Pool Length, the Polyurethane Suit and an Empty Spreadsheet

At two in the morning in Saigon, I opened the analysis file that had just come back and found all nine sections returning the same line: insufficient information. No meet name. No athlete name. Not a single date. The only field still alive was the domain label — swimming. I sat looking at a template built to hold everything: technical analysis, performance data, competition systems, the world swimming map, rules and anti-doping, career trajectory, risk matrix, media narrative, downstream effects. Every cell was as flat as a pool at four in the morning. Sixteen years in this trade have accustomed me to missing numbers, wrong numbers, numbers rounded to look good. An empty spreadsheet is a different animal. It does not lie. It simply stays silent — and silence is the easiest thing to misread in this job. In 2026, aged eighteen, I carried a notebook and a stopwatch to a pool. I recorded a young swimmer's time and read it out with far too much enthusiasm. The coach asked one question: "How long is the pool?" I did not understand why that mattered so much. He explained: over the same distance, a 25-metre pool gives a swimmer twice as many wall push-offs as a 50-metre pool, and every push-off is speed added almost for free. A 1:48 in a short course pool and a 1:48 in a long course pool are two different stories, arguably two different athletes. That Saigon summer, I learned that data also needs watering. Since then, every swimming number I receive passes three filters before I allow myself to say anything about it. The first is pool length and competition format. The second is era — specifically 1 January 2026, when the world swimming federation banned full-body polyurethane suits. The third is event level: Olympics, long course world championships, short course world championships, World Cup, continental meets, or a domestic open. Skip any one of them and the number still looks good, still rounds nicely, still gets published — but it no longer means anything. The era filter is the cruellest one. At Rome 2026, the long course world championships, 43 world records fell in eight days. Forty-three. That was the window in which full-body polyurethane turned a mid-tier swimmer into a world-beater inside a single season. When the ban took effect at the start of 2026, an entire tier of records became the legacy of a technology cycle rather than of a generation of talent. So when someone places a 2026 mark beside a 2026 mark and calls the modern one "slower", I always ask again: slower in which pool, in which suit, in heats or in a final. The third filter decides how I discount a performance. A swimmer going a 200m breaststroke time at a domestic meet in March, mid base-training block, is not measured in the same unit as that same swimmer in a world championship final in July. In the United States, the Olympic Trials are a grinder: only the top two in each event go to the Games, and third place may be a national record holder. Elsewhere, quotas run on A-cuts and B-cuts, where an A-cut grants direct entry and a B-cut depends on allocation. Same time, two selection systems, two entirely different risk profiles. But there is one instrument I use more than all three filters combined: the improvement curve. A fifteen-year-old dropping three seconds in six months might be maturing, might be carrying a heavier training load, might have fixed a turn, or might be improving for reasons that should not exist. The improvement curve is the only thing that separates those four possibilities, and it only works when there is a continuous series of 50m splits across several meets. PPDA is not a number, it is a confession — and in swimming, the 50m split is the same. A first split 1.5 seconds faster than the last over 200m tells a story about pacing; a last split faster than the first tells a completely different story about how the athlete is being coached. In football I once wrote that every goal is a data point, but not every data point is a goal. In a pool, every touchpad finish is a data point, but not every finish is a performance worth analysing. Some records make me trust the process. On 24 July 2026 in Gwangju, Kristóf Milák swam 1:50.34 in the 200m butterfly, breaking Michael Phelps' 1:51.51 set at Rome 2026 — a record born in the polyurethane era, erased in a textile suit. On 23 July 2026 in Budapest, Léon Marchand swam 4:02.50 in the 400m individual medley, breaking Phelps' 4:03.84 from Beijing 2026, also a suit-era mark. Both cases share one structure: a young athlete, a smooth improvement curve, and an old ceiling removed by technique rather than by equipment. On 31 July 2026 in Paris, Pan Zhanle swam 46.40 in the 100m freestyle. Those three numbers hold up because they carry pool length, year, and enough splits to cross-check. Other numbers I have to park. A fourteen-year-old girl breaking an age-group record in a 25m pool at an open meet, with no published splits, no training-load information, and no long course comparison point — that is a news item, not a data point. The gap between those two things is exactly where this trade manufactures wrong stories. In Vietnam that gap is wider. Nguyễn Thị Ánh Viên competed at three Olympic Games and dominated the SEA Games for years; Nguyễn Huy Hoàng won bronze in the 1500m freestyle at the 2026 Asian Games and appeared at both Tokyo 2026 and Paris 2026. But high-level long course meets at home are rare, and most published times are short course or carry no pool length at all. When I compare two Vietnamese swimmers, I usually have to rebuild the context first — otherwise I am comparing two things that do not share a unit. That is why the empty file bothered me more than a wrong one. A wrong file I can fix. An empty file, read quickly, gets understood as "nothing to see here". A blank risk matrix does not mean the athlete carries no risk; it means nobody has been placed in front of the question yet. In this job, a false risk flag can damage a real person's reputation, and a missed one can damage a decision. Both are expensive. But the third error — reading emptiness as a clean bill of health — is the most expensive of all, because it leaves no trace. The strongest temptation at two in the morning is to fill the gaps with general knowledge. I know enough swimming to produce three pages that read perfectly plausibly about a real athlete, a real meet, a real injury — with not one line that can be verified. Sixteen years in the industry taught me that this is the most dangerous genre, because it is not wrong syntactically. It is only wrong factually, and the reader has no way to tell. The contrarian angle I want to keep: in swimming, the most suspicious thing is often an improvement curve that looks too clean at puberty. For female swimmers, the physical transition can stall or reverse results for a season or two before settling at a new baseline. Media love to label a fifteen-year-old "the next Phelps" or "the new Ledecky", but the redemption rate on those labels is very low — not because the athlete weakens, but because the label is applied before the body has finished speaking. When the curve breaks, the public calls it a fallen prodigy. Nothing failed; a forecast was simply issued too early, on too small a sample, in a phase where everything is still moving. Numbers do not lie, but they know how to hide something. Sometimes they hide by not showing up at all. If I have to draw one task for the next cycle, it will not be hunting more records. It will be hunting the missing rows: pool length attached to every time, year attached to every record, splits attached to every age-group breakthrough, and a dedicated empty column for everything that has not yet been verified. An honest spreadsheet is not one with no blanks. It is one that dares to leave blank what it does not know — and comes back to fill it in later. The question I leave for myself: across everything I read this week, how many numbers actually arrived with a pool length and a competition year? If the answer is very few, the problem is not in the pool.