The File in the Wrong Drawer: How a Puebla Rain Bulletin Landed in the Football Ledger

In my Khulna filing cabinet there is a separate drawer labelled “wrong room.” That is where I keep records that look like one thing at first glance and turn into another when the tape rolls. For seventeen years I have written minute, law and decision into a referee’s notebook; no file reaches that drawer until I have watched it at least three times. On the morning of October 2, 2026, a new file dropped straight into it. The report came from the Mexican state of Puebla. The state’s education secretariat, SEP Puebla, announced that in-person classes would be suspended at all schools across 173 municipalities because of heavy-rain forecasts. The civil protection authority issued its recommendation: stay indoors during electrical activity, gusty winds and intense rain, and continue study online. The notice carried no minute, no law, no decision — the three things I have never written an episode without. At first glance this is not football news. Yet the system that processed the bulletin stamped a label on it — “Football.” A school-closure notice with no relationship to any pitch walked into the football analysis ledger. The first lesson of refereeing is that you cannot announce a decision before you have classified the incident. Law 12 has three steps — careless, reckless, excessive force. The same tackle draws three different sanctions: free kick, yellow, red. Classify it wrongly and every remaining calculation flips. A data pipeline’s domain label is the same kind of classification decision. Every classification carries a price; the wrong class means the wrong reader, the wrong analyst, and a staircase of wrong decisions. What the Stage-1 analysis found is the raw material of this piece. All eighteen information points describe class suspensions. There is no formation, no xG, no transfer fee, no FFP, no league table, no manager under pressure, no broadcast revenue, no dressing-room health. What exists is a secretariat notice, a civil-protection recommendation, an instruction on remote learning. In other words, the analyst’s hands hold no football, only a wrong label. The curious thing is that the error itself is the file’s only concrete fact. The analysis rated it one star out of five — one for sporting value, one for industry value, one for reference value. A file with so few stars usually sits in the drawer, and precisely for that reason it should not. I went back to the 2026 frames to see what the naked eye missed. France against Australia, the 58th minute. Referee Andrés Cunha walked to the monitor and came back with a penalty — Josh Risdon’s hand blocking Antoine Griezmann’s shot. It was the first VAR penalty in World Cup history. That day the broadcast said “no hand”; Law 12 asked whether the arm had made the body unnaturally bigger. Two accounts, one incident, one monitor. In that database of 64 matches I recorded 29 penalties, 12 own goals and 29 VAR reviews. The numbers accumulated only after I had watched each contested incident three times. That habit taught me the rule: a label is a claim, and a claim needs evidence beside it. At the 2026 Confederations Cup, the Chile 0-1 Germany final was refereed by Milorad Mažić. Across that tournament I reviewed 16 matches and logged 42 yellow cards, 3 red cards and 5 penalties — 2.63 yellows per match. That became my archive’s first baseline. For three months I refused to live-stream. I later published 24 episodes, each citing the Laws and minute marks. That slowness taught me that nothing goes to print until the label matches. I learned the same discipline in 2026, the day I first sat behind a microphone at Bangladesh Betar. After taking over Krira Jagat as editor in 2026, and later as founding sports editor of Prothom Alo in 2026, I watched archives tell wrong stories simply because files sat under the wrong heading. That long editorial road taught me that classification is not a harmless chore. Now the same chore is due here. The “Football” label makes a claim — that this file is about football. But there is not a single pitch inside it. It is the referee’s moment when the crowd screams penalty while the ball is still lying by the corner flag. The database does not shout; it waits for you to ask the right question. Ask the wrong one and it quietly hands you the wrong answer. The question here should have been: is there football in this record? The answer was no. Nobody asked. Still, a real football story is buried under the label, and it matters more. 173 municipalities is a large slice of Mexico. Where schools close, pitches are wet. In sustained rain, grassroots weekend schedules drown, youth-academy sessions move, and postponed fixtures pile into the following week to create congestion. From the academy supply chain to broadcasting, a small tremor travels. Another detail is buried too: had any match been scheduled in those zones on October 2, the local football association could have invoked a force-majeure clause to postpone it. But no such match appears in the file — because the file is not about football. Now let me borrow one metric, with conditions attached. My referee log’s original domain is match discipline; its job is to measure the density of cards and penalties per match — not data errors. So it does not apply here directly. A resemblance survives nonetheless: one false positive in the log inflates the whole count. One wrong label inflates a whole analysis the same way. Better to stop the analogy there; pull it further and cleverness outgrows the evidence. One more thing is buried in this file, invisible at first look. Shutting in-person activity across 173 municipalities in a single day means a specific organisational capacity of the state. Civil protection and the education secretariat worked together, received a forecast, made a decision, and told the public. A state that can do this in a day can also handle security logistics for a major sporting event. That is a usable signal for football administrators. A tracking signal has also accumulated. If education articles keep arriving under a “Football” tag, this is not an accident but a pattern. Caught early, the fix is quick; caught late, trust in the archive drops. Now the other side. When the crowd says penalty, I look for the angle the crowd cannot see. The same applies here. Everyone blames the algorithm. But a referee knows the error usually is not in the whistle; it is in the decision tree. A pipeline that never once asked “is there a pitch in this record?” is not an algorithm’s fault — it is a design gap. Emotion says: decide now. The rule says: verify first. In football that tension returns in every contested moment; in a data pipeline it is quieter still, because there is no crowd there, only output. Someone will say: no football, so no story. My answer: the absence is the story. An archive that cannot tell a school from a stadium cannot be trusted in the 88th minute. Excessive caution is its own danger, though. Throw out every unfamiliar label as an error and you lose the real signal too. So draw the line: here the error is clear, because there is not one mark of football inside the file. In Khulna, I learned that a new feed can change the old rules. Here the reverse happened: the old rule — “label what you see” — broke under a new feed. It is like monsoon Khulna, where one day of rain rewrites a week’s schedule. One task is simple: correct the domain classification. Then audit Stage-1 output regularly — if education articles keep arriving under a “Football” tag, the pipeline needs a fix. One practical job remains for the October 2 window: cross-check local football schedules across Puebla’s municipalities. Knowing in advance which match slips and which pitch floods makes the shouting afterwards quieter. I trust the replay, the rulebook, and the long walk to the monitor. The referee’s closing question: if a pipeline cannot separate a classroom from a pitch, what other files is it quietly filing in the wrong drawer?