21:14. Lane four, women's 200-metre butterfly final. The scoreboard on the wall of the National Aquatic Centre clicked over, the stands rose, and the applause rolled down onto the water. On the screen in front of me, the data file that had just finished downloading ran to fourteen pages, every page dense with ruled tables and headings, and every content cell reading "insufficient information".

No 50-metre splits. No reaction time. No 150-metre turn time. No stroke rate, no distance per stroke. No pool length, 25 or 50 metres. No competition year.

The nine-dimension report I build for every race was still there: technique, performance, competition system, world landscape, rules and anti-doping, athlete pathway, risk profile, media narrative, industry ripple. Every section had a frame. Every frame was empty. The only populated field was the domain label: swimming.

At 22:40 I filed the empty report with a single line of note: no data, no speculation, update pending the raw file. It was the first post-race assessment of my career that I published without a single performance metric in it.

Readers of the morning bulletin saw only a very short piece. They did not see the other fourteen pages.

In swimming, a finishing time without a pool length and a competition year is analytically meaningless. A 25-metre pool doubles the number of turns relative to a 50-metre pool, and records are ratified under two separate systems that cannot be compared directly. The 2026-2026 polyurethane full-body suit era pushed a whole generation of marks to levels that, once the rules were tightened from 2026, forced analysts to screen results before any comparison. A 2:08 in a 25-metre pool and a 2:08 in a 50-metre pool are two different stories, separated by years of training.

The annual season applies a different kind of pressure. There is no Olympic medal to trade for, only heats and finals week after week, and audiences follow all forty-two events of the national championships. They expect the analysis that same night. The newsroom expects it that same night. The only thing that can be written that night, when the split file is missing, is an impression piece: who won, who cried, who embraced a coach at the end of the lane.

I do not write impression pieces, and the reason is not that I hold feeling cheap.

The data chain of a swimming final has four eyes: the touchpads at the wall, the electronic timing system, the cameras on either side of the lane, and the referee's cross-check sheet. Those four eyes produce reaction time, 50-metre split series, turn times, and, where force plates exist, stroke rate. Remove the second and third eyes and the remaining two cannot reconstruct a speed curve. And the speed curve is the only thing that lets anyone say "she won it in the last 50 or lost it in the first 50".

Based on my experience covering finals at the National Aquatic Centre over the past five seasons, I hold one rule: no split series, no technical judgement — description only.

Every analytical dimension has a minimum input threshold. The technical dimension needs an event name, an athlete name, and at least one of: split series, reaction time, stroke rate, underwater distance, turn time. The performance dimension needs a time, a date, a pool length, a meet tier. The competition-system dimension needs a meet name and a position in the four-year cycle. The world-landscape dimension needs a country. The rules and anti-doping dimension needs an incident or a testing record. The athlete-pathway dimension needs an age. The risk dimension needs a subject to attach risk to. The narrative dimension needs a headline. The industry-ripple dimension needs a brand or a market.

That night, not one dimension had its minimum inputs. And this is where my trade turns uncomfortable: a fourteen-page template already exists, and its existence creates pressure to fill it. That pressure does not come from the newsroom. It comes from the structure of the page itself.

The likeliest thing to happen in this situation is a silent substitution. The analyst reaches into memory, digs out previous seasons, and fills the empty cell with a figure that sounds entirely plausible. In 2026, a Vietnamese-language post about a sixteen-year-old female swimmer spread very fast, and it contained a second-50 split so beautiful that it was reshared thousands of times. That figure does not exist in the official record. It took a week for her coach to publicly deny it, and several more weeks for the original post to disappear from the groups. A sixteen-year-old girl learned her first lesson in media: people will assign you numbers you have never swum.

Women's swimming has a physiological breaking point that no results sheet ever records: the puberty barrier. Between fourteen and sixteen, the body changes, the wingspan changes, body composition changes, and a great many upward performance curves go flat and then descend. In Vietnam, a fifteen-year-old who breaks an age-group record is routinely written up as a "phenomenon", a "gem", "the future of the nation's swimming". Two seasons later her name is no longer on the training list, and nobody writes a single article about that disappearance.

The scouting network finds prodigies and manufactures lottery tickets at the same time. Families change schools, change homes, pour money into a centralised training slot, and nobody tells them how low the conversion rate is from age-group champion to continental-level athlete. I do not have that conversion figure for Vietnam. And I will not invent it just to make the article look fuller.

In 2026, at Suncorp Stadium in Brisbane, I published a prediction built on xG and running distance, and a male commentator smirked: "Sweetheart, football isn't mathematics." By full time the prediction had held. But what I remember is not that I was right; it is that I had the data to be right. The day Germany collapsed in Kazan taught me the other face: with data, you can still be wrong. Kazan is the day I learned that a 99 percent probability can still die on the betting table. Those two lessons combine into one principle — data does not guarantee me the truth, but without data I have no right to speak.

Close to midnight, the head coach of one delegation said to me on the phone: "If you don't have the numbers, say you don't have the numbers. Don't guess on my behalf. My athletes read everything you people write."

The empty report I filed still had content. It was a map of limits in three zones. The first zone is confirmed data: empty, and I stated plainly that it was empty. The second zone is ambiguous data: times shown on the electronic board but not yet cross-checked against the referee's sheet, therefore unusable for conclusions. The third zone is the one that must rest on judgement: the state of the water, the noise of the hall, the way an athlete walks to the podium with shoulders very slightly uneven. The winner's name was on the board. I did not put her name in the piece, because every technical figure of hers was missing, and a name without data only manufactures one more small, unverified legend.

Publishing an empty report looks like hiding behind the iron shield. I know that feeling, because I have used data as a shield often enough to recognise the difference between a careful analyst and a frightened one. But the greater risk sits on the opposite side: the swimming world is turning split charts into a new form of divination. A heat-mapped chart, colour-coded by 50-metre segment, looks extremely scientific, and it conceals what the athlete was actually doing inside the team's system — who set the pace, who was left as bait, who had to swim an outside lane alone from the start. A beautiful split chart may simply reflect that the rivals in the same lane were weak, not that the winner was superior.

This is where I have to argue against myself. "Insufficient data" is not allowed to become a permanent refuge. I set a seventy-two-hour limit. If the federation has not published the raw split file within seventy-two hours, the story changes in kind: from an analytical problem into a governance problem. At that point the thing worth writing is no longer how fast the girl swam, but how a sport stages a national championship and cannot release its own data.

I do not trust emotion. I trust a data series longer than your emotion. But when the series is empty, emotion is the only thing left, and it is also a form of data — we simply have no instrument for it. The winner of the women's 200-metre butterfly told me that over the final fifteen metres her ears filled and she could no longer hear the crowd. No device among the four eyes of the data chain records that detail. It cannot replace a split, but it explains why a split may not repeat.

Data has no gender, but the people who read it do. A split sheet belonging to a seventeen-year-old female swimmer is read by a coach hunting for a training slot, by a parent weighing whether to sell a motorbike, by a sponsor hunting a face for a campaign. One data series, four readings, four different consequences. That is why I do not rush to fill an empty cell.

Limits of the data: this report does not conclude who swam well. It concludes only that we do not know, and records precisely where we do not know. Three things lay beyond my measurement that night: the actual water temperature against the announced figure, whether lane four was disturbed by the hall's sound system, and whether the referee logged a missed touch on the return length. None of the three was confirmed. I left them as open questions rather than turning them into plausible-sounding hypotheses.

The signals to track in the next cycle are specific. First, whether the federation publishes the raw split file within seventy-two hours, and whether that file carries pool-length and competition-year tags — the cheapest test of whether a swimming nation is building a real data culture or only building an image. Second, this season's cohort of girls aged fifteen to seventeen: who holds an upward curve through the puberty barrier, and who is abandoned by the press after a single season. Third, the number of articles about the national championships containing at least one complete split series.

If a sport cannot publish its own splits, is it short of data, or short of the will to publish? I cannot answer that yet. But I know the first step toward a decent analytical culture is not buying more timing equipment. It is practising the hardest sentence to say: I do not have the numbers.