Empty Table, Silent Discipline: Notes on Swimming Data Analysis in Vietnam

At two in the morning I opened the analysis file for a domestic swimming meet. The table had twelve rows and six columns. Every cell was blank: no final time, no 50-metre splits, no stroke count per lap, no gap between two swimmers at the turn. I stared at that white space for a while, then noticed my hand was already resting on the keyboard. A writer's first reflex is to fill the gap. It is also the worst reflex I know.

I pulled my hand back in time. But had the file arrived on a Sunday evening with four hours left before deadline, I might have filled it in for real. A guessed interval in one column, a qualitative remark in another. Nobody would be able to verify it, myself included, because I would have had nothing to verify it against.

Domestic swimming has a paradox built into its data. International meets where Vietnamese athletes compete usually come with full electronic timing, splits published on federation pages, and multi-angle video. At domestic meets, what exists is usually just the final time on the electronic board plus a referee's handwritten sheet. If you want splits, you have to punch them yourself from video, and the video does not always cover all eight lanes.

Two streams of pressure meet in the middle. A major tournament season compresses everything: readers want to know immediately why their swimmer finished fourth, coaches want to know immediately which segment cost time, and the training staff need the number before the next morning's session. The analyst sits exactly at that intersection, holding a file that may not be full.

I started out in 2026 at Thanh Nien newspaper as a swimming reporter, when notes were kept in a notebook and results were read off the electronic board through binoculars. What I carried from swimming into analysis is a habit: measure distance and rhythm first, measure emotion later.

There are two kinds of missing data in this trade, and they are not the same. The first is sparse data: a few empty cells, but enough of the frame remains standing. With sparse data I allow myself to infer, provided I label the confidence and state plainly that I am inferring. The second is null data: no sample at all. With null data, every conclusion is a product of imagination. The honest result of an empty table is the phrase “cannot be assessed”, not a line of speculation written smoothly.

My process has three layers and none of them may be skipped. Layer one is raw data: times, splits, turn counts, kept in their original units, not rounded to look neat. Layer two is cross-verification with at least two independent sources: the organiser's electronic board and video, or the referee's sheet and federation data. When the two sources disagree, I note the discrepancy rather than picking whichever is more convenient. Layer three is writing. Only when layers one and two agree do I give myself the right to conclude.

Numbers only retell; tactics begin with mistakes. Knowing that a swimmer's first 100 metres were faster than the last 100 is a fact. The tactical question sits elsewhere: which of the four turns broke the rhythm, and whether the break came from fading arm force or from a cut in underwater dolphin kicks. A table full of numbers can still be useless if it does not help answer that.

My experience watching domestic meets suggests most arguments about Vietnamese swimming results are not about the final time but about how that time is sliced. A distance swimmer such as Nguyen Huy Hoang lives on his 100-metre split series. Lose one mid-race split and the pacing curve breaks, taking with it the ability to separate two very different scenarios: going out too fast, or turn technique degrading lap by lap. For a sprint event such as Vo Thi My Tien's, losing one detail loses almost everything. The absence is not neutral. It removes the right to ask the correct question.

My 2026 mistake reminded me that data is a mirror, not a lamp. That World Cup, I wrote about Belgium's 3-4-3 against Brazil and stated the team had made 21 successful presses. The real figure was 14. A reader flagged it that same night and I had to correct it. The error was not a shortage of data. The error was that I trusted my memory faster than I trusted the source. Since then every piece costs roughly three extra hours of verification, most of it spent punching splits from video and checking them against the official sheet before touching the interpretation.

During the 2026 shutdown I stayed home rewatching Liverpool's 0-3 defeat to Watford and measured the average distance between the defensive line and the goalkeeper: 28 metres, against roughly 15 metres in their wins. What matters is that I measured it without any official dataset. I punched it from video. That is an important anchor: some gaps exist because nobody bothered to measure, not because measurement is impossible.

The hardest part of an empty file is locating where it went empty. Before concluding anything I always check the collection stage: did the source file arrive, was there an encoding fault, were athlete names truncated. In many cases the data still exists; only the pipe carrying it snapped. An empty result is itself information: it points at the break, and the break usually sits in collection rather than in analysis.

The blind spot lies in the incentive structure, not in competence. Nobody in this trade loses points for a table full of numbers. A piece with twelve figures looks far more professional than one containing the words “cannot be assessed”, even when the second is right and the first is wrong. That structure pushes writers toward disciplined invention: subtle enough not to be caught, concrete enough to read smoothly.

I have fallen into that trap in a different way. Once the story is already imagined — the swimmer fading home because of a fast start — hunting for numbers becomes hunting for evidence for that story. I do not believe in intuition. I believe in how many variables that intuition has been loaded with. For someone who has watched thousands of races, intuition is rarely wrong, but it is not evidence, and it is not allowed to replace evidence.

The counter-intuitive view: publishing an empty result does not damage an analyst's credibility. It moves the problem back where it belongs. A coach who reads it will know to fix the record-keeping before the next session, instead of fixing a swimmer's technique on the basis of a rootless guess.

What I want to try next season is not predicting who breaks a record. I want to build one table with no cell filled by guesswork, even if that makes the piece shorter and less appealing. If I must choose between a piece that reads smoothly and a piece that reads correctly, I choose the latter. Will my readers choose the same, when both pieces sit on the same news feed?