The Blank Page in Vietnam's Swimming Records

23:47. The extraction file opens: nine header rows, nine empty value cells. Article title: none. Source: none. Information points: not one item. The line labelled athletes instructs me to identify them from the information points above, but there is nothing above to identify them from.

I sat there another ten minutes before typing a word. Outside, Nha Trang had turned off its lights. In my head another morning surfaced: five o'clock, a training-centre pool, chlorine in the air, water slapping the wall, a coach clicking a stopwatch on every 50-metre repeat. That place had raw data, sweat, real people. On the screen sat a blank page, perfectly formatted — every field present, every section in place, nothing inside.

The distance between those two images is the whole story of this piece.

Vietnamese swimming is a sport of small sample sizes. Nguyễn Thị Ánh Viên, from Cần Thơ, took bronze in the women's 400-metre individual medley at the 2026 Asian Games in Incheon, Vietnam's first Asian Games swimming medal. Four years later in Jakarta, Nguyễn Huy Hoàng, from Quảng Bình, took bronze in the 800-metre freestyle. At Tokyo 2026, Huy Hoàng reached the final of that same event. Three milestones spread across seven years, and all three stood on a data foundation far thinner than that of neighbouring swimming nations.

The thinness is not a shortage of athletes. It shows up every time someone has to assess a single lane and must rebuild a nine-dimension frame from scratch: technique (start, underwater segment, turns, finish, stroke efficiency), performance data (position against the world record, all-time list, season ranking), competition system (event tier, selection mechanism, cycle), the world map by event, the rules and anti-doping framework, the athlete's career trajectory, the risk profile, the media narrative and expectations, and finally the ripple into the industry.

Those nine dimensions serve very concrete questions that coaches and team leaders actually need answered: is this swim repeatable, is the improvement physiologically plausible, is this meet worth trading a training cycle for, and where does the swimmer's shoulder-injury risk currently sit.

When the file comes back empty, all nine questions come back empty with it.

I re-ran the extraction three times. The result did not change. Nine dimensions, thirty-seven cells, all carrying the same value: insufficient information. For someone trained to cross-check, this is the moment to stop and read the emptiness itself.

An empty data field is itself information — information about the process that produced it. The void tells me three things a full table of numbers never would: the source-fetch step failed, the analysis step was blocked, and nobody along the chain noticed. The frame still looked handsome. The format was still complete. Only the content had vanished, and the vanishing was utterly silent.

That is the kind of error I once made myself, at a far smaller scale. In 2026, the only female data analyst in the technical analysis room at Sanna Khánh Hòa BVN, I miscalculated a striker's sprint distance against Hà Nội FC: 1.2 km instead of 0.8 km. A specialist in the room said plainly that women do not understand tactics. After the match I audited all 14,000 GPS samples the club had collected over three months and found three more systematic errors coming from the synchronisation software. A small GPS deviation is enough to teach you: verification is everything.

That lesson transfers to swimming almost intact. Swimming is a sport that splits data ruthlessly: reaction time, the underwater segment after the start, stroke count per length, average distance per stroke, turn time, speed over the final 15 metres. Every one of those cells can be empty for a very specific technical reason — a sensor out of sync, a camera at the wrong angle, a referee hand-timing in place of electronic equipment. When a cell is empty, an inexperienced reader fills it with feeling: probably a slow start, probably fading at the end. I do not do that. I write unknown and leave it.

In 2026, sent to support data analysis for the national sports channel during the World Cup in Russia, I collected expected-goals figures for all 64 matches. Croatia reached the final but in the knockout rounds created only 5.3 xG, while Denmark, Russia, England and France combined for 7.1 xG. Croatia scored eight goals from 5.3 xG, an overperformance of 51 percent. Croatia 2026 was no miracle — it was xG written into history. That 2,000-word analysis was one of the first Vietnamese-language pieces on xG.

The pandemic season of 2026 taught me to measure a league by recovery indices, not by points. With the season suspended from March to September, I built a model on GPS data from 365 V.League players across three seasons, 2026 to 2026, combining high-intensity distance above 25 km/h, acceleration counts and injury history. When the league returned, I predicted that the three teams pressing hardest would see injury risk rise 23 percent. My club cut training load by 15 percent and lost no key player.

All three stories end with a confidence column in the spreadsheet. All three begin with refusing to fill an empty cell with intuition.

Data does not tell stories; it records everything so that I can tell them. But when data records nothing at all, the only honest move left is to tell the story of where it went silent.

The industry reflex is to fill. An empty file is instantly converted into a headline: Vietnamese swimming lacks data. That conclusion sounds reasonable, travels well, and is nearly useless, because it merges three different problems into one sentence: missing sources, missing extraction process, missing reviewer. Only the first needs money. The other two need discipline. And the price of filling a gap wrongly is usually paid by the athlete.

There is an attractive correlation I have to refuse: strong swimming nations all have thick data systems, so it is easy to infer that thick data produces medals. That correlation is not causation. Thick data is a by-product of a development system with enough time, money and people; it does not create swimmers on its own. Read backwards, it would have us buy software before we buy pools and coaches.

I also have to audit myself. The nine-dimension model I use is a frame written by humans, and it may be the very thing that quietly swallowed the data: a required field declared with the wrong type, an analysis step returning empty without raising an error. When the model breaks, the first suspect is the model, not the athlete.

The signal worth watching in the next cycle is not on the results board. It sits in whether a file dares to refuse publication as a blank page — and in who audits the auditor. If a Vietnamese swimming dataset can return nine empty cells without anyone in the chain noticing, that says more about us than any index about a single lane.