The Blank Cell in Vietnamese Swimming Data

The spreadsheet I opened on Saturday night had 1,842 rows. The final-time column was complete: one swimmer, one event, one meet per row. The split-time column was entirely empty. All 1,842 cells carried the same value: N/A. In my system, N/A stands for "insufficient information — cannot assess." It is an analytical result, not a surrender. I use it when I am forced to admit I do not know, rather than filling the gap with a judgement that merely sounds plausible. From the men's 200m freestyle final at a recent national championship, I have the finishing times of all eight lanes, accurate to the hundredth of a second. I do not have the four 50m splits. I know who won. I do not know why he won. That is where this season's problem begins. Swimming is the best-measured sport in the entire competitive system. The result is a real number. Electronic touchpad timing records to a hundredth of a second. The international federation archives results by year, by distance, by course. No sport starts with a bigger data advantage. That advantage stops at the border of the international system. When a Vietnamese swimmer competes abroad, the data exists in full inside the organiser's system: splits, reaction time off the blocks, turn times. When that swimmer comes home to race domestically, the record disappears. The result survives as an A3 scoresheet taped to a corridor wall, a photograph posted on a personal page, or a scanned PDF with no text layer for a machine to read. Based on years of tracking domestic and international swimming result sheets, I estimate that most lanes at national meets leave behind no split record at all once the meet ends. Heats vanish first, finals second, and about three months after competition day what remains is a single final time with no context attached. There is one universal principle anyone working with swimming data must know: the 25m short course and the 50m long course are two different performance spaces and cannot be compared directly. Over 200m, a 25m pool gives a swimmer 8 lengths and 7 turns; a 50m pool gives only 4 lengths and 3 turns. Every turn carries a push off the wall, and at elite level a good turn saves half a second or more. For events from 200m upward, short-course times are typically one to two percent faster than long course. Over 50m, the comparison can flip, because the long course has no turn at all. Vietnam has very few competition-standard 50m pools. Most of the athlete pool trains daily in 25m water, and many domestic meets are held in 25m pools. The domestic performance curve and the international performance curve of the same swimmer are therefore drawn on two different axes, and no conversion factor is accurate enough to join them. The Nguyễn Thị Ánh Viên generation left behind a complete international record. Born in 2026 in Cần Thơ, she swam the London 2026 Olympics before turning 16, spent a long training period at the United States Olympic Training Center in Colorado Springs, and won more than twenty SEA Games gold medals. That was a stroke of luck for the data. When she left the international stage, most of Vietnamese swimming's record fell back to the domestic level, where splits are not kept. Nguyễn Huy Hoàng is the next case with a thick international record, having taken bronze in the men's 1500m freestyle at the 2026 Asian Games according to the organiser's official results in Jakarta-Palembang. Hoàng Quý Phước, Phạm Thanh Bảo and Trần Hưng Nguyên belong to the following group. Beneath them, in the 14-to-18 age band, I have no way to reconstruct a full development curve, because nobody kept the splits for that age group. If I had to pick a single smallest unit of swimming data, I would pick the split, not the final time. The final time answers one question: who finished first. The split answers the question a coach actually needs: how is this swimmer distributing speed, and is that distribution sustainable. Take two swimmers who both finish a 200m freestyle in 1 minute 52 seconds. The first opens in 25.5 and fades evenly, closing in 30.2. The second opens in 26.8 and builds, closing in 27.6. On the scoresheet the two lines are identical to the hundredth. In a sporting record, the two trajectories run in opposite directions. The second swimmer has a better aerobic base and better pace control. The first has higher raw speed but is spending above threshold in the first half. At 16, the second is a long-term asset. At 16, the first is a question that needs two more seasons before it can be answered. With final times alone, both read 1:52.0. I have no way to tell them apart. A result sheet compressed into one line has erased its own analytical value. Every shock has its own probability. We call it a shock only when we have not yet checked the table. Example: a young swimmer suddenly breaks the national record in the 400m individual medley. My first question is always which course the new record was set in, and which course the old record was set in. If the new mark came in a 25m pool and the old one in a 50m pool, most of the "leap" sits in the number of turns, not in fitness. If both came in the same course, that is when I start calculating. But to calculate, I need splits. And I do not have them. The next indicators are stroke rate and distance per stroke. The product of the two gives average speed; to swim faster, a swimmer must raise one or both. In the data I have gathered at domestic meets — mostly by direct observation and hand counting, because there is no measuring equipment — the stroke rate of Vietnamese swimmers tends to sit below the international baseline, while distance per stroke is not bad. That points to a sound endurance base and a shortfall in explosive power. I will state plainly: this is an observation with no calibrated reference data behind it, and it should be read only as a hypothesis to be tested with optical data in both courses. The data lost most completely is the heats. Finals scoresheets print eight lanes; heats are usually read over a loudspeaker or taped up and taken down. The swimmer who placed ninth in the heats — the one who just missed the final — does not exist in any record I can trace. Yet the group from ninth to sixteenth is exactly what signals the flow of an age cohort. A national team is not built from eight people; it is built from forty, of whom eight rise. The follow-on consequence is that I cannot establish the peak age of a Vietnamese swimmer. In swimming, women typically peak earlier than men, and puberty is a technical barrier: the arm span lengthens, the body's centre of mass shifts, and the feel for the water has to be rebuilt from scratch. For men, the peak usually arrives later and can extend into the late twenties. To test any of this on Vietnamese data, I need an eight-year time series for the same group of swimmers. I have less than a fifth of the data required. Swimming is the sport with the strongest analytical tools and the weakest application of them. I once built an expected-goals model for Vietnam's U20 side at the 2026 U20 World Cup. That model had error. Every shot was assigned a probability, and the probability was wrong, forcing constant recalibration. Swimming does not have that problem: the touch on the wall is fact, not estimate. A model built on correct swimming data would be more accurate than any football model of the same tier. But it only runs if the data exists, and Vietnamese swimming is missing precisely the cheapest data there is to collect. A shot happens once. Its trajectory lasts for years. In swimming, the equivalent sentence is: a single touch of the wall lasts one hundredth of a second. But four touches of the wall, fully recorded, repeated across eight seasons, show me a career that is climbing or one that has gone flat. The first counter-intuitive angle: publishing more data can cause harm if it is published without context. If a federation releases splits without operating conditions attached, the public will read every split as a verdict. A swimmer who closes three seconds slower than he opened will be called spent. But if that swimmer raced another event twenty-five minutes earlier, or is in a deload week after a heavy training block, or has just come through a fever, that closing split does not measure sporting quality. It measures the schedule. Operating conditions are input parameters, not footnotes. I always require four pieces of information alongside the data: course type, time of day, the swimmer's event sequence within the session, and days since the previous competition. Without those four, a split is just a meaningless string of characters. The second counter-intuitive angle concerns infrastructure. A correlation is easy to spot: Vietnamese swimming's SEA Games results rose during a period when more standard pools were built, then levelled off. But correlation is not causation. To claim that 50m pools produce medals, I must point to the specific physical and behavioural mechanism: more long-course pools lead to more long-course meets, racing long course more often leads to a habit of pace distribution without turns, and that habit leads to better international results in exactly the events from 200m upward. That is a three-step chain, and every step can break. Until a step is verified, the honest way to write it is: the two variables are related. The third counter-intuitive angle: there is a share of variance that data will never explain, and that share involves the stands. When the stands fall silent, home advantage dissolves into a number close to zero. In 2026, when competitions were suspended and training pools closed, I reviewed the load data of twenty-nine athletes at a club in Saigon. High-speed running distance rose by roughly twenty percent in the period before muscle injuries appeared. We re-split training into four load thresholds; the following season, muscle injury cases fell by about thirty percent against the previous season. With no crowd and no shouting, athletes still had to absorb volume, but their bodies were not given rest at the right rhythm. In swimming, crowd noise has a real physical effect. Over the final fifteen metres of a lane, shouting changes breathing rhythm and stroke rhythm. It is a variable. When that variable is switched off, a result can no longer be compared with its own previous season. Every 2026 performance record needs its own annotation: no spectators. And one more blind spot: Vietnamese national swimming records are recognised in the 50m course, while most of the training population works in 25m water. The confidence interval between those two curves is far wider than any ranking table can accept. Over the next twenty-four months, I am watching three signals. The first is the share of domestic meets that publish official splits with the course type attached. If that share passes half, athlete development models become viable again, and we can talk about talent with data instead of with feeling. The second is the number of 50m pools entering operation and the number of long-course meets held in them. This is a slow signal, measured in years, but it is the only signal that changes the axis of international comparison. The third is the 200m freestyle time curve of the under-16 group, measured long course, sampled every six months. If that curve slopes upward at the third and fourth splits rather than the first, I will know the aerobic base is being built in the right place. I sit far from the lane so that I can see a single swim more clearly than the referee does. Most people look at a results table to understand a meet. I look at a results table to understand ten years. If four seasons from now I still have to open a spreadsheet with an N/A cell, then the problem with Vietnamese swimming will not be in the lane.