Vietnamese Swimming and the Empty Data Framework of the Annual Season

In the national swimming team's roster for the coming annual season, there is a gap few people notice: most athletes are still judged by medals, not by process data. I once sat in the stands at a national swimming championship, timing each lap myself to record split times, and realised that behind the lane, the coaching staff held only a single sheet of final results. No stroke rate, no cadence count, no heart-rate data per 50 metres. After a week of competition, all that remained was a string of times squeezed onto one page. That is why I always begin any swimming analysis with a simple question: what are we measuring, and what are we missing? In a sport decided by hundredths of a second, the answer to the second question usually matters more than the first. A season measured in medals Vietnamese swimming has for years operated on a results-first model. National sports training centres, gifted-athlete schools and a handful of private clubs together raise the athletes, but most resources are poured into preparing for SEA Games, ASIAD and Olympic qualifiers. At that level, the measure of success is medals and qualifying slots, not the quality of accumulated data. Nguyen Thi Anh Vien is the clearest example of both the achievement and the limit of this model. For nearly a decade she was Vietnam's most reliable source of regional medals, with a SEA Games gold count among the highest in the country's sporting history. When she retired, the gap she left was not merely one competition slot. It was the gap of a system that had never built a data layer thick enough to transfer experience from one generation to the next. Later, Nguyen Huy Hoang emerged as a pillar in the distance events, tied to SEA Games medals and one podium finish at the ASIAD. Tran Hung Nguyen, Vo Thi My Tien and a few other young faces continued to represent the next generation. The list grew longer, but the structure of results barely changed: a few individuals account for most of the medals, while the rest play a supporting role. That is a predictable structure, and precisely because it is predictable, it is no longer good news. The talent conversion chain: where data gets dropped I usually reconstruct a swimmer's journey as five links: talent identification, provincial youth training, the national youth team, the national team, and the international stage. At each link, a decent data system should record at least four groups of metrics: time and splits, technical parameters including stroke rate and distance per stroke, physiological parameters including heart rate and lactate concentration, and training load broken down by intensity threshold. In Vietnam, most data is recorded only at the final link, and usually consists of a single column of times. The other three groups almost vanish. The consequence is that when an athlete improves or declines, the coaching staff can only explain it with unmeasurable words: effort, spirit, form. Those words are not wrong, but they do not help forecast anything. Imagine two swimmers finishing a middle-distance race with times a few hundredths apart. The first swims with a high stroke rate and a short distance per stroke, an energy distribution prone to collapse over the final 50 metres. The second swims with a low stroke rate and a long distance per stroke, more efficient but less able to accelerate suddenly. Same result, opposite futures. With only a column of times, a selector cannot tell who will go further over a longer distance, and who should be moved to a sprint event. In swimming, the stroke index, the product of stroke rate and distance per stroke, is the cheapest and most useful tool any centre can measure with one camera and one frame-counting application. Yet it rarely appears in national-level training reports. Operating conditions are another blank variable. The same athlete, with the same training load, can produce different results in a 25-metre pool and a 50-metre pool, simply because the number of turns differs. Converting a short-course time to a long-course time is not a simple subtraction, but a problem that depends on each swimmer's turn quality and underwater phase. A centre that does not record the first 15 metres and the last 15 metres of every lap will never know whether its athlete is losing time at the turn or in the closing sprint. The annual season also poses another test: patience. No single meet decides a swimmer's fate in one evening. Progress in swimming unfolds over many months, where one hundredth of a second may demand a large block of training to win. For that reason, mistakes in data recording do not produce immediate consequences. They accumulate quietly, like an unexpressed shoulder injury, until a season passes and nobody knows why an athlete has stood still. At the pool, controversy rarely erupts the way it does on a football pitch, but it exists in more discreet forms: a penalised start, a touchpad sensor that fails to register, a record confirmation delayed by missing video data. Such incidents rarely make headlines, yet they are the technical reasons a result becomes doubtful. The neglected part: load and injury data Another area is equally empty. Swimming is a low-impact sport but by no means a low-overload one: swimmer's shoulder, breaststroker's knee, butterfly lower back. These injuries do not appear in a single session. They accumulate over weeks, and the warning signal usually lies in load data before it lies in pain. During the period when football had to pause because of the pandemic, I reviewed GPS data for nearly thirty players at a club in Saigon and found a familiar pattern: high-speed running distance spiked in the roughly two weeks before a muscle injury appeared. Swimming has no running GPS, but it has equivalents: anaerobic-threshold volume, sprint repetitions, total metres per week. Without measuring those, we only detect an injury once it has happened, and we call it bad luck. A single lap appears once. Its trajectory stretches over many years. The counter-view: more data does not automatically produce medals Here I have to argue against myself, because that is the only way not to fool myself. Suppose we build a complete data system: split times for every lap, stroke index for every session, heart-rate variability for every athlete. Will the medal count rise? Not necessarily. First, correlation is not causation. If regional medals increase over the next few years at the same time as investment in data, that does not prove data produces medals. Both may rise because of a third variable: funding, or a particularly gifted cohort, or a change in selection methods. To say that data leads to performance, I must point to a specific mechanism, for example that early detection of overload reduces the number of weeks of interrupted training, and that consecutive training weeks correlate with time improvement. Without that mechanism, I am only permitted to say the two things are related. Second, imported data easily creates false precision. Training models from Europe and North America were built for systems with widespread 50-metre pools, dense competition calendars and a complete school-sports base. Vietnam has a tropical climate, a limited number of standard pools, and a selection system tied to sports gifted-athlete schools. Applying a foreign model wholesale without recalibration produces beautiful numbers in reports and wrong decisions on the pool deck. Third, there is a share of variance that numbers cannot explain. I sit far from the lane to see the race more clearly than the referee, but I cannot measure what happens inside the head of a 17-year-old stepping onto the starting block at a first international meet. At those moments, I have to note clearly that my judgement carries a wide confidence interval, rather than pretend the model has covered everything. Every shock has its own probability. We call it a shock when we have not yet checked the table. What is worth watching this season The signal I am waiting for is not another medal. It is a national split-time database, updated at every meet, open to coaches and athletes alike. When a young swimmer can look back at their own third lap from a meet two years ago, they are more likely to understand what has changed. When a coach can compare a trainee's stroke index against a regional benchmark, they are more likely to choose the right event. A generation of athletes fades when the table of its numbers no longer has any readers. Vietnamese swimming has produced individuals capable enough to keep the sport on the regional map. The question for this annual season is not who will shine next, but whether the next one will inherit a data framework thick enough, or start again from a blank page.