One Wrong Label and a Systemic Error in Combat-Sports Reporting

The "Information Points" field in a combat-sports analysis file returned zero. No fighter names. No event. No organization. No bout date. The only surviving element was a classification label: martial_arts. I spent two days re-checking that file. The one conclusion I could defend in front of an editor: the extraction layer failed, and any professional judgment built on top of it is fiction. The surviving label is the real story, because it reproduces a mistake combat-sports media makes every day — collapsing MMA, boxing, kickboxing, Muay Thai, grappling, taolu and sanda into a single word, then measuring all of them with one ruler. In any sports-analysis workflow, subject classification is the step that must be completed first. Three subject classes demand three different logics, and there is no way to blend them while keeping accuracy. The first class is modern competitive combat sports: MMA under the Unified Rules, professional boxing, kickboxing, Muay Thai, grappling. The source data here is bout outcome, finish rate, defensive efficiency and opponent quality. The second class is performance martial arts — taolu. No opponent, no rounds, no referee on the mat. The score comes from a judging panel with separate panels: Panel A scores movement difficulty, Panel B scores performance quality, Panel C applies deductions for compulsory errors. The third class is sanda, a sparring discipline with its own ruleset: three rounds of two minutes, scored on legal strikes, throws and pushing the opponent off the platform. A file that cannot determine which class it describes cannot select the correct analytical framework. With the wrong framework, the error does not stay in a few details. It spreads into every conclusion. The score sheet is where this shows most clearly. In taolu, an athlete can win national titles for years without a single bout on record. Placing that person in a win-loss table is a meaningless operation. The only measurable outputs are movement difficulty and performance quality across judging rounds. The reverse holds too: placing an MMA fighter on a taolu scorecard is meaningless in exactly the same way. In sanda, defence is not measured by takedown-defence percentage as in MMA. A kick landing to the head, a clean throw and a push off the platform carry three different point values. To assess a sanda fighter, I have to read the per-round score sheets, cross-check foul counts and deduction totals. In grappling, ADCC and IBJJF operate on technical points plus advantages. An athlete can win a match without finishing the opponent. Placing that athlete beside an MMA fighter with a 70 per cent knockout rate puts two measurements of different origin on one table. Professional boxing uses a mandatory 10-9 scale, where a round almost always has a winner. MMA uses 10-9 as well but permits drawn rounds and clearer point deductions. The two systems look identical on paper; they diverge in how the result is read. I started an investigation once with a single skewed figure in a payroll sheet. It ended in an unnumbered room. The same chain applies here: one skewed data field at the top drags every downstream error with it. Testing labs do not know the athlete's name. That is why I trust them. A standard scoring system works the same way: Panels A, B and C do not know who is performing. They record movement, difficulty and error. When a report attaches an "unbeaten record" to a taolu athlete, the writer has injected a name into a system designed not to need one. Based on my experience tracking competitions, the rate of framework error in combat-sports reporting shifts with the season. Around the SEA Games and national championships, the error spikes: taolu athletes get described in fighting language, sanda fighters get graded with statistics from international professional circuits. This file recorded three defects at the first layer: no information extraction, no entity extraction, no subject-class determination. All three are fixable with a single mandatory field. An empty stadium looks clean. The dressing room does not. Reports work the same way: the presentation is clean, the classification underneath is not. A contract usually runs one page. A dirty contract carries an appendix. Competition rules follow the same pattern: the first page talks about sporting spirit, the appendix talks about score panels and deduction values. The other side has a case, and it is not weak. General audiences do not follow score sheets; they follow stories. For a broadcaster, one shared analytical frame across all martial arts saves airtime and holds viewers through advertising blocks. But the cost of that convenience is paid in specific decisions. A scout reading an MMA record to evaluate a grappler will skip precisely the skill set that needs measuring. A sponsor signing a taolu athlete on the basis of a "win rate" will misprice the asset. A betting market opening a sanda line with a boxing model will post a wrong price on the first line. The largest risk sits in holding correct data inside the wrong framework. A file that returns zero carries the status "unknown." Any newsroom without a mandatory subject-classification field before writing is publishing conclusions it cannot verify.