Mislabelled Tags and False Signals: Why the Transfer Window Is Football Data's Harshest Test

Every transfer window begins the same way. It starts with a status update fewer than twenty words long.

This year was no different. An account with seven hundred thousand followers posted: "Deal done, medical this week." No player's name. No club. No number. Just a verb in the past tense and a belief pre-installed in the reader.

Three hours later, the post had been shared fourteen thousand times. Eight hours later, three major sports outlets were running it. Two days later, the agent denied everything — but by then it was too late. The name was on the list, the shirt had been edited, and a three-thousand-word tactical breakdown of where the player would line up had already gone out.

I sat and read the whole sequence back, from the first status update to the final bulletin. What stopped me was the tag, not the content.

Today I want to tell a story about mislabelled tags. About false signals quietly entering football's analytical systems. And about something I have believed for nine years in this job: a correct number placed in the wrong slot does more damage than an outright lie.

The Journey of a Piece of Information

To understand why transfer rumours have such vitality, you have to understand their route. A piece of transfer information passes through at least five stages.

The first is the agent, whose motives are clear: inflate the price, create pressure, or simply keep a client in the media's eyeline. The second is the journalist, who receives the fragment and needs a source to hang an article on. The third is the aggregator, where the information is trimmed into a headline. The fourth is the fan, who turns it into emotion. The fifth is the data model — the stat tables, the indices, the player-evaluation tools — where a player suddenly gets assigned to a club that never signed him.

At each stage, the information loses a little accuracy and gains a little belief. By the final stage, the tag has replaced the content. Nobody asks what the original fragment said. They only ask what it was labelled as.

In data analysis, this phenomenon has a dry name: a labelling error. It happens when a dataset is assigned to the wrong subject. The consequence is not a small margin of error. The consequence is that every conclusion downstream becomes rubbish — while still wearing the shape of a serious conclusion.

I once saw this at the smallest possible scale. In 2026, aged sixteen, I worked as a volunteer statistician at a national U19 match in Ho Chi Minh City. The game was U19 Hanoi against U19 SHB Da Nang. I was assigned to log forty-seven fouls and twelve offsides. In the 78th minute, I spotted a foul in the box that the referee had missed, leading to a controversial goal.

I built a comparison table against the IFAB Laws, noting the minute, the position, the running direction of both players, and sent it to the organisers. Nobody replied.

At the time I thought the problem was that the organisers hadn't read it. Years later I understood the problem lay elsewhere: my data table had been mislabelled. I sent it as a technical stat sheet, but the recipient read it as the complaint of a sixteen-year-old boy. Same content, two tags, two entirely different fates.

When a Wrong Tag Breaks an Entire System

In professional football, labelling errors are everywhere; they are simply rarely called by their real name.

The first type is a position error. A winger is recorded in a striker role, and suddenly he is judged inefficient because his goal count is low. A defensive midfielder is tagged as a deep-lying centre-back and compared with players operating in a completely different job.

This is where I want to state a clear view on a trend that is flattening football. The inverted winger is becoming the default. Every young wide forward is coached to cut inside, shoot with his weaker foot, and turn into a false striker. Arjen Robben is the perfect model of this archetype, and his success has made a whole generation get judged through the same lens.

Traditional wingers — the ones who hold the touchline, who stretch the pitch horizontally, who live on crosses as David Beckham once did — are being undervalued in a distorted way. They are undervalued not because they are worse, but because the metrics used to measure them were designed for a different role.

A traditional winger can lead the league in chances created from wide areas, but if you measure him by goals and successful dribbles, he will look ordinary. The tag has already decided the verdict.

The second type is a competition-level error. A player's numbers in a second division cannot be placed beside a top-flight player's numbers without an adjustment coefficient. Yet in the comparison tables published daily, the two columns sit side by side, with no note, no warning.

The third type is a sample error. This is the most dangerous, because it is the most subtle.

In 2026, I joined a research project on how empty stadiums affect match outcomes. I gathered data from eighteen rounds and found that the home win rate had fallen from forty-two percent to thirty-one percent. I wrote the report excitedly. My lecturer read it, then asked a single question: have you compared it with the previous five seasons?

I hadn't. When I did, the picture changed completely. The drop was not large enough to draw a conclusion, because the sample was too small and squad quality dominated. I nearly published a wrong conclusion based on a correct sample.

Since then I have written a rule for myself: never use a single season to assert a trend. I don't believe in luck. I believe in a number repeated a hundred times.

The fourth type, and the one I meet most often in daily work, is the camera-angle error.

Camera Angles and the Reference Point

Every slow-motion replay carries its own truth. My job is to find the truth that cannot be argued with. But "cannot be argued with" is a relative concept in a sport recorded by dozens of cameras at dozens of positions.

Semi-automated offside technology has changed the game. It rebuilds the line using motion data from every joint of every player, and in most cases it produces an answer so precise it is hard to dispute. But it still depends on a reference point set by a human: the instant the ball leaves the passer's foot. A frame off by a few hundredths of a second can turn a valid goal into an offside, or the reverse.

That reference point is a tag. And that tag is applied by a human being.

This is why I always tell colleagues: the referee's decision is only the endpoint. The real journey lies in each angle. If you look only at the endpoint, you understand nothing. You are merely agreeing with or objecting to a predetermined result.

The Things That Don't Happen

There is another kind of data the naked eye never sees, and I spend most of my time reading it: data about the things that don't happen.

A striker who never receives the pass does not appear in the assists table. A defender who drags the whole back line right to open space on the left gets no metric at all. A run that pulls the opposing centre-back out of position to open a gap for a teammate gets credited by the scoring system to the scorer, not the creator.

The moment the naked eye misses, the data never forgets. But only when you know how to ask the data the right question.

This is the boundary between analysis and guesswork. The good analyst is not the one with the most data. The good analyst is the one who knows which data not to trust.

In the transfer window, this pressure is higher than at any other point in the year. Clubs must commit tens of millions of euros based on datasets assembled from different sources, different leagues, different ways of tagging. A small deviation in position labelling can turn a five-million player into a fifty-million failed signing.

I have one non-negotiable rule: one number, one translation.

Before or after every number I publish, I must include a sentence explaining where it came from, how big the sample was, what the context was, and how it might be misread. If I can't write that sentence, I drop the number.

The Non-Numeric Grey Zone

There is a paradox I have to admit, and it makes my job more complicated than its cold exterior suggests.

Not everything in football is measurable. A fractured dressing room does not show up in the league table. The confidence of a striker in crisis is not inside his expected-goals figure. A centre-back half a step slower because of a dull calf ache keeps a clean tackling metric until the first serious mistake.

I devote at least one paragraph in every analysis to that grey zone. Not to appear humble, but because dropping it is lying.

The Analyst in the Dressing Room

Over the past decade or more, data departments have become a formal part of professional football. Big clubs hire dozens of specialists, and data models now inform everything from recruitment to tactical choice.

This is real progress. But it comes with a risk few mention: analysts' conclusions are often detached from the actual rhythm of the match.

A model can say that full-back X is the best transfer target, based on twelve metrics. But that model was not in the dressing room at half-time, did not hear the captain shouting, does not know that this player needs a full-back beside him who knows how to cover. The model is right statistically and wrong humanly.

I am not saying data should leave the dressing room. I am saying the person reading the data must walk into the dressing room before delivering a conclusion.

Contract Structure: Where the Truth Lies

When a deal is announced, the most-quoted number is the transfer fee. It is usually the least meaningful number.

Two deals with the same nominal fee can be completely different in substance. The first consists of a small upfront payment, with the rest made up of variables tied to appearances, goals, and European qualification. The second is a four-year contract with a rising salary and an unusually low release clause. In accounting terms, the two deals are entirely different. In the headline, they are identical.

The structure of the release clause and the wage bill is the real story. The transfer fee is only the tag.

Youth Development and the Tag Called Results

At youth level, the most dangerous tag is "results". A U18 coach is judged on wins, so he picks physically strong players, a safe style, and short-term outcomes. Technical players who are late to develop physically get cut from the squad, and many of them are never seen again.

The physicalisation of the U18 age group is eroding football's technical soil. It does not happen because anyone is cruel. It happens because the tag "results" is applied to a developmental phase where results should not be the main yardstick.

This is where data can help, if used correctly. Instead of measuring a U18 team's win count, measure how many of its players are in the first team five years later. But the second measure takes five years to pay out, while the first pays out in ninety minutes. People always choose the faster payout.

Social Metrics and the Illusion of Importance

There is another kind of mislabelling I meet every day when reading the news: engagement used as a measure of quality.

An article with a million views does not mean a million people understood it. A player whose follower count spikes after a goal has not necessarily improved correspondingly. Social media measures spread, not accuracy.

The problem arises when those numbers are used as the basis for real decisions. A club choosing a player for his following adds yet another deviation to the labelling chain. By now the data has passed through too many hands, each adding a tag, and nobody remembers the original fragment.

Verify First, Speak Second

I have a habit many colleagues consider slow: I never comment on an incident before watching at least four angles. Four is not an arbitrary number. It is the minimum number of angles for an incident to appear in enough dimensions, to eliminate the illusions created by a single camera.

In an era when everyone wants to respond in three seconds, waiting another three minutes is a competitive disadvantage. But I have learned that disadvantage exists only in the short term. Over the long term, the person who speaks correctly slowly always beats the person who speaks wrongly fast.

Emotion Is Not Evidence

In every major controversy, the most predictable thing is the reaction, and the least predictable thing is the evidence.

When a controversial decision is made, within five minutes social media already holds thousands of verdicts. Within five days, the number of people who actually rewatch every angle can be counted on one hand.

The crowd is not short on passion. The crowd is short on the patience to verify. And that impatience, repeated often enough, becomes a form of fake data — a signal that isn't real but spreads faster than the truth.

This is the counter-intuitive point I want to stress: in football, crowd emotion is not evidence. It is just another data point, to be cross-checked like any other.

If an offside decision is given and the whole stadium objects, that does not prove the decision wrong. If the whole stadium stays silent, that does not prove the decision right. Noise only indicates the level of interest, not the level of accuracy.

The empty stadium taught me this: noise never scores.

But I also have to say something many in the profession avoid. Absolute consistency is an ideal, not a reality. Two fouls identical in form can differ in context: speed, angle of approach, degree of danger, match situation. A referee who penalises a similar incident in the tenth minute and lets one go in the ninetieth is not necessarily biased. That can be game management.

This is the grey zone of game management that pure data never fully captures. Admitting it does not weaken the analysis. On the contrary, it makes the analysis more honest.

What I Want to Leave Behind

The story of that twenty-word status update at the start is not a story about a rumour. It is a story about how we process information in a transfer window ruled by speed.

What I want to leave behind is not advice to avoid the news. It is a small habit: every time you read a number, ask where it came from, what it was tagged as, and whether the conclusion behind it would still stand if that tag were wrong.

Football does not change because you look at it more closely. Football changes because you look at it more correctly.

And if you had to choose between a compelling story told in three seconds and a dry truth that takes three days to verify, which would you choose?