Football Data and the Name-Collision Disease: When the System Has No Idea Who It Is Talking About

Football Data and the Name-Collision Disease: When the System Has No Idea Who It Is Talking About

Opening

In 2026, in a meeting room in Chengdu, I opened a scouting file sent over by a European data company. Among hundreds of rows, one stopped me: “Jordan — midfielder, 28 years old.” Where the club should have been listed, they had left a blank.

What made me pause was not that the player was unknown. It was that the system did not know who it was talking about. In the world’s football databases, “Jordan” shows up at an alarming frequency: Jordan Pickford, the Everton and England goalkeeper; Jordan Henderson, the midfielder who wore Liverpool red; Jordan Ayew, the Ghanaian forward; and the Jordan national team itself. Four different entities sitting on one string of characters, in one classification column.

An algorithm that matches strings will not tell them apart. When a news site writes about “Jordan” without a surname, the system mislabels it, then multiplies that error through every layer of processing. That is the first of three errors that are slowly eroding football’s information flow from within.

Context

The football data industry has exploded over the last fifteen years. From companies like Opta and StatsBomb to countless scouting startups, every match now generates thousands of data points: passes, expected goals, pressing intensity, distance covered. Clubs pay millions of dollars a year to buy those data points, and media outlets sell them back to audiences as analysis.

Running alongside that official data stream is another, far noisier one: transfer rumours. Every transfer window, hundreds of stories appear each day, most of them sourceless, most of them unverified. And now, both streams are pushed through the same machine: aggregation algorithms, automated accounts, and language models rewriting the news at a pace no newsroom can keep up with.

At first I thought the problem was speed. Then I realised it ran deeper: these systems exchange data that no one in the chain is able to verify. One number goes in, three articles come out, and the identity of the subject is distorted from the very first line.

Before 2026 I watched football with my eyes. After 2026, I watched it through numbers that know how to cry. And those very numbers are crying because they have been labelled with the wrong name.

In Vietnam the problem is even more acute. Vietnamese player names are brutally short and repetitive. Surnames like Nguyen, Tran, Le and Pham combined with common middle names produce hundreds of individuals bearing the same name. In a youth-player database, “Nguyen Van Nam” can appear at five different academies, and no identifier distinguishes them. When news sites aggregate data from these sources, they unwittingly blend the records of different children into a single profile.

Which means the name-collision problem is not merely a Western technical issue. It is the issue of every football culture with short, common personal names — which is to say, almost all of East and Southeast Asia.

Core Analysis

Three systemic errors, and how they amplify one another.

Error One: the name collision. In data science this is called the entity-resolution problem — determining which entity a name actually points to. In football the problem is harder than in most fields, because player names are short, repetitive and constantly abbreviated. “Rice” might be Declan Rice of Arsenal, or an amateur player. “Walker” is Kyle Walker of Manchester City, or anyone else. “Banks” was once the legendary Gordon Banks, and is also an obscure seventh-tier defender.

A system that only matches strings will fuse these entities together. The result is that you might read a scouting report in which a goalkeeper’s numbers are assigned to a midfielder, and a defender’s defensive metrics are placed beside a striker’s goal record. The reader has no way of noticing, because both are named “Jordan.”

In the big leagues, people have started using unique identifiers for each player — FIFA codes, or those of data providers. But most of the football content fans read every day does not pass through a system with identifiers. It passes through journalism, through social media, through aggregation accounts — places where the name is still the only thing identifying a human being.

And when a name collision happens, it does not stop at one article. It spreads. A wrong metric is cited in an analysis, the analysis is quoted by another outlet, and within weeks the wrong number becomes the consensus. No one traces it back to source, because the source vanished long ago.

Error Two: self-reported metrics.

In football, huge numbers of figures are supplied by the subject himself or his agent, not by an independent body. A transfer “worth 80 million euros” usually includes add-ons that may never be triggered. A player who “runs 13 km per match” may be quoting an internal GPS system that was never published. A “100 million bid” may simply be information released by one side to apply negotiating pressure.

These numbers, unverifiable as they are, have remarkable longevity. They get quoted again, folded into consensus, and eventually become “fact” in the sense that no one remembers the source. When a club says what a player is worth, and no one holds independent data to contradict it, that number becomes the benchmark by itself.

I once witnessed exactly this in China’s second tier. A player was advertised with a “92% pass completion rate.” Impressive — until I watched the footage and realised most of those passes were sideways and backwards. A high completion rate, but not a single decisive ball into the box. Technically correct, semantically wrong.

That is the nature of the second error: a metric that is not false, but placed in a context that drives the reader to the opposite conclusion. And when that metric is attached to a mistaken identity — Error One — the two combine into a profile that is entirely wrong about a person who does not exist.

This leads to a paradox I have not resolved. Clubs pay for accurate data, but the data they use most to make decisions is the least accurate kind: the assessment of an individual scout, the advice of an agent, and market rumour. Not because they lack good data, but because good data cannot answer the question they need answered: will this player fit our dressing room? And that question, to this day, no data system has answered.

Error Three: the rumour bubble.

This is the error I care about most, because it bears directly on how audiences form expectations. In every transfer window, the volume of rumour around a deal is usually inversely proportional to the quality of evidence behind it. The bigger the deal, the fewer people who actually know what is happening — and the more who write about it.

The mechanism is simple. When two parties negotiate, they are usually bound by a confidentiality agreement. Silence is maintained. But that very silence becomes the raw material for rumour: lacking verified information, people infer from indirect signs — a player left unregistered, a coach answering evasively, a photograph on social media. These inferences, stacked together, create the illusion of evidence.

As I have said before: transfer rumour is not information; it is the emotional state of the market. It does not tell you what will happen; it tells you what people want to happen. And those two things are frequently different.

There is a lesson here from my own experience. In 2026, when the whole media praised Germany as World Cup favourites, I wrote that Germany would be eliminated in the group stage. I analysed their duel-win rate in central midfield, and the fact that the coach had no Plan B when trailing. The piece was mocked. Then Germany lost 0-2 to South Korea and went out. Crowd consensus is not evidence; it is only an echo.

But I do not want to convert being right once into a rule. Because next time, the crowd may be right, and I may be wrong. The only reliable thing is not going against the crowd, but holding to verifiable data.

In 2026, when stadiums worldwide closed because of the pandemic, I found a variable that data models had not yet accounted for: crowd noise. In the Bundesliga, the home-win rate during the no-spectator period fell by nearly 12% compared with matches played before crowds. A variable outside every pass and expected-goals metric, yet directly affecting results. It showed me that football data is always missing a dimension — and that people tend to fill the missing dimension with assumption, rather than by admitting it.

The Contrarian Angle

If these three errors are so serious, why do they persist?

The answer is not technology. It is demand. Fans do not merely want to know the truth; they want to be told a story. An accurate but dull report will lose to a compelling but false rumour. So the pressure on information systems is not to become more accurate, but to become more engaging.

I ask myself whether I am exaggerating the severity of the problem. Perhaps most rumours are harmless, and most data is correct. Perhaps the three errors I have described are peripheral phenomena, not the central disease. I leave that possibility open, because I have learned that scepticism must also be applied to my own argument.

But one thing cannot be denied: when a single string of characters can point to four different entities, and no one in the distribution chain re-checks the identity, the system is no longer conveying information. It is recycling error.

And here is the truly counter-intuitive point. People assume the solution to a data problem is more data — more metrics, more models, more algorithms. I do not believe that. A system with little data that knows exactly who it is talking about is more useful than a system with millions of data points that cannot tell four people of the same name apart. The quality of identity matters more than the volume of numbers.

Takeaway

My prediction, made public so it can be verified: within the next two seasons, at least one major transfer will publicly collapse because of a data error — either a player misidentification, or an inflated self-reported metric. When that happens, do not blame the algorithm. The algorithm only does exactly what it was taught, using the dirty data we hand it.

The 0-6 in Sichuan was not a defeat; it was a door into the world of data. That door is still open. The question is whether we walk through it with a clear identity, or keep letting the system call all of us by the same name.