When the Splits Sheet Is Blank: Swimming's Data Gap Sits Exactly Where It Hurts Most

A nine-page technical assessment sat on the screen, and in every field that needed filling, the system returned the same line: insufficient information to assess. No athlete name. No distance. No stroke. No time, no date, no long course or short course designation. The engine ran through all nine dimensions — technique, performance, competition system, world landscape, rules and anti-doping, athlete career, risk, narrative, industry ripple — and stopped at the same sentence: no conclusion can be drawn.

What made me sit with that screen was not the emptiness. It was the honesty of it.

In more than a decade of covering swimming, I have read thousands of wrong analyses. The worst one was never the one that answered incorrectly. It was the one that answered everything, including questions for which no data existed.

Swimming has the strangest data infrastructure of any sport I have worked in. At the top tier, every lane is measured to a hundredth of a second, reaction time appears instantly on the electronic board, and official results reach the World Aquatics servers minutes after the touch. One tier down, the picture blurs faster than most people would believe.

A national championship in Southeast Asia usually publishes finals times and nothing else. The splits sheet — the 50-metre splits — vanishes. Reaction times vanish. Whether the pool was 50 metres or 25 metres sometimes survives only as a footnote nobody reads.

The 2026 annual season pushes that problem to the surface. The Asian Games in Nagoya in September and October are the biggest target of the year for both Vietnamese and Chinese swimming. Most of the meets before them are selection events, where organisers have no obligation to publish detailed data and no budget to do so.

Based on my experience watching races across pools from Beijing to Southeast Asia, roughly seven out of ten races I want to examine in depth end with me scrubbing video and measuring by eye. Eyeballing carries an error margin of about 0.1 seconds. In swimming, 0.1 seconds is the gap between a medal and a seat in the final.

The split void

Pan Zhanle's 46.40 seconds at Paris 2026 is public data and nobody disputes it. What matters sits in the part that was never published: what the first 15 metres underwater looked like, how many dolphin kicks, where the breakout happened, what the stroke rate was through the first 25 metres and how much it decayed through the last 25.

A 100-metre freestyle race breaks into fragments: reaction off the blocks, the underwater segment, the acceleration after surfacing, the main swimming segment defined by stroke rate and distance per stroke, two turns, and the finish. Coaches routinely say the underwater segment sets the momentum for the entire first 50 metres, yet that same segment is the least recorded part of the race.

Public splits at major meets reveal two or four of those fragments. The rest sits on a coach's laptop, or does not exist.

The consequence is not small. With only a total time, an analyst cannot separate an athlete who is fast because of the start and the underwater work from an athlete who is fast because of high stroke rate and aerobic durability. Those are two different athletes who need two different training programmes, two different racing strategies, and who will age along two different career curves. Folding them into a single total time is the fastest route to a wrong forecast.

The conditions void

A 50-metre pool and a 25-metre pool produce two different time systems. A turn in a short course pool saves considerably more energy, so the same swimmer can sit seconds apart between the two. Conversion coefficients exist, but they are statistical averages, not physics. A swimmer with excellent turns gains more from short course than the average; a swimmer with poor turns loses more. Converting without knowing that is manufacturing a systematic error.

Then there is era. The men's 200-metre freestyle world record is still 1:42.00, set by Paul Biedermann in Rome in 2026, during the polyurethane suit period before the ban. That timeline is accurate, fully documented, and has stood for more than fifteen years while training methodology has changed completely. Comparing that figure directly with a 2026 swim is comparing two different objects.

The third factor is physical state on the day. A swimmer who goes 1:58 in the 200-metre individual medley in March and 1:56 in July may have improved very little. He simply tapered. Conversely, someone who holds the same time across both dates while doubling training volume is showing a strong progress signal. Without training-cycle data, two swims that look identical on the clock tell opposite stories.

The sample void

Swimming has the poorest sample size of any timed sport. A world-class swimmer races their main event at most five or six times a year, and only one or two of those come at full taper. Every model built on that foundation is working with what statisticians call too small a sample to conclude anything.

At Paris 2026, Leon Marchand won four gold medals, including an evening when he swam two finals back to back in two different strokes. The most analysable part of that night was not the two winning times but the recovery window between the two starts — heart rate, blood lactate, muscle temperature, actual rest duration. Almost none of it was published.

An injury is where every analytical model has to bow its head — and it is also where I have learned the most. A butterfly swimmer's shoulder can reshape an entire career curve in six weeks, and no performance metric warns you in advance. When medical data is withheld, the analyst is forced to infer from the outside, and outside inference about injury carries the highest error rate in this profession.

Some discoveries do not come from luck, but from the willingness to read the movements the crowd skips past. When splits are missing, I read other things: how often an athlete posts a personal best in the final rather than the heats, the gap between heats and final, the stability of reaction time across meets, and the tendency to swim the back half faster than the front. All four are computable from public data, and they say more about physical capacity than a single total time ever will.

The void is itself data

The industry's reflex is to fill the gap with story. No splits, so we talk about character. No recovery data, so we talk about spirit. No injury data, so we talk about hunger. Those words carry weight, but they cannot be verified, cannot be refuted, and therefore cannot be wrong. A statement that cannot be wrong is a useless statement in analysis.

The contrarian position, and the one I believe: an honestly marked empty field is worth more than a plausible number filled in to look complete. The absence of data is itself information. When a meet publishes no splits, that tells you what tier the meet occupies, how much the federation invests in measurement infrastructure, and how broadcast rights are valued. When a federation publishes everything, that tells you it is building a swimming culture that wants to be analysed seriously.

A doping case disclosed three years late is the clearest illustration of what a void costs. During those three years, what filled the space was not fact but imagination — and public imagination is always cheaper than an investigative file. By the time the file opened, most parties had already built their positions and there was no room left for detail. That is a double loss: one side loses the right to be judged by data, the other loses the right to be trusted by data.

Data does not judge, but it points me toward the questions everyone else forgot.

The most dangerous analyst in this sport right now is not the cautious one. It is the one with a confident number for every lane, every distance, every meet — because to do that, they must fabricate exactly where data does not exist, and fabricate in the places that are detected last.

I once mispronounced a footballer's name at a World Cup, and from that I rebuilt the entire way I watch a match. The lesson was not about memorising names. It was that I had trusted memory instead of trusting a system I could check. Swimming lacks exactly that system at the base layer: a mandatory, public, traceable archive for every race counted as an official result.

What to watch before Nagoya

When the Asian Games open in Nagoya in September, the question I care about is not who swims fastest. It is which federation publishes full splits for every round, and which one returns a sheet of finals times and leaves the rest to commentary with wings.

A swimming nation that wants to be taken seriously has to start by giving others the right to judge it. For Vietnam and China, the distance between those two choices is not large in money, but it is very large in time — because data not published today cannot be reconstructed tomorrow.

If next year's national championship published full splits across every round, how many of the conclusions we currently hold as true would have to be rewritten from scratch?