Three Scoring Systems Under One Name: The Classification Gap in Vietnamese Martial Arts

Page 41 of my third notebook holds two numbers side by side. On the left, 9,730. On the right, 2-1. Both were recorded at 21:40 on 15 May 2026 in Hanoi, roughly two hundred metres apart as the crow flies. The first figure is a taolu routine score for a single athlete, marked on a 10,000-point scale. The second is the scorecard of a combat bout, expressed as the number of rounds three judges awarded to two different corners.

I sat between two arenas and realised I had just written two different units of measurement onto the same page. Not out of laziness. Because in my head, and in the heads of almost everyone in that press room, both were filed under a single Vietnamese word: vo thuat, martial arts.

That was the moment I understood that the problem in the field I have covered for twelve years does not lie with the athletes, does not lie with the coaches, and largely does not lie with the money. It lies in the frame. We are measuring three different things with one ruler, then expressing surprise when the results do not add up.

Numbers do not lie. We simply have not asked them the right way.

Context: one name, three systems

In Vietnamese, martial arts is an umbrella term. It covers at least three groups of disciplines whose competitive logic differs at the level of first principles.

The first group is performance-based: wushu taolu, taekwondo poomsae, vovinam forms and sparring patterns, traditional martial arts routines. The athlete steps onto the floor alone or as part of a team, with no direct opponent, and a panel of judges awards a numeric score.

The second group is judged combat: boxing, kickboxing, muay thai, taekwondo kyorugi, karate kumite, judo, pencak silat, sanda. Two people on the mat, referees present, scorecards in play, and a binary win-or-lose outcome at the end.

The third group is pure binary outcome: MMA, wrestling, freestyle wrestling, and most grappling systems. There is no scoreboard visible to the crowd minute by minute. There is only the result, and the method by which it arrived.

These three groups use three different counting systems. Yet in newspapers, in medal tables, in television bulletins, they are compressed into a single line: Vietnam won a martial arts gold medal.

I began tracking that compression in 2026, when I started out as a journalist in Australia before moving to Vietnam. At first I assumed it was mere shorthand. Then in 2026, when the pandemic shut down every competition and I spent eight months at home coding data, I saw that the compression was far from harmless.

Based on my experience covering bouts, a small distortion in classification cascades through the entire chain behind it: how targets are set, how athletes are selected, how coaches are paid, how local budgets are allocated, and how headlines are written.

Core: the three systems and what each optimises

System one: relative panel scoring

In the performance group, the number is not the outcome of an event. It is the outcome of a panel.

A wushu taolu routine is marked on two main axes: technical difficulty and execution quality. The difficulty score depends on the movement group the athlete registers in advance. The execution score depends on how the panel rates the quality of the performance. The scale is usually expressed out of 10,000 so that decimals can be separated without premature rounding.

The critical point is this: a performance score is a relative quantity. It only means something when you know who judged it, under which ruleset, and at what moment. A routine scoring 9,730 at a 2026 regional games and a routine scoring 9,730 at a 2026 national championship do not say the same thing, because the average difficulty of the entire field has shifted.

In my notebook, every performance entry carries three mandatory fields: the judging panel, the ruleset applied, and the field average for the same event. Without the third field, 9,730 is meaningless. It is like a student scoring 8.5 with no one knowing whether the exam was hard or easy.

In other words, the first system measures performance quality against a drifting standard. It does not measure who is stronger than whom in any combative sense.

Much of the online argument about whether Vietnamese martial arts are superior or inferior is really an argument between two different standards within the same scoring system. Fans compare a domestic routine with a foreign routine, but the two athletes competed at different events, were judged by different panels, under different versions of the rulebook. That is comparing two temperatures measured by two uncalibrated thermometers.

System two: accumulation inside a time window

The judged combat group runs on different logic. The final result is binary, but the path to it is cumulative.

Amateur boxing scores by round, based on clean and effective strikes, with different coefficients for different punch groups. Taekwondo kyorugi awards points per technique, with a substantial weighting gap between body and head strikes, and electronic systems often intervene directly in recognition. Karate kumite scores technical points plus accumulated penalties. Judo takes another route entirely: a technical score decides an outright win, while accumulated penalties lead to sanctions and ultimately to the end of the contest.

All three belong to group two. But they optimise three different things. Amateur boxing optimises the volume of effective scoring blows within a short window. Taekwondo optimises the selection of the highest-coefficient technique. Judo optimises the creation of a single decisive moment.

If you take the medal counts of those three sports and add them into a single figure for combat martial arts, you have blended three distinct training objectives into one index. A country can be very strong in judo and very weak in amateur boxing, and the total still looks impressive.

In my coding notebook, every bout in group two carries four fields: discipline, ruleset, number of rounds or minutes, and the ratio of landed to thrown strikes. The last field is the one I use most, because it tells me whether the athlete won by controlling the contest or by exploiting a single moment.

People remember the goal. I remember what led to the goal.

Applied to martial arts: people remember the gold medal. I remember the effective-strike rate and the number of times an athlete was pushed into the corner in round three.

One pattern I have logged fairly consistently across regional games is this: combat disciplines with high technique coefficients tend to produce a wider gap between the leading group and the rest, compared with disciplines where points distribute evenly across rounds. The statistical reason is simple. When a single technique can deliver nearly half a bout's worth of points, the variance of outcomes rises, and high variance means results are harder to predict but also that a technically superior athlete can open a gap faster.

This is information a medal table never displays. A medal table shows who finished first. It does not show who optimised the scoring system better.

System three: binary outcome and finish rate

The third group, MMA, wrestling and grappling systems, follows the strictest logic of all. No visible scoreboard. No rounds accumulated on cards. Only wins, losses and method of finish.

Here the most important metric is the rate of finishing a bout before the limit, combined with the quality of the opponents finished. An athlete at 8-1 with six early finishes against opponents holding winning records cannot be placed beside an athlete at 8-1 whose opponents have never won a bout.

In Vietnam this group is still relatively young in terms of data infrastructure. Domestic professional events have made genuine organisational strides, but public bout-level data has not been standardised. I once spent nearly three weeks reconstructing a single athlete's competitive record from scattered reports, photographs of scorecards and reposted clips. None of those sources had been cross-checked.

That is why I keep one rule: every number that appears in a piece must be verified against at least two independent sources. The rule costs time, but it is the boundary between analysis and guesswork.

First consequence: the medal table as a broken ruler

The medal table adds three systems into a single figure. It is not arithmetically wrong. It is semantically wrong.

A gold medal in a three-person team performance event is the product of a judging panel. A gold medal in a weight-class combat event is the product of three consecutive wins in a single day against three athletes already separated by weight. A gold medal in amateur MMA is the product of controlling one opponent within a fixed time.

Adding those three and calling it martial arts achievement is convenient for headline writers, but it destroys any possibility of learning from the data.

I once tried the reverse. In 2026, when the season was cancelled mid-campaign and I sat down with a database of 242 matches involving the club I was following, I separated performances by scoring group to see how the picture changed. The result confirmed something I had sensed for years: once separated, the ranking of units shifted considerably across events. Some provinces were very strong in the performance group and nearly absent in the combat group, and vice versa. On the combined table, they looked identical.

Before asking who won, ask who kept discipline.

Discipline here means classification discipline. Whoever maintains a clean separation between the three systems is the one who can actually read their own data correctly.

Second consequence: selection and the athletic life cycle

The three systems also generate three different sporting life cycles, and this is the least-discussed part.

Performance athletes peak in a way that depends heavily on registered technical difficulty and psychological consistency. Their careers can be long, because they absorb no direct impact. But they are fragile in the face of joint injuries and major rule changes governing compulsory movement groups. A single rule change can invalidate years of accumulated work.

Judged-combat athletes have life cycles constrained by weight class. Weight cutting is routine, and it is the single largest health risk in the entire field. I have watched athletes walk to the scale with an unsteady gait, then compete twenty-four hours later. No notebook can fully record what happens in the interval between those two moments.

Binary-system athletes carry the highest cumulative head-impact risk. The number of blows to the head across a career is a figure nobody in Vietnam is tracking systematically.

Those three life cycles are currently pooled into a single support regime, a single nutrition regime and a single recovery regime. That is one reason I consistently push back when people talk about athletes needing to prove themselves after injury. For a performance athlete, returning early is mostly a technical question. For a combat athlete, returning early is a neurobiological one.

Contrarian angle: the misunderstanding from outside

There is a very common misunderstanding, and I have encountered it in all three countries where I have worked.

It goes like this: Vietnam's martial arts medal count at regional games is very high, therefore Vietnamese martial arts are very strong; and when world-level events produce no matching results, that proves something is wrong with the athletes or the coaches.

Both halves rest on the same false assumption, that a thing called martial arts ability exists independently of the scoring system.

What actually happens is more complicated. Regional games change their event lists with each edition. Sports are added and removed. Weight-class groups are expanded and trimmed. Host nations that are strong in particular groups actively structure the event list to their advantage.

The consequence is that a country's total regional medal count depends heavily on how many of its strength sports appear on the programme. It is not a linear measure of capability.

This sounds like an excuse. It is not. It is an observation verifiable by the simple act of comparing event lists across two consecutive editions and checking that against medal distribution. I have run that comparison three times, and each time the results showed a significant dependence on event-list structure.

The second misunderstanding concerns media. When a performance athlete posts a high score at a domestic event, the report typically states that the athlete is at continental level. But domestic performance scores and continental performance scores are awarded by different panels, under different rulesets, at different registered difficulty levels. There is no basis for inferring continental level from a domestic number, however high that number is.

The third misunderstanding follows directly from the first two. Because we read the medal table as a linear measure, we allocate resources as a linear measure: the sport with more medals gets more investment. But the cost of producing a medal differs sharply across the three systems. The cost of securing a place in a team performance event is markedly lower than the cost of developing a combat athlete capable of competing internationally over several years.

When we optimise for medal count, we are unconsciously optimising for the lowest cost per unit of medal. That is an administratively rational choice and a professionally wrong one.

There is one more layer, about development structures. Athletes in Vietnam are typically tied to localities, and movement between units is real, visible in the entry lists of national championships. When a promising young athlete moves from a small province to a major centre, the former unit usually has no compensation mechanism matching the cost it invested. I see the same structural problem I track in the football transfer market: the developing side bears the cost, the using side captures the value. And as in football, that structure keeps smaller units permanently in the business of producing semi-finished goods.

Every season is a chapter. I am only the one holding the bookmark.

For Vietnamese martial arts, the book is still being written in three parallel counting systems, and the bookmark holder has not yet agreed which system to number the chapters by.

Takeaway: internal signals to watch

There are three signals I will be logging over the next twelve months.

First, the standardisation of bout-level data in the combat group. I do not need a beautiful scorecard. I need a public database with absolute dates, opponent names and method of finish. When that exists, finish rate and opponent quality become comparable metrics. If it does not exist by the end of next season, I will hold my conclusion: every assessment of this group is an estimate reconstructed from reposted clips.

Second, the separation of scoring systems in performance reporting. A report split into three sections carries far more information value than a single combined table. This requires no money, only process. If a federation publishes an annual report in three separate blocks, that is a signal it understands the classification problem.

Third, athlete safety data in the binary group. Head-impact counts, mandatory minimum rest days after a stoppage, and pre-licence medical screening. This is a category where absence of data does not mean absence of risk. It only means we are not measuring.

The bout ends, but the data does not.

Three systems will keep running in three different arenas next season. I will still carry two notebooks. One for what I see on the floor, and one for what I need to verify before I write. The only way to read a field made of three different measurements is to refuse to add them together, even though adding them is always easier and always yields a nicer-looking number.