The 'martial_arts' Label Classifies No One

The 'martial_arts' Label Classifies No One

At eleven at night the handover file opened on screen. The information field came back empty. No fighter name. No event. No governing body. No publication date. Only a single surviving label line: martial_arts.

To an ordinary reader, an empty data field is a technical matter. To me it is a red flag. Across seven years of building investigative files, I have never seen a dossier immaculate from its first line that ended up factually correct. The doping file sat on an assistant coach's old hard drive. Modification date: the night before the play-off. The anomaly was never in the content. It was in the blank.

That night, eight analytical dimensions were assembled to read an article about martial arts. All eight ended on the same sentence: insufficient information. The more telling detail sat elsewhere — the label.

Three worlds under one word

Martial arts is not one discipline. It is three worlds stacked under a single word.

The first is professional combat sport: MMA, boxing, kickboxing, Muay Thai, grappling. The measures there are win-loss record, finish rate, takedown success, striking differential and weight class. A 24-year-old with 12 professional bouts is read with a completely different toolkit than a 34-year-old with 40.

The second is traditional and performance martial arts, taolu above all. There is no opponent. No extra rounds. Scores come from movement difficulty, precision, stability and the judges' technical scales. Applying a win-loss table here misreads the discipline from the root.

The third is sanda. A hybrid, with genuine contact but scoring, grappling rules and fault counting that differ sharply from MMA.

Three worlds, three logics. The martial_arts label does not say which world the file belongs to. That is the break point.

The blank spreads downward at a very even tempo

When subject classification is left blank, every layer beneath collapses on the same beat.

The evidence layer falls first. With no information points extracted, there is no factual unit to hold onto: no named fighter, no gym, no coach, no federation. The technical board is left with empty columns for style matchup, finishing ability and record quality.

The athlete layer falls next. No name, no age, no bout count, no injury history. The career-age curve cannot be drawn. Weight-cut risk — one of the mandatory red flags — cannot be screened, because weight class and weigh-in results are missing.

The organisational layer is entirely blank. Whether the file concerns a UFC, ONE or PFL-tier event; one of boxing's four major sanctioning bodies; a kickboxing circuit such as Glory or K-1; a grappling circuit such as ADCC or IBJJF; or a multi-sport games taolu context is undetermined. No hierarchy diagram can be built without a top unit.

The commercial layer is blank too. With no gate, broadcast or purse figures, the revenue structure cannot be decomposed. And the governance layer — the one that decides every ruling on the rules — has no authority to cross-check against.

What I want to stress sits here: a blank classification label does not produce a gap in information. It produces false information wearing valid clothing. When an analyst is forced to fill every column, they fill it with the toolkit closest to hand. The industry default is the MMA and boxing toolkit — record, knock-out rate, output. Applied to a taolu athlete, the result is an analysis that sounds highly professional and is wrong at the level of substance.

Based on my experience watching fights over many years, one pattern repeats: the most serious error never comes from dirty data. It comes from clean data placed in the wrong frame.

Three times I worked a file, I reversed the familiar order. I began an investigation with one anomalous figure in a payroll sheet. I ended in a room with no number. But before reading that anomalous figure, I had to establish whose payroll it was, in which league, in which season. In 2026, the 37 test samples in a leaked file from the Moscow anti-doping laboratory only became meaningful once I knew which playing position and which half each sample belonged to. In 2026, 14 pages of medical documents about a winger only became a dossier after I separated permitted treatment from procedural violation.

The laboratory does not know the player's name. That is why I trust it. But the laboratory always knows what type of test that sample was. Classification comes first, conclusion comes second. There is no exception.

At the risk layer the gap is most visible. Brain health, cumulative strike exposure, concussion history — no data. Weight-cut incidents — no data. Post-career security — no data. And here is the line I want on the record: the absence of a risk rating must not be read as the absence of risk. The correct status of this file is unknown. Not neutral.

The charitable reading

There is a charitable reading, and it is not unreasonable. A data pipeline is a technical system; technical systems fail. A blank field is an operations fault, not an editorial one. Fix the pipeline, re-run it, and the full dossier returns.

That argument is correct on the technical half and missing on the human half. The same blank classification happens every day at the copy desk, except nobody calls it by its fault name. A story about a traditional martial arts event gets written on an MMA template. A sanda report gets graded by boxing criteria. A performance athlete gets assessed on win rate. The writer does not see themselves filling a blank field, because the default toolkit is already sitting in their hands.

Here the pipeline's own caution accidentally produces a form of validation. The file returns nothing but empty values, so no sentence is fabricated. It sounds like a clean result. But clean here means clean because it was never written, and that white space will be filled by the next person with an assumption.

I have seen the consequences of that kind of filling in another role. In 2026, when a club in Tianjin announced dissolution, 28 players and 14 staff lost their income. Most reports simply recorded the disappearance of a team. Nobody touched the bank statements, where 11 million renminbi had moved through three shell subsidiaries. The stadium was clean. The dressing room was not.

The charitable reading also overlooks one detail: martial_arts is a top-level label that merges three different logical systems into one cell. Fixing the blank field does not fix the wrong label. Those words will keep flowing down into every layer beneath, and every layer will pick its own toolkit rather than wait for classification.

What remains

What is needed is not a re-run of the pipeline. It is one mandatory field: subject classification. Decide first whether the file concerns professional combat sport, traditional and performance martial arts, or sanda. That field determines the entire toolkit behind it, and there is no shortcut around it.

A file with no fighter name can still be archived. A file with no classification must not go to press. If that night's file had carried a single line stating which world the subject belonged to, eight analytical dimensions would not have ended on the same sentence. When a combat sports report does not state which world it is describing, is the reader receiving analysis — or receiving a data field that has been filled in to look complete?