Vietnam's Swimming Data Gap: Every Medal Comes With an Empty Scoreboard

In June 2026, at the SEA Games in Singapore, Nguyen Thi Anh Vien walked onto the starting blocks eight times and won eight gold medals, setting eight Games records. I was in Saigon, logging each swim into a spreadsheet. On the fourth day of competition I noticed my file held only finishing times. No 50-metre splits. No stroke rate. No underwater kick count. No turn times. Eight gold rows, and an empty explanation column.

Eleven years later I still open that file whenever someone asks where Vietnam's next swimming generation is heading. I can only answer in probabilities, and I know those probabilities rest on sand, because the underlying record does not exist. That is the central problem of this sport in Vietnam, and it is not underwater.

I entered the trade in March 2026, on the sports desk of a Saigon newspaper, covering swimming. Twenty years later I sit on the other side of the profession, building data models for football teams. In between came the 2026 Under-20 World Cup and an xG analysis of Vietnam's youth side, the 2026 World Cup and a passing-chain model of the eventual champion, and 2026, when a load-management algorithm I designed was applied to 29 players at a Saigon club.

Returning to the pool with a football toolkit produced a dry paradox: swimming is the most measurable sport of all, and in Vietnam it is the most thinly archived. No sport records its outcome more precisely, down to a hundredth of a second, and no sport leaves the process behind that outcome so poorly documented.

Vietnamese swimming data lives in three disconnected places. The electronic timing systems at senior national meets produce printed or exported result sheets that almost never include public 50-metre splits. Coaches' hand timers at junior meets carry an error between two presses on the same swim of up to three tenths of a second, exactly the margin between gold and fourth in a 50-metre race. And the memory of the people in the trade, myself included.

At international level the picture is different. World Aquatics' database stores split times, seasonal rankings and a conversion points table that allows results across distances and eras to be compared. The Vietnam Aquatic Sports Association has both the authority and the technical resources to run something similar nationally. Nobody has treated it as mandatory.

To see the cost, break a 50-metre race into variables. Total time equals reaction off the blocks, plus underwater time, plus surface swimming divided into stroke cycles, plus turns in longer events, plus the finish. Each variable is separately measurable. Stroke rate and distance per stroke are independent inputs, and their product produces speed. Speed is the output of a two-variable function, and if you only record the output, you cannot fix the function.

The physical mechanism is specific, not an inference from statistics. Drag rises with the square of velocity. When a swimmer lifts stroke rate without holding distance per stroke, the added drag exceeds the added propulsion, and the finishing time slows even though the water feels faster. This only becomes visible with two parallel data series. Domestic meets give us one: the finishing time.

Based on my experience following domestic meets for two decades, most technical conversations in Vietnam stay at the level of sensation: my arm felt heavy, I was short of breath over the last 25 metres. Those descriptions are accurate, but they do not accumulate into a database that can forecast.

Now the more uncomfortable part: the structure of Vietnam's medal output.

In my tracking, across the SEA Games cycles from 2026 to 2026, more than sixty per cent of Vietnam's swimming gold medals came from no more than two athletes. On the men's side the gold supply concentrates in a very thin group: Nguyen Huy Hoang, who raced the 800-metre freestyle at the Tokyo 2026 Olympics, alongside Tran Hung Nguyen in the medley events and Pham Thanh Bao in the sprints. A generation earlier, Hoang Quy Phuoc carried the 100-metre butterfly. On the women's side, gold depended almost entirely on Nguyen Thi Anh Vien for a decade.

That concentration should be read as a risk index. When one source supplies more than half the total, the variance of the total equals the variance of that source. The next SEA Games will not depend on the quality of the system; it will depend on whether one person can swim.

When Anh Vien reduced her senior racing schedule and moved into the later phase of her career, the women's gold column thinned faster than any junior results table could fill it. That was not a surprise. It was a thin-tailed distribution, and thin tails always snap first.

Every shock has its own probability. We call it a shock only when we have not checked the table.

The next part carries a lesson from a different sport.

In 2026, when competition stopped, I reviewed GPS data for 29 players at a Saigon club. A pattern repeated: high-speed running distance rose by roughly twenty per cent in the period before a muscle injury. The cause was not running more; it was the slope of the load curve exceeding what the tissue could absorb. I proposed splitting training volume into four load thresholds. The following season, muscle injuries fell by about thirty per cent against the previous one.

That mechanism transfers to swimming, though not by strapping a GPS unit to a swimmer's back. The equivalent variables in a pool are a drop in distance per stroke at a fixed stroke rate, and rising variance in turn times. Both appear weeks before shoulder damage. The shoulder is where swimming pays its bills, and in Vietnam we measure neither signal.

So injuries arrive as shocks. A swimmer is at peak form, then one morning cannot lift an arm overhead. Nobody saw it coming, because nobody logged the curve before it.

I propose a framework I call the hydrodynamic profile: four thresholds tracked quarterly for each swimmer from age ten to retirement. The technical threshold covers stroke rate, distance per stroke, turn time and underwater time. The load threshold covers weekly volume and the gradient of its increase. The physiological threshold covers lactate after standard sets and heart-rate variability. The competition threshold covers the gap between training and racing times, in heats and in finals.

None of these require expensive technology. They require one rule: every swimmer carries a single identifier, and every measurement attaches to that identifier rather than to a meet.

There is one more place where Vietnamese data picks the wrong people.

Between ages eleven and thirteen, selection relies on the national youth results table. That is the single most maturation-dependent measurement in sport. A child who matures early has a longer arm span, greater height and larger lung volume, and wins at twelve. After puberty the advantage compresses as the others catch up. Select on the twelve-year-old results table and you systematically choose the swimmers who will be caught at seventeen.

The mechanism here is biological, not statistical. The anthropometric edge from early maturation is temporary, and any selection system that does not adjust for it will keep losing the athletes with longer development curves. European federations use two tools for this: the World Aquatics conversion points table, and predicted adult height calculated from parental height. In Vietnam, I have not seen either applied widely.

I sit far from the pool so that I can see the lane more clearly than the referee.

Now the part I consider most important, and it argues against my own proposal.

More data does not mean better decisions. Vietnamese swimming does not lack stopwatches. Every coach at every pool owns one and presses it thousands of times a week. What is missing are stopwatches wired to each other. An isolated number creates no knowledge; a series of numbers with an identity key does.

Even with a series, I have to state the limit clearly: correlation does not guarantee causation. If training volume and medal counts rise in the same year, I cannot conclude that one produced the other. To claim load leads to medals, I must identify the physical and physiological mechanism linking the two variables, and that mechanism must be testable against a control group. At the scale of one national team, the sample is so small that most relationships fail any reliability test.

I also have to concede variance my model cannot explain. SEA Games 31, held in Hanoi in May 2026, is the example. The stands were full, the noise was enormous, and some swimmers went under their own best training times while others were two seconds slower than in the heats that same day. No variable in my table captures that. In my model the crowd is a switch that can be turned on or off; in reality the crowd is a share of variance that cannot be quantified.

And I have to say something the Vietnamese swimming community may not like: Anh Vien's formula does not exist. She is an outstanding data point in a distribution, not a mould for casting the next one. Any effort to reproduce a specific individual copies the past rather than forecasting the future.

A tactical era dies when nobody reads its data table any more.

So what are the signals for the next cycle?

If within the next two seasons the Vietnam Aquatic Sports Association publishes 50-metre split data for every national final, that is a stronger signal than any medal. If a training centre begins using the World Aquatics conversion points table to compare results across distances, the door to forecasting opens. And if a young swimmer appears whose times at fifteen are unremarkable but whose quarterly improvement curve is steady, that athlete is more likely to last at the top than the junior champion.

Most people read a results table to understand a SEA Games. I read the lane to understand ten years.

The probability I assign to my own forecast: sixty per cent that Vietnam will have at least one open national swimming dataset within three years, provided an independent working group does it rather than waiting for instructions. The remaining forty per cent covers what I cannot measure: funding, relationships, and the inertia of a system still watching medals instead of curves.