The Empty Wind Column: Four Checks Before Trusting Any Vietnamese Athletics Number

May 2026, Phnom Penh. Nguyen Thi Oanh stepped onto the track for the 1500m, then came back roughly twenty minutes later for the 3000m steeplechase. Most spectators remember two gold medals. I remember an empty box on the results sheet: the wind-reading column.

Three numbers open this story. +2.0 m/s — the tailwind threshold beyond which a mark is still ratified as a performance but no longer eligible as a record. 40 mm — the maximum sole thickness for a plated racing shoe, the device that introduced a systematic performance dividend every cross-era comparison must deduct. And close to 0 — the number of wind columns fully completed in most domestic Vietnamese results sheets I have had reason to revisit.

Vietnamese athletics is in a phase where marks appear faster than our ability to verify them. An Asian medal, a long jump beyond 6.50m, a 400m hurdles split under 57 seconds — any of these can become a headline within ten minutes. But a headline is not data. And data, in order to speak, needs a frame.

Method: four identity lines, four metadata lines

When I build an athlete file, I write four identity lines before I write any number: who, which event, which competition, which date. Those four are necessary conditions. Sufficient conditions are four metadata lines: wind reading, altitude and surface, footwear, and the number of times that mark has been repeated within the season.

Based on my experience tracking domestic and regional athletics meets, most athlete files handed to me by newsrooms contain only the first line. The name is correct. The other three are usually missing, and the four metadata lines are almost always missing.

That does not make the mark meaningless. It makes the mark not yet usable as evidence. This is the distinction I want to spend this article on, because in three years I have seen it ignored more often than any other analytical error.

Wind — the first bracket

In sprints and jumps, wind is the decisive variable. A 6.60m long jump with a +1.8 m/s tailwind and a 6.60m long jump with +3.4 m/s are different events in substance, even though the scoreboard shows the same number. The first sits inside the legal envelope. The second still reveals ability, but cannot establish a record, and cannot be compared with a jump in an indoor arena with closed airflow.

At many domestic meets, the anemometer is either not placed correctly or not recorded in the official report. The result is a gap in the file, and that gap is usually filled with an assumption. The default assumption — that the air was still, that conditions were normal — is an assumption that favours whoever produced the mark. That is why I never accept it.

For Bui Thi Thu Thao, the long jump gold at the 2026 Asian Games in Jakarta with 6.55m is a traceable landmark: named competition, named date, named round. That is the kind of data I can put on the table. Most jump and sprint marks reported in Vietnamese media do not carry that same completeness.

Terrain, surface and heat

Altitude above 1000m assists sprints and jumps while penalising endurance events. Vietnam has almost no track at that threshold, so this is a variable we rarely have to deduct. But there is an inverse variable we routinely forget: temperature and humidity.

A 5000m race in Hanoi in June, under 35 degrees and 80 percent humidity, cannot be read on the same scale as a 5000m race in Europe in September. Without correction, we conclude the athlete ran slower — when in fact they ran in a harsher environment and their result is worth more than the raw number shown.

This is what I call the inverse distortion: it does not inflate a mark, it strangles it. And it is more dangerous, because nobody sues a number that was undervalued.

The surface matters the same way. My Dinh stadium on a May evening and a new synthetic track at a provincial training centre are two different surfaces. At 400m, the difference between them can reach several tenths of a second — enough to change a SEA Games final placing.

The footwear dividend

This is the most neglected variable in every Vietnamese record debate. A racing shoe with a rigid plate and a supercritical foam midsole generates a systematic dividend, not for one athlete but for an entire generation. When we compare a 2026 mark with a 2026 mark, we are comparing two technologies.

I am not saying the new marks are not real. I am saying that a cross-era comparison without deducting the equipment dividend is methodologically invalid. It is like comparing two motorcycle racers across eras without caring about engine displacement.

In Vietnam, access to high-technology racing shoes is uneven. Some national-team athletes are sponsored, some provincial athletes buy their own, some train in training shoes. Those three groups should not sit in the same comparison table. When a young provincial athlete breaks through suddenly after moving to a major centre, the first thing I check is not pharmacology — it is equipment and training conditions.

Sample size: one race is not one level

This is what I want to say to everyone working in sports journalism. One mark does not constitute a level. Two marks in two weeks do not either. You need at least three independent repetitions, at three different competitions, within the same training cycle, before you are entitled to say "this athlete has reached a new tier".

I do not believe in luck; I believe in what has been repeated enough times.

There is a quick check I use constantly: the gap between this season's best and the personal best. If the season's best equals or exceeds the personal best abruptly at age 29, that is a story about maturity or about something else. If it sits 3 percent below the personal best at age 24, that is a story about building a base, and there is no reason to panic.

Distance never lies, it is only that we have not been patient enough to listen.

Splits: the most expensive data and the most discarded

A results sheet gives you a finishing time. It does not give you a story. The story lives in the splits — every 200m in an 800m, every 400m in a 1500m or 5000m.

A 5000m runner whose first and second halves differ by 4 seconds is distributing evenly, with reserve. A runner with the same finishing time but a 22-second differential is borrowing in the first half and paying interest in the second. Two athletes, one identical number, two different futures. Read only the finish time and you will pick the wrong athlete for a championship with heats and a final.

At minute 70, the crowd sees collapse; I see a structure being rebuilt.

In sprints, what gets discarded is reaction time and maximum velocity reached. A 100m athlete with a 0.19-second reaction time is losing a margin that conditioning cannot fix — it has to be fixed with starting technique. But the results sheet has no column for it, so nobody fixes it.

The counterintuitive angle: correlation is not causation

International sports analytics uses a screen for anomalies: if an athlete's one-year improvement exceeds roughly three times their own historical average annual improvement, that is a red flag requiring review.

I want to state plainly what many people abbreviate: a red flag is not a verdict. It is an instruction to look further.

A 20-year-old moving from a provincial coach to a national centre, fed according to a nutrition plan, monitored with load-tracking devices, recovered in ice baths and sent to three international competitions in a year — that athlete can improve two seconds over 1500m with nothing unusual happening. A 27-year-old with the same coach, the same training volume, improving by the same margin — that requires questions.

The correlation between a performance leap and a violation does not exist; the correlation between a performance leap and a change in training environment does. Anyone working with data has a duty to separate those two correlations before writing.

And there is a more dangerous error, one I encounter repeatedly in reports sent to me: reading "not assessed" as "low risk". When a file has no injury data, no testing data, no competition schedule data, the correct conclusion is "cannot be assessed". Not "clean". Not "fine". Emptiness is not evidence of innocence, and it is not evidence of guilt either.

This is why I always write the words "not assessed" rather than leaving a box blank. A blank box will be filled by the next reader with the most favourable assumption. An explicit line will not.

Another reading of decline

In athletics there is a habit of reading a slower second half as a sign of downward trajectory. I read it differently, provided I have split data from at least two prior races.

If a 5000m runner three months ago had an 18-second differential, two months ago 12 seconds, and this month 7 seconds while the finishing time was 20 seconds slower — that is not decline. That is someone learning to distribute effort, paying in finishing time to buy a more sustainable race structure. That price is usually repaid within two to three cycles.

Conversely, an athlete holding the same finishing time while the split profile steepens is living on an old base. They will not decline gradually. They will collapse in a single evening.

When the numbers speak, all I have to do is listen.

The road into a major championship: two doors, not one

For world championships and the Olympic Games, World Athletics opens two doors: hitting the entry standard, or accumulating world ranking points. The weighting between the two shifts by event and by cycle.

This means an athlete can reach a major championship by racing many smaller meets consistently rather than through one explosive performance. It also means a SEA Games medal does not automatically open either door. Two systems, two logics.

The third door — and the most volatile — is national selection. A centralised trials system of the "one afternoon decides everything" type creates an entirely different risk profile from a multi-stage evaluation system. Same athlete, two risk levels. This is the variable anyone forecasting for a national team must model first, not last.

Every number is a confession the race cannot deny.

Signals for the next cycle

Three things I will track in the coming cycle, and three things I suggest anyone working with Vietnamese athletics data should track alongside me.

The first is the wind column. Not because wind changes marks, but because the existence of a wind column signals that the organisers have a measurement process. A meet with a complete wind column is a meet you can trust on the other columns.

The second is split data. A meet that publishes splits is producing data reusable for years. A meet that publishes only finish times is producing headlines, not files.

The third is the repetition rate. Within a season, how many times did an athlete perform within 1.5 percent of their own best? That number describes consistency, and consistency is the only thing that reliably predicts a final.

I still keep a notebook with handwritten lines from a coach in Nha Trang who times athletes by hand on lane four every morning. He has no anemometer, no analytics system, no software. But he records the date, the time, the temperature, the athlete's own feeling after every run — and he has done it twelve months a year, three years running.

Across everything I hold on the province's young athlete cohort, the most trustworthy dataset is that notebook. Not because it is precise. Because it is complete.

An empty wind column can hide a record. A complete handwritten notebook can reveal a champion. The job of the person reading data is to know which one is in their hands.