A Blank Page in the Lane: The Craft of Reading Numbers and the Honesty of Data

The data table returned a blank line.

That is the sentence I wrote in my professional notebook at 3:17 a.m., after running a query on my swimming analytics system. I was looking for the 50m split of a swimmer in the heats of a domestic meet. The result: nothing. No finish time, no stroke rate, no kick count. Just a grey line of text, exactly what data engineers call a null value.

In more than twenty years of following swimming, I had grown used to a data table always having something to read. Even the smallest meet leaves traces: a touch time, a lane number, a heat ranking. Yet that night, an entire lane became empty space. Instead of panicking, I sat still. Because I know something that newcomers to the trade often forget: a blank line is also data. It tells you that something has broken, somewhere, in the chain of information collection. And if you rush to fill it with a fabricated number, you have not fixed the fault — you have merely hidden it.

Every shock has its own probability. We call it a shock only when we have not yet checked the table. But there is another kind of "shock," a quieter one: the moment you discover that your system holds no data at all, while the entire sports world is arguing over a result that nobody can actually verify.

That is the subject of this article. Not a medal, not a record. But the empty space behind the medal — and how Vietnamese sport is handling it.

Context: an ecosystem of data as thin as a pool wall

To understand why a blank line matters, you need to understand how swimming data is born.

Unlike football, where every match has dozens of cameras and hundreds of automatically recorded metrics, swimming is a sport of single numbers. One athlete, one lane, one duration measured in hundredths of a second. The entire value of the sport lies in the precision of measurement. You cannot analyse a swim if you do not have times split by each 50m. You cannot compare two swimmers if one raced in a 50m pool and the other in a 25m pool. And you absolutely cannot conclude who has improved without the operating conditions: water temperature, altitude above sea level, lane assignment, the time of day of the race.

In Vietnam, the swimming data ecosystem has three tiers. The first tier is international meets — SEA Games, Asian Games, the Olympics, the Asian Championships. There, data is supplied by the international organisers, relatively complete and traceable. The second tier is national and junior meets. This is where data starts to thin out. Electronic touchpads may exist, but exporting the data into an analysable format is not always done. The third tier is grassroots meets, school meets, provincial meets. At this tier, data exists almost only on paper, in photographs of scoreboards, or in the memory of coaches.

Because of that three-tier structure, a blank line is not a rare event. It is an everyday event. The problem is that not everyone dares to say so.

I once sat in a meeting where I was asked to "estimate" an athlete's time at a meet with no electronic data. I asked: estimate based on what? The answer was: based on the coach's feeling. I stayed silent. Because I know that a good coach's feeling is a valuable source — but it is not data. It is a hypothesis. And a hypothesis, if treated as a certain number, will poison every conclusion downstream.

Body: anatomy of a blank line

Let us begin with the physical mechanism. When an athlete swims, their body generates data at three levels.

The first level is the level of result: the final touch time. This number is measured by electronic touchpads, or by hand timing if the meet does not have the equipment. The error between a touchpad and a hand timer can reach several tenths of a second — in a sport where medals are decided by a hundredth of a second, that is an enormous gap.

The second level is the level of splits. Each time an athlete turns at the wall, the touchpad records a split. In a 50m pool, a 400m freestyle swim has seven intermediate splits. In a 25m pool, that number doubles. This split structure is what allows an analyst to see the rhythm of a swim: where the athlete accelerates, where they fade, how they finish.

The third level is the level of technique: stroke rate, the number of arm cycles per length, turn time at the wall, the underwater glide distance after the start and after each turn. This is the richest tier of data, but also the hardest to collect, because it requires dedicated cameras or wearable sensors.

When a blank line appears, it can come from any of those three tiers. And the analyst's job is to determine where it came from — before concluding anything about the athlete.

This is where many people in the trade go wrong. They see an athlete with a good result at the SEA Games and immediately infer that the athlete is improving. But "improving" is a comparative word. It requires two data points at two different times, under the same operating conditions. If one point is in a 50m outdoor pool and the other in a 25m indoor pool, you are comparing two things that cannot be compared. Short-course pools produce faster times because there are more push-offs from the wall. An athlete can "break a record" simply because today they swam short course, while the old record was set long course. That is not improvement. That is a comparison error.

I still remember the feeling when I first managed to draw a complete split chart for an athlete. That was the period after 2026, when I began introducing xG — the measure of chance quality — into football analysis, and realised that the same principle could be applied to swimming: do not look at the final result, look at the process that produced it. In swimming, the final result is the touch time. The process is the seven intermediate splits. And the real story lies in the splits.

A 400m freestyle swimmer can finish with an unchanged time across three consecutive meets. On the surface, they are standing still. But if you look at the splits, you may see two entirely opposite scenarios. Scenario one: they start slower but finish faster — meaning their base fitness and ability to distribute effort are improving. Scenario two: they start faster but fade in the final 100m — meaning they are burning all their energy too early, and that unchanged time is in fact a disguised step backwards. The same number, two stories. Without splits, you cannot tell them apart.

That is why I treat a blank line as a more serious signal than a defeat. A defeat gives you data to learn from. A blank line takes your data away and leaves a blind spot.

The SWOLF index and the trap of pretty numbers

In swimming, there is a fairly common composite metric: SWOLF. It is the swimming version of a golf score — you add the time to swim one length of the pool to the number of arm cycles in that length. The lower the score, the better, because it means you swim fast with fewer movements.

SWOLF is a good metric. It forces the analyst to look at speed and efficiency at the same time. But it is also a trap, and I have seen many people fall into it.

The trap is this: SWOLF is only meaningful when you compare two swims in the same pool length, the same distance, and the same water conditions. If you compare one athlete's SWOLF in a 50m pool with another's in a 25m pool, you are comparing two different units of measurement. In a short-course pool, the number of arm cycles is lower because there are more turns, so SWOLF is naturally lower. A pretty number may simply be a consequence of pool architecture, not of talent.

This connects directly to our subject. When the underlying data is missing, people tend to replace it with composite metrics — because composite metrics look more complete. But a composite metric calculated from missing data is not a metric. It is a decorative number. And in a sport where investment, selection, and remuneration decisions all rest on metrics, a decorative number can cause real harm.

Based on my experience of following matches and swimming meets, I have learned one principle: better to lack a metric than to hold a false one. Because when you lack it, you know you are blind. But when you hold a false one, you believe you can see — and that is when you lose your way.

Border conditions: the variables nobody records

Newcomers often think swimming is a simple sport: water, lane, time. But swimming has more border conditions than any other sport I have ever analysed.

Try counting them. Water temperature directly affects muscle performance and buoyancy. Colder water contracts the muscles, warmer water slackens them. Altitude above sea level affects the oxygen in the blood — a meet held at altitude will produce slower times, and that does not reflect ability. Lane assignment matters too: the middle lanes carry wave reflection from both sides, while outer lanes are cleaner but farther from the view of the crowd and the officials. The time of day of the race affects circadian rhythm — morning heats and evening finals produce two different bodily states. And of course, the stands.

I always tell my students: when the stands fall silent, the home-pool advantage melts into a number close to zero. Swimming has no notion of "home ground" in the football sense of a crowd screaming at the opponent. But it has another kind of advantage: familiarity with the pool surface, the depth, the lighting, and the feel of the water. When a Vietnamese athlete competes in the very pool where they train daily, that advantage is real, however small. Conversely, when they must compete in an unfamiliar pool, the entire advantage vanishes — and if you do not record the variable of "pool familiarity" in your data, you will unwittingly attribute the drop in performance to form, when the real culprit is the environment.

This is why I never analyse a swim without asking three questions about border conditions: long course or short course, indoor or outdoor, and at what point in the season. Those three questions are cheap but they protect you from wrong conclusions. And if I have no answer, I state clearly in the report: condition data insufficient, conclusion carries a wide confidence interval.

I would rather record a wide confidence interval than present a narrow conclusion built on a foundation that does not exist.

The case of Nguyen Thi Anh Vien and the price of collective memory

One cannot speak of Vietnamese swimming without mentioning Nguyen Thi Anh Vien, born in 2026 in Can Tho. She is the most successful swimmer in the history of Vietnamese sport, with eight gold medals at the 2026 SEA Games in Singapore — a feat no one has repeated to this day. She competed at three consecutive Olympic Games, and for years her name was synonymous with Vietnam's elite swimming.

I mention her not to praise her. I mention her to point out a data problem.

Throughout Anh Vien's career, a large volume of data was recorded: times, splits, rankings. But most of that data sits in the first tier — international meets. When she retired and Vietnam entered a phase of searching for a successor, a question arose at once: is our data foundation sufficient to find a successor?

The honest answer is: not entirely.

Because to find a successor, you need data at the second and third tiers — junior meets, national meets, school meets. You need to know how a fifteen-year-old swims a 200m individual medley, what their splits are, what their stroke rate is, how good their turns are. But that data tier is the thinnest of all. Many of our young talents grow up inside a data blind spot.

This is what I want to call "the price of collective memory." We remember a decorated athlete very clearly, because that memory is nourished by the media. But we have almost no systematic data on those who are on the road to becoming decorated. Collective memory is a form of data, but it is data that cannot be traced, cannot be verified, and cannot be compared. It is like a blank line coloured in with emotion.

The case of Nguyen Huy Hoang and the power of real data

On the other side, there are cases where the data has spoken clearly enough.

Nguyen Huy Hoang, born in 2026, is one example. He drew attention in distance freestyle, and at the 2026 Asian Games in Jakarta he won bronze in the 1500m freestyle — one of the rare achievements of Vietnamese swimming on the continental stage. That was the result of a process, not of a moment.

I remember sitting down to analyse Huy Hoang's splits in the 1500m — the longest event in the pool, where the tactics of distributing effort matter more than raw speed. In this event, the fastest finisher is rarely the leader at the first 200m. The winner is the one who maintains consistency across the splits — with no segment fading too deep. If you only look at the final time, you see one number. If you look at the splits, you see a strategy.

That is the difference between a scoreboard and an analysis.

And this is why I stress the collection of data at the second and third tiers. Huy Hoang did not appear out of nowhere. He rose through junior meets, through selection trials. If those meets leave no data, then our talent-discovery process depends entirely on the human eye — and the human eye, however good, has limits.

The contrarian angle: the temptation to fill the gap

This is the part I want to devote to the hardest thing in my trade.

Not hard technically. Hard ethically.

When you have a blank line, there are four ways to handle it. The first: leave it blank and note it. The second: find a credible alternative data source. The third: estimate and label it clearly as an estimate. The fourth: invent a plausible-looking number.

The fourth way is the most common. And it is common because it is rewarded.

Think about incentives. An article with complete data is more attractive than one with gaps. A selection report with all the metrics is easier to approve than one that says "we lack data." A press conference with concrete numbers inspires more confidence than one that says "we cannot yet conclude." The reward system encourages subtle fabrication.

And the most subtle trap is the trap of correlation. When you see two data series moving together, you very easily conclude that one causes the other. For example: you see an athlete increase their training volume, and at the same time their performance improves. You conclude: more training means faster swimming. But both might be consequences of a third variable — say, the athlete has just moved to a new coach, who both increased the volume and improved their turn technique. If you cannot point to the physical mechanism linking the two variables, you are not allowed to use the word "causes." You may only use the word "correlates."

I once saw a technical report asserting that a certain type of leg exercise was the cause of an athlete's improvement. The report was based on a comparison of two seasons. But it did not check one simple thing: whether the athlete had changed anything else in that same period. It did not check. And so that conclusion is not a finding. It is a belief presented in the form of data.

That is the most dangerous thing a data analyst can do: use the authority of numbers to legitimise a belief.

There is one more layer of complexity I want to confess. In matches and swimming meets with great social meaning — a derby, a SEA Games held at home, an athlete carrying the expectations of an entire region — emotion creates a portion of variance that my models cannot explain. The stands can create a pressure that no metric captures. I have been wrong when predicting that an athlete would shine at a major meet, simply because I treated the crowd as a variable that can be switched on and off, when in reality emotion is a non-linear variable.

Since then, I have learned to write one line into every report of mine: the confidence interval widens when crowd conditions exceed historical thresholds. It is a way of admitting that there are things I do not know — and admitting ignorance is the first step of an honest analysis.

From blank line to signal

Back to the story at the start. After the system returned a blank line, I did three things.

The first: I checked which tier the blank line came from. I traced the data path and discovered that the electronic touchpads were still recording, but the meet's data-export software had not been running the correct version, so the results were not being pushed into an analysable format. The data was not lost at all. It was simply blocked at one stage. It was a pipeline fault, not an athlete's fault.

The second: I recorded that fault as an operational signal. Because if this error happened to one meet, it could happen to a whole series of meets run under the same procedure. A blank line is never an isolated event. It is a sign of a system.

The third: I did not infer anything about the athlete while waiting for the data. I did not guess their times. I did not assess their form. I left that gap empty.

A tactical era dies when its data table is no longer read by anyone. And a sport weakens not when it loses, but when it stops verifying itself.

Takeaway: the signal of the next cycle

So what will shape the next cycle of Vietnamese swimming data?

First, I believe the pressure will come from the lowest tier. Junior and school meets will gradually be digitised, not because organisers want it, but because parents and training centres need data to demonstrate their students' progress. When that demand appears, data will flow on its own.

Second, I think we will witness a shift from result metrics to process metrics. Counting medals still matters, but it is no longer enough to answer the central question of any sport: are we rising or falling? To answer that question, you need splits, stroke rate, turn times, border-condition data. You need the things of which we currently have only a portion.

Third, and perhaps most important, I believe the greatest value of data lies not in predicting who will win. It lies in forcing us to be honest. A good data system does not allow you to fabricate. It does not allow you to say "this athlete is improving" when you do not have two data points to compare. It does not allow you to call a result a shock when you have not checked the base probability table.

Ordinary people look at the medal to understand the lane. I look at the lane to understand the years.

And when I sit far from the pool wall, looking at a blank line on my screen, I do not see failure. I see a reminder: that in a sport decided by a hundredth of a second, the only thing that cannot be measured in a hundredth of a second is the honesty of the person reading the numbers.

The question I leave for the next cycle is not who will break a record. It is: when the data table returns a blank line, will we choose to invent a number, or will we choose to sit still and go looking for the truth?