The Mislabelled File: The Silent Crack Inside Vietnamese Football's Data Archive

In the archive I have built over 53 years in this trade, there is a file labelled "football". I opened it one January morning while reorganising my index for the closing rounds of a domestic season. Inside there was not a single line about football. No club, no player, no scoreline, no recorded minute of play. The file held a court's cause list, a petition seeking an inquiry into a hospital fire, and litigation over an additional toll charge levied on vehicles without electronic tags on the motorway. All of it was judicial administration. The label on it said two words: football. I sat still for a while. Not because of what was inside. Because of the chain behind it: who attached that label, when, under what rule, and how many other files in the data stores that serve Vietnamese football carry the wrong label in exactly the same way. The short answer is: a great many, and most of them will never be discovered, because nobody opens the file to check. From January 2026, when Vietnam's U23 side reached the final of the AFC U23 Championship in Changzhou, the volume of data surrounding Vietnamese football grew season by season. Broadcast rights for V.League 1 were resold and split; GPS vests appeared in training; some clubs hired dedicated staff to log metrics; and thousands of social accounts made a living counting goals and translating foreign reports. The verification layer did not grow with it. Across those seven years, I never once received a player dataset from a V.League 1 club that named the person responsible for each number. No ingestion timestamps. No provenance. Just numbers, standing alone, passed hand to hand as fact. I watch domestic matches in the stands, and at home I still log each game by hand: starting line-ups, substitution minutes, who shot from where. One June afternoon I sat down to cross-check my handwritten sheets against a data file a scouting group had sent me to "take a look at". The two differed by 14 shot attempts. It took me three days of tracing to find the cause: the file had merged a youth-team friendly into the first team's totals, because both matches shared a single fixture code. That was the first time I realised the problem did not lie with the player. The 2026 transfer window taught me that a contract is a signed confession. But to read that confession, the reader needs a clean archive. And Vietnamese football's archive has never been clean. Three mechanisms produce mislabelling, and all three are present in the domestic game. The first is label inheritance. A document does not choose its own label; it receives one from the page, the section, the category code of the place it was harvested from. If a judicial news item sits inside a feed tagged "sport", it enters the sports corpus as a sports article. No intent is required. One misplaced item is enough, and every later copy inherits the error. In Vietnamese football, this mechanism runs strongest at the transfer-news layer. A story about a company liquidating employment contracts, sitting beside a story about a club terminating a player's contract, gets clustered under the same keyword. From there, a corporate redundancy becomes "insider information about a club releasing players", and it travels faster than any correction ever will. The second is keyword collision. Vietnamese has a trait that makes automated classification fragile: many words carry double meanings. "Phat" is both a free kick and a fine. "Hop dong" is both a transfer contract and a commercial one. "Chuyen nhuong" is both a player transfer and an asset transfer. A system that counts tokens will gather all three document types under one label and add them into a single index. The third is the self-unchecking template. When an analytical process is pre-built as a form with fixed fields, the person filling it tends to complete every field rather than stop and ask whether the document belongs to the domain being analysed at all. The "entities involved" field is left blank; the "time sensitivity" field is marked "not assessed"; and a judicial administration file travels the entire pipeline as a football file. The frightening part is not one wrong file. It is wrong files being added together. When a mislabelled record enters a dataset used to evaluate players, it does not vanish. It becomes a row. The row contributes to an average. The average contributes to a trend. And the trend is used to price a human being in a transfer window. Numbers do not lie. The way we grip them in our hands does. In the Vietnamese market this effect has a specific name: price. A foreign striker is presented with "0.6 goals per game" from a league where the metric was aggregated across cup ties and pre-season friendlies. The club pays for that 0.6. The following season the striker arrives, plays 12 matches, scores twice. The crowd calls it the player's failure. In reality, it is the failure of a data cell nobody ever checked. And when an archive is contaminated, the cost does not stop at one bad contract. It stops somewhere else: memory. A dirty archive does not clean itself. Each new season adds another layer, and the older layers become harder to trace. A wrong figure written in 2026 is cited again in 2026, and by 2026 it has become "history". At that point nobody can tell whether we are arguing about a player or about a data-entry error. At 69, I do not believe in spectacular collapses; I believe in the quiet crack from the previous season. A player does not lose form in one afternoon. He loses it across a run of seasons in which his metrics were measured wrongly, compared wrongly, priced wrongly, until he himself begins to believe an inaccurate description of himself. There is a common reflex I want to question. When a national team plays badly, the reflex says we need more data: more metrics, more platforms, more analysts. I disagree. More unverified data does not make decisions better; it makes them more confident. And a confident decision built on dirty data costs far more than a hesitant decision built on direct observation. The second reflex I want to question is the habit of importing standards. We buy foreign metric packages, adopt their terminology, but do not bring their verification layer. Those packages were designed for a league with 380 matches a season and a data system that has run for decades. Placed over a league with fewer matches, fewer sources and an incomplete ingestion process, the result is not a copy of that system. It is a handsome form and a wrong conclusion. I have examined many contracts over the years, under every angle of light, and what I have learned is that the true motive of a deal is on the page before the first match is played. But that page is only useful when the numbers on it are honest. A wrong label at the data layer travels straight into the negotiation layer, and there it stops being a technical error. It becomes money. So what should be done. I am not proposing more machinery. I am proposing a gate. Every record entering the archive of a club, a federation or a newsroom needs three things attached: an absolute timestamp, the name of the person who entered it, and its source. Without those three, the record is not permitted onto an aggregate table. This is the cheapest task in the entire analytical chain, and the most commonly skipped. Second, a simple test before labelling: does the text contain at least a few entities belonging to the domain. A football article must contain a club, a player, a competition or a governing body. If none are present, the label must be suspended pending human confirmation. Third, the trade must accept something it rarely accepts: permission to say "insufficient information". In analysis, the sentence "no conclusion can be drawn" is worth more than a conclusion manufactured to fill a template. Vietnamese football has paid for manufactured conclusions more than once. Before every transfer window, I ask young colleagues the same question: this number, who is responsible for it. If nobody can answer, I strike it out. That question, I think, deserves to be asked in every football analysis room in this country, before the next season begins, and before one more mislabelled file drifts into the archive.