Swimming and the Measurement Problem: What a Lane Still Isn't Telling Us Through Data

In 2026, at the My Dinh Water Sports Palace, I sat in the seventh row with two stopwatches in my hands. A women's 200m individual medley final had just ended. The electronic board showed the finishing time but left the split section blank. The organisers published splits only for finals, not for heats. All the data needed to answer one seemingly simple question — where a swimmer accelerated, where she let it slip — vanished before I could open my laptop. I turned to the colleague beside me, a reporter who had covered swimming for over a decade: "Did anyone capture the heat splits this year?" He shook his head. A tiny GPS drift taught me this much: cross-verification is everything.

That afternoon was not a big event. It is the common denominator of almost every swimming front I have followed across eighteen years, since my early days writing for Thanh Nien newspaper, when I stood by the poolside recording every stroke of young swimmers. Back then I thought my job was to retell the race. Now I know my job is to find where the data goes silent.

Swimming is one of the most heavily measured sports. Every lane is 50 metres long, every wall touch is a signal, every hundredth of a second can decide a placing. World Aquatics (formerly FINA) maintains world and Olympic records for every distance and stroke, in both 50-metre and 25-metre pools. In theory, this is a sport of numbers. In practice, most numbers exist only at the top of the pyramid. The further down you go, the thinner the data becomes, until it disappears entirely at the grassroots level — precisely the level that developing nations like Vietnam need most to know where they stand.

I used to think this was only a logistics problem. Later I understood it was a methodological one.

Splits are what tell us how a race actually unfolded. If you only have the finishing time, you know who won but not why. With 50-metre splits, you see the pacing structure: who went out fast and faded at 150 metres, who swam a negative split by taking the last 50 faster than the first, who held an even rhythm throughout. In my line of work, that is the difference between a line of results and a verifiable story.

The problem sits exactly there: split data is abandoned in rounds that are not televised. I have cross-checked this many times. At major international meets, every swim has splits; at many domestic and regional meets, only the finals do. That means the long-term database used to compare a swimmer across years — the most basic tool anyone in analysis has — is built on a broken foundation.

I learned to read those gaps from a mistake. In 2026, while serving as the only data analyst in the technical department of a football club in Nha Trang, I miscalculated a striker's sprint distance: recording 1.2 km instead of 0.8 km. A specialist in the room sneered that a woman at a desk would never understand tactics. After the match, I re-checked all 14,000 GPS samples the team had produced over three months and found three more systemic errors from the synchronisation software. From then on, every table I built carried an extra column: confidence level. I trust numbers, but only after they pass three rounds of checking.

That principle applies to swimming intact. A split without a source is not data; it is a rumour with a unit of measurement.

In swimming, most of the race is decided before the swimmer's head breaks the surface. The rules allow freestyle, backstroke and butterfly swimmers to stay underwater for up to 15 metres after every start and every turn. In breaststroke, exactly one dolphin kick is permitted during the underwater phase after the start and after each turn. These technical rules turn the underwater phase into a theatre of its own, where skill, hip power and oxygen economy decide the time.

But if you only have the finishing result, you cannot see that theatre. You do not know which of two swimmers sharing a finish time won the opening 15 metres and which made it back over the final 350. To see it, you need speed and stroke-rate data by segment, and in many places that data simply does not exist.

Alongside this is the story of swim efficiency, usually measured by stroke rate and distance per stroke. Two swimmers can post the same 100m freestyle time with opposite strategies: one raises stroke rate to compensate for force, the other lowers stroke rate and pulls a longer distance per stroke. Without a dataset of these two metrics, every debate about "whose arms are stronger" is meaningless. It is the loudest argument in the stands and the emptiest one in the data room.

At the performance level, analysts use a coordinate system of world records, all-time lists and in-season rankings. These are the three reference points for locating a result. In swimming, the era factor is far more complicated than in most sports, because the high-tech suit era around 2026-2026 created a tier of records in polyurethane material before the rules returned to textile. When comparing records, the analyst must screen by equipment era, or they will mix two biologically different competitions into one chart. It is a screening I have learned to question: what suit was this record set in, under which start rule, and in a long or short course?

Because 50-metre and 25-metre pools generate two different physical constants: the short course allows more turns and underwater starts, so times are usually faster. Mixing short-course results into a long-course comparison without a note is a basic error. A lane does not announce its own length; the data reader must say it for them.

Swimming's competition system is clearly tiered, and a result's position depends on which tier it sits in. At the top are the Olympic Games and the long-course world championships; below that, the short-course world championships and the World Cup; regionally, the SEA Games and the Asian Games serve as both battleground and springboard. For Vietnamese swimming, these two fronts define almost the entire four-year cycle.

I remember the career of Nguyen Thi Anh Vien, who was a pillar of Vietnamese swimming at the SEA Games with a substantial medal collection, and Nguyen Huy Hoang, who made his mark in distance events at the Olympic Games. Those names are proof that Vietnam's swimming potential is not lacking. But ask a different question — how that potential is measured and tracked year by year — and the answer is much thinner.

That is when the issue stops being individual technique and becomes a system. An Olympic cycle is divided into Games years, adjustment years, accumulation years and sprint years. A result in each type of year carries different meaning: performances in sprint years are weighted heavily, while those in adjustment years after a major meet are usually discounted. Without locating a result within the cycle, people will inflate a heat qualification in an adjustment year into a breakthrough, then be disappointed when it is not repeated the following year.

Selection mechanisms and entry quotas are also part of the picture. The A and B standards set by the international federation turn every domestic meet into a race for places, and the quality of that race depends on the density of rivals in the same event. A swimmer who is fast domestically but has no equal rival will struggle to know where they stand against the world. This is why teams send swimmers abroad for training camps — not only to train, but to have someone swimming beside them every morning.

Meet density and scheduling leave traces too. A meet lasting many days with two swims per day across heats and finals creates an accumulated load. When I built a recovery-index model for footballers during the seven months the 2026 season was suspended, I relied on high-intensity running distance, acceleration counts and injury history. The pandemic season taught me to measure a competition by its recovery index, not by its points tally. The principle transfers to swimming easily: accumulated high-intensity metres across rounds, plus the number of short recoveries between swims, is a far more useful indicator than the final placing alone.

The world swimming map has long been drawn around a leading group, and the striking thing is not who is winning but the development system behind them. The United States sustains its edge through the college system, where thousands of swimmers train beside each other daily and internal selection is as fierce as a world final. Australia stands out in freestyle events and a coaching tradition tied to the sea. China and Britain have built centralised training centres. What these systems share is depth: they do not have one star, they have a whole tier of swimmers good enough to push each other up.

Against that picture, Vietnamese swimming sits among nations with potential but a thin talent supply chain. Limited squad depth in each event means a talented swimmer can easily end up alone in an event, with no internal rival to drive progress. This is a structural problem that cannot be solved by a single individual training camp slot.

Then there are personnel-movement signals — sporting nationality switches, training-base changes, coaching changes. Each time, a swimmer's data is interrupted. Someone relearning technique at a new centre may lose a season finding their rhythm again. Without tracking that rupture, a dip in performance is read as a decline in form, when in fact it is a transition cost.

At the level of rules and governance, swimming has one of the strictest monitoring systems in sport. World Aquatics works with the World Anti-Doping Agency (WADA) and national organisations, and every eligibility decision passes through a clear procedural process. This places a special demand on the writer: distinguishing between a media allegation and a confirmed violation.

I have one rule: until there is an official conclusion, the only word I use is "ongoing", never "confirmed". A case can be a confirmed violation, a contamination dispute, a procedural error, or merely an online allegation. These four situations carry very different severity and legal consequences, yet in headlines they are usually merged into one. The data analyst has a duty to separate them, even when that makes the article less sensational.

On competition rules, the flashpoints usually lie in technical limits — the number of dolphin kicks allowed in breaststroke, the 15-metre limit in freestyle, backstroke and butterfly, the use of start equipment. A turn judged illegal can erase an entire medal. So whenever a result is disputed, the first thing I do is re-check the rule applicable to the exact stroke and phase — not rely on memory.

On career trajectories, swimming has a fairly distinctive age curve. Most athletes peak in their twenties, with sprint events often earlier and distance events potentially later. For female swimmers, puberty is a significant physiological barrier: changes in body structure affect propulsion and buoyancy, so some junior results do not translate to senior level.

This is what I always place alongside eye-catching junior performances. In my former transfer-analysis work, I once saw a case: a player whose conversion rate was nearly double the league average, mostly from set-piece situations, and I recommended against signing him. They ignored it. His actual output was never repeated. That story is not meant to say science is always right, but that a beautiful number must be placed beside the context that produced it. In swimming this matters even more: a junior result dependent on a developmental phase may not be a forecast, but merely a measurement taken before maturity.

On physical risk, the two classic injuries in this sport are the shoulder (common in strokes with continuous arm pulling) and the knee (characteristic of breaststroke). Beyond that are cumulative training load and big-meet psychological issues. These factors are hard to measure directly but leave traces in the data: truncated training sessions, shifts in stroke rate, sudden drops in performance. The data analyst can see the fractures before the results board announces them.

At the level of public narrative, swimming easily generates durable and misleading labels. "Record night", "prodigy", "the comeback", "medal storm" — these phrases exist to sell tickets, not to explain. They attach to one result and stay for a long time, even when later results do not reinforce them.

The trouble with a label is that it replaces verification. When a swimmer is called a prodigy, people stop asking about the sample behind that performance — sample size, opponents, pool conditions, point in the cycle. The stronger the label, the larger the verification gap. Conversely, a modest result at a low-tier meet can signal a rising development curve that goes unrecorded, simply because it lacks a catchy headline.

The consequence of measuring by labels rather than data is that we do not know where we are in the cycle. We read a regional meet with the mindset of a world meet, or dismiss an international result because it does not fit the story being told.

At the industry level, a swimming front drags an entire value chain behind it. Upstream is the youth training market, the pool system and the talent supply; midstream are the athletes and events; downstream are broadcasting, sponsorship, equipment and derivative markets. Even equipment details tie to performance history: every time the swimsuit rule changes, a whole tier of old records has its comparability re-questioned.

Investment in pools is a quiet, long-term bet. A competition-standard pool does not serve one meet; it creates a training hub, a competitive community, and a database of local swimmers. But those benefits only appear after many seasons, so they rarely enter the balance sheet of a single term.

For Vietnam, I believe the biggest bottleneck in swimming is not at the talent peak but in the data infrastructure — something far cheaper than a pool and yet the easiest to overlook. A system that records splits from the heats, archives by year, and publishes under control would change how the entire swimming community assesses itself, from district level up to the national team.

The counterintuitive point I want to state plainly is this: more data does not automatically make anyone swim faster. This is where professionals like me are most prone to self-deception — believing that with enough numbers, answers will emerge on their own. In many cases the opposite is true: a young swimmer needs protection from too much data, from the pressure to prove themselves every single morning.

I also have to warn myself about old habits. After years of building models, I have seen myself become rigid about my own model — defending it against new data that refutes it. Cross-verification sounds like humility, but it can also become a way of delaying. Sometimes the counter-evidence is already sitting inside my article, and the most honest thing is to display it rather than fold it into a footnote.

There is a fact easily overlooked in the splits story: many of the most important technical decisions are not made in the strongest heat, but in a "just enough" heat by a swimmer conserving energy. If we record only what stands out, we lose exactly the type of data needed to understand strategy. Recording everything is more useful than recording only when it looks good.

In eighteen years in this profession, half my time has been at a desk with data, the other half standing by the pool. For the past three years, most of my time has gone into answering one seemingly simple question: where in our swimming data stream is it going silent, and why. I do not have the whole answer. But I believe in one principle. I trust numbers, but only after they pass three rounds of checking.

Data does not tell stories; it records everything so that I can tell them myself. And if Vietnamese swimming wants to read its own story correctly, the first step is to keep enough traces to read later. Every split discarded today is a question that will never have an answer on some afternoon ten years from now.