When a Swimming Data Sheet Comes Back Blank: Verify First, Conclude Later

In January 2026 I sat at the back desk of a newsroom in Saigon, staring at a national championship swimming results sheet. Three columns of numbers, eleven blank cells. The data clerk shrugged: nothing yet, just write something. I did. The story ran the next morning with a line that still makes my face burn sixteen years later: a young swimmer had come close to a national record. No 50m splits, no back-half time, no sample size, no second source. One claim built on eleven empty cells, and nobody in the room that day objected. That same January, FINA, now World Aquatics, banned polyurethane suits from elite competition, effective January 1, 2026. The world record book was sliced into two eras. Ever since, every swimming figure that crosses my desk has to answer one question before I use it: which side of that cut does it sit on? FOUR PILLARS Strip away the decoration and swimming analysis stands on four pillars. The first is splits, the 50m-by-50m times, the only thing that reveals how an athlete distributes effort. The second is the pair of stroke rate and distance per stroke: going faster by turning the arms over more often, or by lengthening each pull. The third is micro-technical data: reaction time off the block, underwater distance, turn time on the wall, touch time at the finish. The fourth is meet context: a 50-metre long course or a 25-metre short course, heats or final, and the level of the meet that sets an A-cut or B-cut standard. None of those pillars means anything alone. A 25-metre pool produces more turns, so times are always faster than in a 50-metre pool; putting those two numbers side by side is putting two different events side by side. A 15-year-old and a 27-year-old can match on every split and still sit in completely different phases of a career curve. And a record set in 2026 and one set after 2026 both carry the label world record while belonging to two different equipment eras. Vietnam publishes very little of this in a structured, traceable form. Most domestic swimming data survives as PDF result sheets, photographs of electronic boards, or a coach handwritten log. Those formats are honest about how they were made. Nguyen Thi Anh Vien, Vietnam most decorated swimmer at SEA Games level, is a case in point: her journey is recorded mainly through news coverage and images, not through any open database that lets you query a single split. The problem comes at the next step: anyone who wants to turn those fragments into a judgment must rebuild the whole evidence chain by hand, and every rebuild is a chance to be wrong. WHEN THE SYSTEM RETURNS AN EMPTY FRAME Not long ago, an internal system handed me an analysis report. It had headings, a skeleton, all nine analytical sections laid out to template. But the one-sentence summary was blank. The list of information points came back empty. No athlete named, no event, no time, no date. Every field carried the line: insufficient information to assess. The easiest thing to do at that moment is to fill in the blanks. An experienced analyst staring at that frame can spend a few minutes and build a plausible story: a young swimmer breaking through, a record under threat, a race that fell apart in the last 50 metres. The story would read smoothly. And it would have no evidence behind it whatsoever. An empty cell is not a clean cell. That is the distinction most readers miss. In data quality work, no anomaly detected and never checked are two different states, and merging them is a serious error. In my tracking sheet, a row missing data is never allowed to turn green. It stays amber until a second piece of evidence arrives. Take the 200m freestyle. Two swimmers finish within the same span of time. The first blasts the opening 50 and fades. The second swims even and accelerates over the last 50, what analysts call a negative split. Identical results, opposite coaching meanings. The first has a pacing problem; the second has a better aerobic base and a higher pain tolerance. If the sheet only records the total time, the two are written down identically, and every conclusion drawn from that row is a guess dressed up in numbers. If I ever show a worked example here with specific figures, I must label it clearly as an illustration rather than a measurement. Readers have a right to know what they are reading. Katie Ledecky set the 800m freestyle world record in 8:04.79 at the Rio 2026 Olympics, according to official organiser data. That mark was born after the textile era, so it can be compared with other records from the same era. A record set in 2026 under the high-tech suit regime belongs in a different drawer. Sports writers routinely lump both into a single list of the greatest records of all time, and the merge quietly destroys comparability, because both numbers are correct. Competition structure produces data that results sheets never show. A major championship usually runs three rounds: morning heats, semi-finals, final. An athlete may swim three times in a little over twenty hours, with two swims separated by minutes in the semi. How they spend effort in the heats, just enough to advance or all-out, shapes the final more than people assume. An A-cut grants direct entry; a B-cut depends on quota allocation. A swimmer who hits B-cut can still compete, but the pathway there differs sharply in psychological terms and in scheduling. For teenage female swimmers, one variable never appears on a results sheet: the puberty barrier. This is the stage where physical change stalls or reverses performance for one group of athletes while another group surges. A 14-year-old breaking an age-group record offers no guarantee of holding that trajectory at 18. Calling her a phenomenon after one fast swim ignores the single most decisive variable in the event. Two more career risks sit alongside it: swimmer shoulder and breaststroker knee, the two most common injuries in high-volume athletes. And coaching stability, where a mid-cycle change is a far bigger event than a change in training set, and it never shows up on the scoreboard. The rulebook draws fairly clear lines: no more than 15 metres underwater after the start and after each turn, specific regulations on the breaststroke kick, rules on the backstroke start device. On anti-doping, the hierarchy running from World Aquatics through WADA to the Court of Arbitration for Sport demands one principle: a confirmed positive, a contamination dispute, a procedural violation, and a social media allegation are four entirely different tiers. Silence in the data is not evidence of a violation, and it is not a certificate of cleanliness either. It is just silence. The industry around swimming runs on events: the lights only come on for a major championship or a new face. Labels like the next Phelps or the next Ledecky appear on schedule, and their conversion rate into reality is low enough to deserve a proper statistic. The training market, equipment, broadcast rights all wait for a specific trigger. A report with no event in it generates no commercial signal at all. THE COUNTERINTUITIVE ANGLE Most readers treat a blank cell as harmless, because it says nothing wrong. Inside an analytical pipeline, however, the blank spreads: it passes format validation as a valid row, then surfaces in a summary as a verified row. A system that fails loudly gets stopped. A system that returns a beautiful frame with an empty payload goes straight to the decision-maker. I have stumbled on the inverse too. In 2026 my recovery index model predicted that the three heaviest pressing teams faced roughly a 23 percent rise in injury risk. Those three were also the teams training hardest during the shutdown. The model identified a relationship, and the relationship held. But the model assumed workload was the dominant variable; in a season split in half by a pandemic, everything else moved too, from the fixture list to training conditions. The estimate can be right while the mechanism is wrong. Since then I write the limitations section longer than the findings section. In 2026 I miscalculated a striker sprint distance: 1.2km instead of 0.8km. A specialist in the room said women do not understand tactics. I spent three months re-checking 14,000 GPS samples and found three further system errors in the synchronisation software. The cross-verification process I built afterwards became the club internal standard. My career began with a mistake, not with an achievement. SIGNALS FOR THE NEXT CYCLE The next cycle of swimming data should begin with three gates. Gate one: a report with zero information points must be flagged void and raise an error instead of being auto-filled. Gate two: every number must travel with a source, a retrieval timestamp and pool context. Gate three: every model must publish its own limitations before publishing its findings. I believe in the number, but only after it clears three rounds of verification. Data does not tell stories; it records everything so that I can tell them myself. A small GPS drift taught me enough: verification is everything. What I want to see next season is not another record, but a sheet thick enough that the next time a cell is blank, we know exactly what we are missing.