When the Dataset Is Empty: Vietnamese Martial Arts and the Trap of a Single Yardstick

In September 2026, in a press room inside an arena in Osaka, a young reporter sitting next to me asked: "Sir, how many bouts has this wushu athlete won?" It took me almost a minute to answer. The question was grammatically correct but structurally wrong. The athlete we were discussing competes in taolu, the routine discipline. She has no opponent across the floor, no rounds, no overtime. What she has are three judging panels, a difficulty score sheet, and a margin measured in tenths of a point. Yet in the questioner's mind, and in the minds of many Vietnamese sports editors, she sits in the same basket as MMA fighters.

Six months later, an empty dataset landed on my desk. No title. No source. Not a single information point. The eight-dimension analytical framework was intact, but every cell read "insufficient information to assess." The natural reflex of any writer is to fill those gaps with a few names, a few numbers, a few plausible-sounding conclusions. I did not. That decision taught me more than any complete analysis I have ever written.

An empty data cell does not mean "neutral." It means "unknown." Those two words sit very far apart. In this profession, they sit apart by an entire career.

Why the label "martial arts" does harm

In Vietnam, "martial arts" is a label so broad it becomes useless. Underneath it sit at least six competition systems with entirely different logic. Olympic boxing and professional boxing are scored by the number of clean strikes, by three rounds or twelve, by judges' decisions. MMA adds wrestling, locks, chokes, and finishing criteria. Kickboxing and Muay Thai have count rules, knee-strike rules, and different banned zones depending on the sanctioning body. Wrestling and grappling all but abandon the concept of striking, leaving only control points. Sanda is a hybrid species. And taolu lives in a different universe altogether.

The 32nd SEA Games in Cambodia in 2026 demonstrated this more clearly than any theoretical essay. Vietnam's delegation topped the overall medal table. Wushu contributed in both taolu and sanda. Each sub-discipline has its own medal table, its own scoring, its own ranking. But across most of the press, it all collapsed into one line: "Vietnamese martial arts competed successfully."

That line is not wrong. It is merely useless. It merges three different logical systems into one emotional symbol and turns a sport with three rulebooks into an unmeasurable lump of sentiment.

Three subject classes, three yardsticks

Class one is modern competitive combat sports: boxing, MMA, kickboxing, Muay Thai, grappling. Here the unit of analysis is the bout. Wins and losses. Finish rate. Rounds contested. Effective striking rate per minute. Takedown defense. Title defenses. For MMA we add head-impact data, because that is a long-term debt with no insurance policy.

Class two is sanda, the bridge between two worlds. Sanda has an opponent and has rounds — usually three, two minutes each on the international circuit. Points come from three sources: effective strikes, legal throws, and pushing the opponent off the platform. None of those three scoring sources exist in any boxing scorecard. And pushing an opponent off the platform is not a foul. It is a scoring technique.

Class three is taolu, where the concept of an opponent disappears entirely. The score is assembled from three judging panels. Panel A scores movement quality: a rotation a few degrees short, a posture that misses the standard, points deducted straight from the total. Panel B scores performance quality: rhythm, power, continuity between movements. Panel C scores difficulty: rotational jumps with a prescribed number of turns, each level carrying its own point value.

Duong Thuy Vi, for many years a pillar of Vietnam's national wushu team, has not won a single bout in her career. She has no bouts to win. She scores with a sequence of movements that nobody blocks, and loses points because a foot lands a few degrees off.

That is why I call the shared yardstick a technical trap.

The classification error lives in the workflow, not in the audience

People like to blame the audience. The audience misunderstands because it lacks knowledge. I do not buy that. Drawing on my experience watching thousands of bouts across many different disciplines, I find the error sits with content producers, and it sits at a very early step.

My three-source verification process starts with a question about the discipline, not a question about the person. Skip that first step and the three sources become meaningless. I will pick three sources that speak one language, look through one lens, and err in one direction. One source from MMA, one from boxing, one from wushu — that is what genuine cross-checking looks like. Three sources from three outlets all repeating one press release is one error counted three times.

The empty dataset I received six months after that question in Osaka was exactly that kind of diagnosis. It flagged three concrete defects. No information-point extraction. No entity extraction. No subject-class determination. The third is the heaviest, because it poisons every step that follows. Applying win-loss logic to a taolu article produces systematically misleading conclusions even when the sourcing is perfectly clean.

The injury debt differs by class

Taolu athletes pay in joints and tendons. Rotational jump volume accumulates over years, and the knees carry the heaviest ledger. MMA athletes pay in brain tissue. Nguyen Thi Tam, the Vietnamese female boxer who competed at the Tokyo 2026 Olympics, pays in both, plus the pressure of the weight cut. Nguyen Tran Duy Nhat, a multi-time Muay Thai world champion, pays in shin bone and in years of professional competition.

Injuries do not detonate inside a single bout; they quietly accrue as debt across many seasons. An injury file is only worth something when you know how many bouts the athlete has fought, in which discipline, under which rulebook. The same figure of thirty appearances a year means something entirely different for a taolu athlete and an MMA fighter. For taolu, thirty appearances means thirty times the body absorbed jump load. For MMA, thirty appearances means thirty times the head absorbed impact.

The weight cut is a technique, not an administrative detail. In modern combat sports it determines the real physical condition when the first bell rings, and it is the largest medical risk source that is least well recorded. In taolu, the concept of a weight cut does not exist. So a shared "condition" index spanning both classes is an invented index.

The counter-intuitive angle

The industry is loudly calling for a unified martial arts data system. I think that is the wrong order. What must be built first is not a bigger data warehouse but a classification layer rigid enough that every subsequent data point lands in the right place.

A data warehouse without a classification layer only amplifies error. It turns a small labelling mistake into a large publishing conclusion, and that conclusion spreads into headlines, into commentary, into audience expectation, and then returns as pressure on the athletes themselves.

Methodical counter-argument requires me to hunt for evidence against my own position. On this topic, the first piece of evidence is the 32nd SEA Games. Taolu and sanda were reported side by side, and Vietnamese readers followed both. Any hypothesis that audiences cannot tell the two sub-disciplines apart has been refuted by the data.

The second piece of evidence is Martin Nguyen, the Vietnamese-Australian fighter who held the ONE Championship featherweight world title. Analyses of his career are read closely, and readers are already fluent in the specific language of MMA: takedowns, chokes, control. The problem is not the audience's reading ability.

The third piece of evidence is the empty dataset itself. It is not worthless. It reveals that the classification step is being skipped at the production layer. An empty dataset at the production layer becomes a full article at the publishing layer — through fabrication. That is the professional line I refuse to cross.

The rules I apply to every martial arts article

One: determine the sub-discipline before determining the person. Two: state the applicable rulebook explicitly, because one name can compete under three different rule sets within three years. Three: never place the word "martial arts" in the subject position of a conclusion. Four: mark clearly what is a primary fact, what is inference, and what is speculation. Five: when the data stays silent, name that silence out loud.

Every number tells the truth, but a bout never tells the whole truth. Our problem is not a shortage of numbers. Our problem is that we line up numbers from three different worlds in the same row and then act surprised when the sum means nothing.

A forward-looking thought

A map is not the territory; data is not the bout. The writer's job is to know at what scale the map is being drawn and to tell the reader so. Before Vietnam needs a sports data industry, it needs a small, boring classification layer that gets checked three times. Who will build that layer, if not the people sitting in the Osaka press room every night?