The Data Gap in Vietnamese Swimming: What the Medal Table Cannot Tell You

A results file with eight lines

The results file I received after the men's 400m individual medley final at a national meet contained exactly eight lines. Each line held a name, a bib number, a finishing time and an empty notes field. The whole file was under two kilobytes. No reaction time. No 50m splits. No turn times. No stroke counts.

A swimmer covering 400 metres of medley in nearly five minutes performs hundreds of strokes and seven turns. The data trail that produces runs into thousands of points. The only thing kept was a string of characters: 4:28.xx.

Set that beside the official results page for the men's 100m freestyle final at Paris 2026, published by World Aquatics: reaction time, the 15-metre mark, four 50m splits, turn times. Pan Zhanle touched in 46.40 seconds and broke the world record, and every segment of that swim can be reopened and checked.

The gap between the two files is not a matter of technology. Timing systems in domestic pools are perfectly capable of recording splits. The gap is that nobody asks for them.

Data never lies, but it knows how to hide itself. In swimming it does not hide. It simply does not exist.

A sport with no public data layer

I have spent years working with football data, where a match from seven years ago can be reopened to read every pass, every pressing action, every run. Swimming should be easier. A lane is a closed environment: 50 metres, eight lanes, times measured to the hundredth, no deflections off the post, no referee judgement calls at the boundary. A sport with fewer variables has less public data. That sounds absurd until you follow the money.

Data is not free at the collection stage, but it is free at the demand stage. In football, broadcast contracts and betting markets create permanent demand for numbers, so analysis companies pay people to log every action. In swimming, that demand appears for a few weeks every four years, when the Olympics arrive. For the rest of the cycle, nobody pays to know that a swimmer lost 0.4 seconds on the fifth turn.

The result is a data layer so thin it barely exists. Swimming coverage in Vietnam, including my own, usually works with three things: finishing time, placing, and medal count. All three are the final product of a long process for which no record survives.

When COVID shut the stadiums, I reopened the V-League directory. No league is meaningless. It took me four months of tracking 240 players to rebuild physical data, and the biggest lesson was not identifying who was declining. The lesson was this: when data is absent, people default to the prettiest story available. In Vietnamese swimming, that data layer was never built in the first place.

Anatomy of a swim

What belongs in a complete swimming dataset? Reaction time off the blocks, measured in thousandths of a second from the signal. The 15-metre mark, which shows how efficient the underwater phase was. Splits at every 50 metres. Stroke count and stroke rate. Distance per stroke. Turn times. And a fatigue index, calculated as the difference between the final 100 metres and the opening 100 metres.

The central formula of the sport has two variables: speed equals stroke rate multiplied by distance per stroke. A swimmer can go faster by stroking more often, by stroking further, or both. Without those two numbers, every technical claim is speculation from outside the lane.

World Aquatics rules limit underwater swimming after the start and after each turn to 15 metres. That is one of the clearest boundaries in the sport, and it is also where plenty of swimmers lose time without anyone seeing it. A turn that is 0.3 seconds slow, multiplied across seven turns in a 400m medley, costs more than two seconds. On the results sheet, those two seconds appear only as a gap between two names.

In the 400m individual medley, the race divides into four blocks: butterfly, backstroke, breaststroke, freestyle. Each has its own logic. The breaststroke block usually produces the widest spread because the rhythm is slow and depends on the glide phase. With 50-metre splits, I can say precisely which block a swimmer lost the race in, and whether they lost it to technique or to pacing. Without splits, I know only the total, and the total is a spectator's number, not an analyst's number.

In distance events such as the 1500m freestyle, the pacing curve is the story. A swimmer who holds an even pace and a swimmer who empties the tank over the first 800 metres and dies over the last 300 can finish in identical times. Those are two completely different races physiologically and tactically, yet on the results sheet they are the same line of text.

In sprint events such as the 100m freestyle, the whole race lives between second 22 and second 47. Reaction time, the underwater distance inside the first 15 metres, and speed over the final 25 metres decide almost everything. Pan Zhanle touched in 46.40 seconds at Paris 2026, and the analytical value of that number comes from the four splits published alongside it, which allow the first half to be compared with the second. A world record can be interrogated using its own data. A national gold medal cannot, because there is nothing to interrogate.

Léon Marchand won four individual gold medals at Paris 2026 in the 200m butterfly, 200m breaststroke, 200m medley and 400m medley. Read through the medal table, that is the story of a dominant athlete. Read through the splits, it is the story of a training group that calculated its energy distribution precisely across four events crammed into a single week. Two readings, two levels of understanding.

Stroke rate and the trap of the effort number

In football, distance covered and sprint count are packaged as effort metrics. Running without purpose still produces handsome numbers, and people keep forgetting that. Swimming has an equivalent.

A swimmer stroking 52 times a minute looks more committed than one stroking 38 times a minute. Stroke rate on its own will always favour the faster churn. But speed is the product of rate and distance per stroke. The swimmer who churns fast with a short stroke is actually less efficient. Publishing stroke rate without distance per stroke is an actively harmful editorial habit, because it teaches the audience to look in the wrong place.

This matters for Vietnamese swimming in one specific way. Young swimmers are usually judged on results at 15 or 16, when distance per stroke is still short because height and arm span have not finished developing. A swimmer powering through the 100m freestyle with a high stroke rate can win at 15 and lose the position at 19, once their peers grow their stroke length. Without a dataset tracking both variables across seasons, we cannot separate potential from puberty.

Based on my experience covering swim meets, for years I had to reconstruct splits by hand from video at 25 frames per second, and the cumulative error of that method reaches nearly half a second per 50 metres. Half a second is the distance between a medal and fourth place at regional level. In other words, even when I did the work of rebuilding the data, I still did not have enough reliability to conclude anything.

A champion's squad is not built in the wallet, but in the way time gets compressed into metrics. Vietnamese swimming has not managed that compression. Nguyen Thi Anh Vien is one of the most decorated athletes in Vietnamese sport at the SEA Games, with a medal collection stretching across many years. Nguyen Huy Hoang won silver at the 2026 Asian Games in the men's 1500m freestyle. Tran Hung Nguyen and Pham Thanh Bao are names that have established themselves at regional level. We know their names, we know their medals, and we know almost nothing about the shape of the swims that produced them.

A swimmer's elite career lasts roughly ten years. When that curve turns downward, no dataset exists to answer whether the cause was age, injury, the training environment, or a change of coaching staff. A retired footballer still has match footage to move into commentary, analysis or coaching. A retired swimmer has only collective memory, and collective memory fades quickly.

The counterintuitive part

The medal table is the most misleading document in the entire sports information system. It is presented as a neutral record, but it is an editorial product. Who decided medals should be ranked by delegation? Who decided only the top three should be listed? Who decided the personal best of the eighth-place finisher should be dropped? Every one of those choices shapes what the public believes happened at a championship.

The comfortable explanation for the data shortage is to blame the press. I do not buy it. Reporters write what their newsroom pays them to write, and newsrooms sell what the audience reads. The problem sits in the incentive structure: nobody in that chain benefits from publishing splits, so splits do not get published. Changing that behaviour requires someone to pay the cost upfront, and the only parties able to do so are the federation and the network of training centres.

People look at the price tag; I look at the curve. Plenty of deals die before they are announced. In swimming, an athlete's commercial value is priced off medal counts, which means priced off past results. Pricing off past results works well in a stable environment and fails in a fast-changing one. A sponsor signing the swimmer who just won SEA Games gold is buying a medal, and the numbers show that medal predicts very little about the following two years unless an improvement curve comes with it.

There is another trap I have to keep reminding myself about. A swimmer races only three or four times a year in their main event, once per meet. Four data points a year is far too few to separate signal from noise. I set the threshold in advance: I only draw trend conclusions after at least three consecutive seasons moving in the same direction, and only when the margin exceeds measurement error. Electronic timing error at international meets runs to 0.01 seconds, while the error of inferring from four swims a year can be dozens of times larger.

There is one thing I should say plainly about my own work. When I re-ran the analysis pipeline for this topic, the input extraction step returned an empty result. No information points, no entities, no subject to analyse. At first I treated it as a system failure. After checking again, I understood: the empty result was the finding. A pipeline designed to analyse swimming, run against publicly available swimming data, could not find anything to analyse.

Signals for the next cycle

There are four things I will be watching over the next 18 months, and all four are verifiable signals rather than gut predictions.

The first is the appearance of splits at national championships. If a single organiser prints four extra columns into the results file, the entire analytical baseline behind it shifts within a season. The marginal cost is close to zero, because the data already exists inside the timing system.

The second is a results database searchable by athlete and by season. Right now, rebuilding the career trajectory of a young swimmer means stitching information out of dozens of articles, most of which list only the final time. A proper database turns that from weeks of work into minutes.

The third is a career transition programme for athletes after retirement. The competitive lifespan of an esports professional is shorter than a footballer's, yet their youth development and post-retirement support systems are close to nonexistent. Swimming shares that weakness, with a variation: an elite swimmer's peak window is even narrower, and with no recorded data from their lane there is no raw material for a later move into coaching or performance analysis.

The fourth is the shape of the curve among the younger cohort. A national record at 16 means little if distance per stroke does not increase across the next two seasons. That is the single indicator, in my view, that can forecast where Vietnamese swimming lands in the regional order during the coming cycle.

Luck is something I do not have. I have probability and data thick enough to work with. But probability can only be calculated when there is a sample, and a sample only exists when somebody agrees to write it down. The question for the next cycle is not how many Vietnamese swimmers are good enough for continental level. The question is whether, by the end of that cycle, we will hold enough data to answer the first question accurately.