rewarding fast, confident judgements, the bad labels would disappear. If the crowd started asking "based on what", the body-language experts would have to answer more seriously. The machine is not at fault. Those who feed it are.

I know this sounds like blaming the audience. But I believe it. In ten years of observing the industry, I have never seen a trend persist without the demand behind it. The frenzy of reading people exists because we crave knowing what others think. We want to believe that looking is understanding. It is a comfortable illusion, and it sells a great deal of advertising.

As for me, I choose the harder path: I read every figure, I hold the draft, I cross-check three seasons before concluding anything about a person. I am not perfect. I still err. But I record my errors, and next time I correct them.

That is the only thing separating an archaeologist from a treasure hunter.

The archaeologist digs slowly, takes notes, and accepts that the soil beneath may refute what he believed. The treasure hunter only wants to find gold, and will call anything shiny gold.

Modern football is full of treasure hunters.

That night, after closing the data file, I sat a long while before the screen. I thought of Holmes, of the six blinks, of the experts, of the 80-second clip. I thought of the nineteen-year-old midfielder who stayed up caring for his grandmother, of the boy who could not look a coach in the eye, of everyone named too quickly.

And I thought about how my data called Holmes a footballer for exactly one moment.

A small moment. But it reminded me that even the systems that seem most objective can fall into the old trap: seeing something, and naming it without understanding what it is.

If a machine can call a biotech fraudster football, then a machine can also call a real talent a failure. And if we keep trusting labels without digging beneath, we are building an entire football on geological layers that have never been checked.

A question I leave for myself, and for anyone who reads this far: of all the judgements you believed this week — about a player, a coach, a person — how many did you actually verify, and how many did you simply name?

I have no answer for you. I have only a promise: tonight, I keep digging.