When the Swimming Data Sheet Returns Zero: The Six Minimum Fields That Keep a Report Standing

The analysis file landed at 6:12 in the morning, nearly two days ahead of the deadline I set myself, and weighed exactly 41 kilobytes. I opened it with the confidence of someone who had just finished the hardest part of the day. Nine sections. Nine tables. More than forty data cells. Not one cell contained a swimming metric.

The "Athlete" column read N/A. The "Event and distance" column read N/A. The "Information points" field was entirely blank. The final two fields — "Entities involved" and "Source quality" — contained a verbatim instruction: identify from the information points above. There were no information points above. The instruction had been left in the output field, which means the system copied the assignment instead of answering it.

I sat still for four minutes. Then I did what I have done for nineteen years whenever a data sheet comes back empty: I closed the file and did not write a story from it.

An empty analysis file is not a thin article. It is proof that the extraction stage died before it ran — and any conclusion built on top of it is fabrication wearing the costume of intelligence.

My process has two layers. The first extracts from the source text: facts, entities, timestamps, source quality, time sensitivity. The second is where I ask professional questions — technique, performance, competition systems, international context, rules and anti-doping, athlete careers, risk, public narrative, industry ripple. The second layer cannot live if the first returns zero. A good analyst at layer two without data from layer one becomes a fiction writer, and that is the worst failure mode in my trade, because it is quiet.

I came to swimming before I came to football. In 2026 I was a swimming reporter for Thanh Nien newspaper, standing at the pool edge counting laps with a handheld stopwatch, writing splits into a paper notebook because automated tracking systems did not exist yet. That is where I learned the thing that later became the spine of everything I write: a swim only means something once you know how many metres long the pool was. The same athlete, the same distance, a 25-metre pool and a 50-metre pool produce two entirely different stories. Beginners compare those two results directly, and it is the single most common error in the whole sport.

In 2026 I wrote a blog post about Atlanta United using an xG model and an editor rejected it for fear readers would not understand the charts. When an editor says no, I learn to listen to the data — not to prove I was right, but to present data in a form other people can use. I published it on my own blog; a Belgian analyst shared it; it drew over two thousand reads in 48 hours. The lesson was not the two thousand. The lesson was that correct data nobody can read is worth zero, and wrong data beautifully presented is worth less than zero.

World Cup 2026 taught me the second lesson. The whole newsroom was watching Brazil and Germany. I analysed tracking data and saw Croatia holding an average PPDA of 8.2, with Modric covering 10.6 km per match at a pace that barely dropped in the second half. I wrote that Croatia would reach the final, and colleagues laughed. When Croatia beat England in the semi-final, the desk apologised and republished the piece. Being right too early is also a form of rejection — and I keep that feeling when I analyse swimming, because in swimming a national record broken at a domestic championship routinely runs six to eighteen months ahead of the international forecast.

In 2026, when the pandemic stopped football, the Bundesliga returned to empty stadiums. I recognised a perfect natural experiment and compared nine seasons of data against 93 matches without crowds: home win rate fell from 41.3% to 34.7%, average goals from 3.1 to 2.7. Empty stands, but the numbers still knew how to score. I wrote a twenty-page report but delayed it two months because I wanted to perfect the model — once again, perfectionism nearly killed a finding. Since then I set deadlines two days early and add a "data limitations" section to every piece.

The 2026 transfer window was a lesson in cross-verification. The data showed Kalvin Phillips, post-injury, dropping successful presses per 90 from 18.4 to 14.1, while Tyler Adams of RB Leipzig sat at 17.8. The big outlets hesitated; I worked with a European data broker to confirm, then was first to report that Leeds would buy Adams and sell Phillips to Man City for 45 million pounds. Every transfer is a problem waiting for a solution — but only when you have at least two independent sources pointing at the same number.

Back to swimming, where I now report for the US market. Swimming is the hardest sport I have covered to verify, for three reasons. First, a swim result is a raw absolute number: no xG, no advanced metrics, no transfer valuation to soften the judgement. Someone swims the 100 breaststroke in 59 seconds — that is it. There is no grey zone in which to say "the model indicates". Second, swimming runs two parallel frames of reference (long course, short course), which makes bad comparisons the default. Third, the four-year Olympic cycle imposes an enormous discount rate on every result, and most reporting ignores that rate entirely.

From there I built a list of six minimum fields. If a source does not supply all six, I do not analyse. I call it the six-field threshold.

Field one: athlete name. Obvious, apparently — but in swimming a name carries a nationality and a governing federation, and the federation determines the development system. An athlete raised in the US NCAA system has a different trajectory from one raised in a centralised national system. The NCAA gives swimmers year-round competition in an extremely dense internal field, but ties them to an academic calendar and caps training hours. A centralised system allows greater volume but less competitive friction. Without a name and a nationality, I do not know which story I am reading.

Field two: sex. In swimming this is the field with the greatest professional weight that outsiders treat as paperwork. The reason is the puberty threshold. For female athletes, puberty is the most frequent career cliff in the entire sport: structural body change stalls or reverses improvement for twelve to thirty months. A fifteen-year-old breaking an age-group record guarantees nothing about a national record at nineteen. Any report that applies a "prodigy" label to a pre-puberty female swimmer without an age and sex coordinate is a report that has not been checked.

Field three: age. Age gives me a position on the career curve: early fame, rising, peak, veteran, comeback. It also tells me which expectation is reasonable. A 23-year-old improving 1.2 seconds in a season is normal; a 28-year-old improving 1.2 seconds is a signal requiring investigation, because it usually comes from a technical change, a training-plan change, or something else. Without age I cannot separate those three.

Field four: event and distance. The 100 breaststroke and the 1,500 freestyle are two different sports running on the same surface. Distance determines the energy system: 50 and 100 metres lean on the phosphagen system and anaerobic glycolysis; 400 and up lean aerobic with complex pacing allocation. The same athlete can win a 200 and fail in a 400 with nothing abnormal happening. If the article does not state the event, every comparison after that point is meaningless.

Field five: the time, with pool length. This is the field I check first, always. A 50-metre pool is the Olympic standard and the standard on which World Aquatics ratifies world records; a 25-metre pool has twice as many turns over the same distance, and because a turn is a wall push that generates momentum, short course is almost always faster. World Aquatics — the global governing body, renamed from FINA in 2026 — publishes two separate record lists and does not permit conversion between them. A report that says "broke the record" without stating long or short course is a report that has not finished its sentence.

Field six: meet name and date. This field gives me two things at once. First, the meet's position on the pyramid: Olympics, World Championships, World Cup, continental, national. Second, the phase of the four-year cycle. A result in a post-Olympic year has lower reference value than the same result in a pre-Olympic year, because the athlete's training structure is entirely different: the post-Olympic year is for adjustment and experimentation, the pre-Olympic year for acceleration. Without meet and date, I cannot apply any discount rate to the result.

Those six fields are the floor. Above the floor, I start asking harder questions.

The first is split structure. In swimming, splits are the per-50-metre intermediate times within a race. They are the most important raw material for reading tactics and the most frequently ignored. Two swimmers finishing in the same time can have completely opposite split structures: one goes out fast and fades, one goes out slow and surges. Looking at the final time, they look identical. Looking at splits, they are two different stories, and each story forecasts a different future. A negative split over 200 metres signals a good aerobic base and disciplined pacing; a large positive split over the same distance signals a coming choice between speed and endurance.

The second is how the speed was produced. In swimming, speed is the product of stroke rate and distance per stroke. When an athlete improves, I always ask: did rate rise with distance held, or did distance collapse in exchange for rate? The first is durable improvement. The second is improvement with an expiry date. Many reports praise a breakout swim without knowing it was bought by losing efficiency per stroke — and it will disappear after one training cycle.

The third is the start and the underwater phase. In modern swimming, the first fifteen metres after the start and after each turn is the real contested zone, because the rules cap underwater distance. An athlete losing 0.3 seconds on the start can make it back with better stroke mechanics — or never make it back in a sprint. Without reaction-time and 15-metre split data, any technical judgement I offer is educated guesswork.

The fourth is qualifying context. In the US, an A-cut earns direct entry to the Olympics and World Championships, while a B-cut depends on quota allocation. But the standard matters less than the selection mechanism: the US trials take the top two on the day, regardless of world ranking. A swimmer ranked third in the world can stay home because two people were faster that morning. That mechanism is fundamentally different from federations that select by comprehensive evaluation. The same result, two mechanisms, two entirely different meanings — and without a named federation, I do not know which mechanism I am reading.

The fifth is the equipment-technology marker. From 1 January 2026, the world governing body then still called FINA banned polyurethane swimsuits, ending the high-tech suit era of 2026–2026. Any record comparison crossing that line must be flagged and era-adjusted. I have seen pieces compare a 2026 time directly with a 2026 time without a word about suits. That is not a small error. It is a comparison of two different sports.

The sixth is injury load. Swimming is a sport of repetitive, not impact, injury. Swimmer's shoulder is the signature injury of high freestyle and butterfly volume; breaststroker's knee comes from the torsional load of the whip kick. An athlete swimming five events at a major meet carries many times the load of a one-event swimmer, and the cumulative risk is non-linear. Without an event entry list, I cannot build a risk map.

The seventh is rules risk. Swimming has disqualification points spectators rarely see: false starts, non-simultaneous two-hand touches in breaststroke and butterfly, breaching the underwater limit after the start or after a turn, backstroke turn violations. A medal stripped over a hand position is real and happens regularly. A report that never mentions that branch is ignoring a risk path with non-trivial probability.

All of those questions require data. That is why I did not keep writing from an empty file.

My contrarian angle lies elsewhere, and it has nothing to do with swimming.

Sports news rewards the appearance of completeness. A nine-section report, each section with a table, a conclusion line, an evidence block, looks exactly like a real report. Readers cannot tell from the outside. Neither can editors, if they do not read the evidence section closely. And because complete form is rewarded, the pressure to fill the cells is real and very strong.

I fell into that trap in a milder form. In 2026 I had enough data but still delayed two months to perfect the model. That is the other face of the same coin: when you believe the product must look flawless, you either delay or you invent to fill. Both are failures. Time discipline — a hard deadline, publish when sufficient, state what is missing — is the only way out.

One more thing I learned from that empty file itself: emptiness is also data, but it is data about the process, not about the athlete. The instruction text left in the output field tells me the extraction stage stopped before inference, rather than encountering unreadable text. That is a diagnosis with real operational value: re-run on the same source and it will probably succeed. But it does not give me a single second of swim time to write about a swimmer. The two kinds of information must be kept strictly separate, because mixing them is the fastest way for a process diagnosis to be read as a professional judgement.

The same logic applies to anti-doping. Swimming carries a history of testing controversy long enough that the temptation is always present: one abnormal improvement and someone starts typing. I refuse. There are four distinct levels to separate — confirmed violation, contamination dispute, procedural issue, and pure public-opinion allegation. Without an athlete name and a described proceeding, I cannot classify into any of them. And in this sport, an anti-doping allegation attached to a name is almost never fully cleaned off, even after exoneration. Suspicion must never be promoted to fact merely because it feels plausible.

Another counter-intuitive point: more metrics does not mean more rigour. I have seen swimming analyses dense with numbers — stroke rate, distance per stroke, efficiency index, composite technique score — with not one line stating what device measured it, over how many swims, and whether there was a control. That is decorative data. It produces a feeling of rigour without producing predictive power. Three numbers with clear provenance are worth more than thirty with none, and I have to remind myself of that every time a piece starts to swell.

The last trap is reading correlation as causation. An athlete changes coach and improves — the press writes that the new coach is a genius. Without a control group I cannot separate the coaching effect from the age effect, the four-year cycle effect, the training-plan effect, the psychological effect after a disappointing season, or simply regression to the mean after a peak season. I do not argue with emotion, I present a chain of data — and the chain here is enough to say there is a correlation, not enough to say there is a cause. My model indicates a range of possibilities; it does not declare a conclusion.

So what does it take to move from a range of possibilities to a conclusion?

A set of six fields: athlete name, sex, age, event and distance, time with pool length, meet name with date. Those six unlock the technical, performance, competition-system, career and narrative dimensions simultaneously. It is the highest-leverage input expansion I have seen in this trade, because it is small enough that a poolside reporter can collect it in a thirty-second interview.

A second layer of three optional but valuable fields: per-50-metre splits, coach name, and the event entry list at the meet. These lift analysis quality by a step and a half. And a third layer of two fields reserved for deep analysis: reaction time with 15-metre splits, and a twelve-month injury history.

On the process side, I have added an output validation gate. Three conditions trigger automatic rejection: an empty information-points list; field values that match the prompt instruction verbatim; and any quantitative conclusion without an absolute timestamp attached. The gate does not make me write faster. It stops me having to retract, and over nineteen years the number of retractions caused by an unfounded conclusion has fallen from several a year to near zero.

On the editorial side, I propose one simple standard: every swimming report that publishes a time must state the pool length in the first sentence containing a number. Not in a footnote at the end. In that sentence. This change costs nothing and erases the sport's most common comparison error.

Based on my experience watching hundreds of swims from the pool deck and thousands of result lines, I would put the share of US swimming reports that state pool length in the first sentence at under one third. That figure is not an accusation. It is an indicator of where the process is being compressed to make deadline.

The match ends, but the data is still in stoppage time. For me, the stoppage time of a swim is not the last fifteen metres in lane four. It is the period after publication, when a reader asks me which pool that time was swum in. If I cannot answer in one sentence, the piece is not finished — however many words long it is.

Amid the noise of the stands, I choose to sit with the numbers. But numbers are only worth sitting with when they have at least six columns. An empty sheet I close, and that day I write nothing at all — which is the most correct editorial decision I could make before seven in the morning.