From a Mexican Pension Card to V.League: The Cost of a Misapplied Data Label

At 3:47 in the morning, as my data table completed its last refresh of the day, a single line surfaced among thousands of others. Classification label: football. Right beside it, the headline of the source document concerned something entirely alien to a ball — the "credencial blanca", the identification card that Mexico's Instituto Mexicano del Seguro Social (IMSS) issues to retirees and pension recipients.

I read all twenty-two information points extracted from that text. Not one club. Not one player. Not one coach. Not one match. Not one xG figure, not one PPDA value, not one pass-completion-under-pressure metric. The entire content revolved around administrative procedure: documents to submit, photo specifications, a national mixed commission on retirements and pensions, and a list of discounts at Tiendas IMSS stores.

One mislabelled data row. Small as a speck of dust. But I once spent four months re-watching all 26 matchdays of a single season because I believe specks like that decide the quality of an entire information system. Every prophecy begins with a table nobody bothers to read.

A data label is infrastructure, not paperwork

To understand why a row like that matters, you have to understand how a modern sports desk actually operates. Every day, thousands of documents arrive from hundreds of sources: club press releases, wire copy, aggregator summaries, federation notices, notes from press conferences. Nobody reads all of it. The system tags automatically, classifies by topic, and routes everything into separate processing streams: tactics, transfers, club finance, medical, competition law.

The label is the infrastructure. It is like the touchline on a pitch: nobody sees it on a television feed, but remove it and the match becomes an unending brawl with no boundary. When a document about a Mexican pension card is tagged "football", it is not merely in the wrong place. It becomes raw material for the next decisions: a round-up edited around it, a metric computed from it, a prediction issued because of it.

In Vietnam, this infrastructure is being built faster than it is being checked. V.League now has event data, xG tables, transfer metrics, statistical platforms that would have been unthinkable a decade ago. But data infrastructure is not only sensors and APIs. It is also taxonomy discipline — and no technology vendor sells that alongside the licence. Based on my experience watching matches in both England and Vietnam, the hardest part of data work has never been collection. It is deciding where a thing belongs.

The evidence chain of a mislabelled document

Back to the twenty-two information points. Pull them apart and examine each one, and the picture becomes clear.

The subject of the document is an administrative procedure: the white card issued by IMSS to people who have retired. It sets out the card's purpose, the benefits attached to it, and most importantly the conditions for obtaining it. One technical detail stands out: this card does not replace official identification, and holders may still be asked to present an INE voter card or a passport. Another point states plainly that the card is not mandatory in order to receive the monthly pension.

Those two sentences are the mark of a disciplined source. A careless document would not attach its own scope limitations. It suggests the original text likely derives from an official IMSS FAQ — a reasonably reliable primary basis, with only the journalistic layer left unattributed.

On procedure, the text describes filing in person before a national mixed commission on retirements and pensions, a document list including a payroll receipt, a photo requirement with a validity window, followed by review and issuance. None of it touches sport.

What is telling is that the classifier picked up the signals "credit" and "discount" — words that overlap with the vocabulary of football finance. This is the error type I call a keyword false positive. A word appears in the right shape but entirely the wrong context. In football data this error is not rare: "clause" in an employment contract read as a release clause, "option" in a commercial notice read as a purchase option, "credit" in a banking story read as a club loan.

And here is the point I most want to press: the most serious fault in football data is not a wrong number, but a right number filed under the wrong category. A wrong xG can be caught by comparing it against another model. A pension document sitting inside a football category cannot — because it contradicts nothing already there. It simply persists, and over time it skews everything built on top of it.

Every piece I write carries at least three advanced metrics with verifiable sourcing. Here the verification is not a number but a classification question: which vertical does this document belong to? Answer that wrongly and three correct metrics mean nothing.

One further detail makes this row more notable than usual: the text carries no effective date. Without a timestamp, a reader cannot know whether the photo rule or the document list still applies or has been superseded. In football data, a metric without a time stamp is similarly hazardous: a team's PPDA sample at matchday 5 cannot describe that same team at matchday 20, once the fixture list has thickened and squad depth has shifted.

In 2026 I spent four months re-watching 26 matchdays of the season Hanoi FC won. The figure I measured was an average PPDA of 9.8 — the highest in the league, reflecting how aggressively that side pressed to win the ball back inside the opponent's third. My first analysis was dismissed by colleagues as dry, academic, unfeeling. I did not change style. I added xG comparison tables and squad-length data to the next three pieces. By the end of the year several clubs had begun copying that pressing model, and the article was suddenly being shared widely among players.

The lesson from that season was not the number 9.8. It was that I had to check every single matchday by hand because there was no trustworthy data table to lean on. That work took four months for one club's season. Applied across the whole of V.League over ten seasons, it would take thirty-three years. That is why data classification quality is not a small technical matter. It is a question of whether a league's information resources are sustainable at all.

The counter-intuitive angle: the fault is not in the bad row

The first reaction most people have on hearing this story is to blame the algorithm. I think that conclusion is wrong, and wrong in a convenient way.

The tagger works on keyword and context probability. It behaves exactly as designed. The document contains "credit", "pension", "national commission" — and in some training sets those words have appeared in articles about club finance and federation governance. The system did not make a logic error. It faithfully reflected the ambiguity of the language it was taught.

The real fault lies elsewhere: in the incentive structure. A newsroom is measured by output volume, not by classification accuracy. Nobody is praised for spotting a mislabelled row. Everybody is credited for publishing ten more pieces in a day. In that environment, cross-checking is a cost and error is free.

Spectators can leave the stand, but the number stays in its seat. In this case the number sat in the wrong seat for months without anyone noticing, because nobody was paid to sit next to it.

I must argue against myself here. What would change my view? If the data showed mislabelling running below one in a thousand with no measurable downstream consequence, this would be statistical noise and undeserving of an analysis. But in the systems I observe, classification-layer errors tend to multiply as they pass through successive aggregation steps, because each step trusts the one before it. That is why I hold to the cautious position.

The transfer market is not a game of sentiment; it is a game of maps being redrawn. And a map with a wrongly drawn boundary sends the whole convoy astray, even when every individual vehicle in it is running beautifully.

Signals for the next cycle

V.League does not lack numbers. It lacks people who know how to frame those numbers into a window. The story of the Mexican pension card is not a football story. It is a story about what materials we are using to build those windows.

Three signals I will be tracking. First, whether Vietnamese football data platforms publish their tagging methodology openly — because a process that is not published cannot be challenged. Second, whether clubs begin requiring unique record identifiers for every transfer entry, instead of letting each party name the same deal differently. Third, whether supporters start asking about the source behind a number, rather than only asking what the number is.

We go looking for the future of football while it already sits in pasts that have never been encoded. A player expresses emotion; ten seasons are needed to make a system. The trouble is that the system is only as good as its weakest data row — even when that weakest row is talking about a pension card in a country half a world away.