Martial Arts and the Data Void

On the night of July 12, 2026, at a small arena in Nishi ward, Osaka, the third round of a 67 kg kickboxing bout ended with a right hook. The stands erupted. The referee raised the winner's arm. Within three minutes, the promoter posted a single line online: win by knockout, round three.

No stat sheet. No strike accuracy. No successful defence count. When I asked for official data, the media officer answered flatly: 'We don't collect it.' I stayed until nearly midnight, rewound the recording and counted by hand. My figures differed from the crowd's impression: the winner had actually lost the first two rounds on effective strikes. The ring told a story, but almost no one wrote it down.

That void repeats across combat sports. Football has Opta and StatsBomb, hundreds of metrics standardised and delivered to broadcasters within seconds of the final whistle. Basketball has Second Spectrum tracking every step. Boxing, kickboxing, MMA, Muay Thai and vovinam, by contrast, leave data in the hands of promoters, one format per event, mostly unpublished.

In Vietnam the picture is more fragmented still. A fighter may have three parallel records: one held by the domestic federation, one by an international event he once competed in, and one assembled by fans online. The three rarely agree on the number of bouts, rounds, or even results. I once cross-checked a young fighter's records and found a four-win discrepancy between two sources. Which one was correct I could not confirm, because neither traced back to an original document.

This forces a choice on the writer. When data is empty, we either name the void or fill it with plausible-sounding guesswork. For years I have chosen the first path, even though it makes my work slower and less dramatic.

Before discussing missing data, it is worth looking at what does get recorded. A fight usually preserves four fields: winner, loser, method, and round. Four fields. A football match leaves thousands of data points; a fight leaves four numbers.

The most compelling part sits outside those four fields. Fans remember a fight for its pace, for the way one fighter switched stance after the second round, for the number of strikes he missed and was countered on. None of that is stored. I once built a manual tracking sheet for twelve consecutive bouts at an indoor tournament: strike rate per minute, number of retreat steps under pressure, the exact moment each fighter began to fade. After the event I compared it with four major articles. Not one carried even a third of those figures. The judgements in those pieces rested mostly on the writer's feel from a front-row seat.

In martial arts, the fight does not generate its own evidence; the evidence must be rebuilt by hand. That is the core difference from team sports, where machines and data providers do the work for people.

Then comes a second layer: even when data exists, it can vanish in transit. I once received a results file from an international event. The processing pipeline ran, but returned empty columns: no title, no source, no type, no summary, no information points. The only thing that survived was a domain label — 'martial arts'. A full dataset had passed through the system and come back as zero. For me that was a lesson in the fragility of information: data does not persist automatically just because it was once produced.

The third layer is judging. A professional fight is usually decided by three judges under the 10-9 must system, meaning the winner of a round takes 10 points and the loser 9, unless a point is deducted. The system publishes no reasoning. Three judges can watch the same round and return three different scorecards, with no obligation to explain. When the scorecards are released, they are three rows of numbers without argument. A clean strike, a tidy defensive move, a proactive stance change — all dissolve into a single figure.

Injury is the last layer, and the most concealed. Injury does not explode in one fight; it quietly accrues debt across seasons. A fighter who leaves the ring with his arm intact may still have accumulated, over three years, micro-damage to shoulder and knee that no dataset records. I once tracked a group of fighters for two years: the number who had to reduce training intensity after each competition cycle rose steadily, yet no public medical record reflected it. When a fighter collapses in the fifth round of a big fight, the media usually call it a surprise. To me it is an invoice from earlier seasons, paid on time.

For that reason my process has to build its own data. Based on my experience tracking fights, every fighter I write about must have a 'predictive index': one specific variable I believe will decide the outcome — strike rate in the first three minutes, defensive rate when pushed against the ropes, or the moment balance is lost after the fourth round. That index must be measurable, must come from at least three independent sources, and at least one source must come from a different market. That is the three-source rule, and it exists for a simple reason: three sources sharing one frame of reference are not three sources, but one source repeated three times.

Here a paradox appears that took me years to accept. The more data I pour into martial arts, the worse I read fights. When every strike is counted, people begin to believe the fight lives in the spreadsheet. But a fighter can win on total strikes landed and still lose the bout, because combat sports score by round, not by aggregate volume. One strike at the right moment can erase twenty scattered ones.

I once met a genuine counter-example after staking my belief on activity data. I assumed the more aggressive, pressing fighter would win, until I watched one concede position, accept being pressured, and then finish with a precise blow in the final round. The activity data said he lost; the scorecards said he won. Since then I have added a counter-evidence step: before publishing, I must find three pieces of evidence that refute my own position.

The map is not the territory; data is not the fight. The silence of martial-arts data is not merely a technical flaw; sometimes it is the trace of something that cannot be measured. And when the numbers fall silent, the writer must name that silence instead of decorating it with invented detail.

If Vietnamese martial arts wants a credible generation of analysis, the first task is not buying technology but building a minimum data standard: original records preserved, fighter profiles unified, and the reasoning behind decisions published. A martial art matures analytically only when it accepts that the fight is larger than any spreadsheet — but the spreadsheet still has to exist first.