The Wind Gauge, the Shoe, and the Data Gap in Vietnamese Athletics

The Small Box Beside the Track

The scoreboard clicks over: 10.47 seconds. Three metres away, a small box the size of a cigarette pack displays the wind speed: +2.3 m/s. Two numbers sit side by side on the same board, in the same instant, for the same athlete. One of them will enter the record book. The other will erase it.

Nobody in the stands looks at the box. The crowd looks at the clock. The coach looks at the clock. The parents, the journalists, the person filming from the infield — all of them look at the clock. The wind gauge is only mentioned when someone wants to discredit another person's performance, and so it carries a bad reputation: the spoiler, the one who arrives late to strip away the joy. But the box is not at fault. It merely records a condition. The problem is that we have grown used to reading a performance as an absolute fact, when a performance is only a conditional fact.

I read conditions for a living. For years, my job has been to tell coaches that the number they just saw does not mean what they just thought. It is an unpleasant professional habit, and in Vietnam it is even more unpleasant, because here sports data is still often treated as decoration: nice to have, fine to lack, as long as the medal still comes home.

The day football stopped, I began counting steps again.

In 2026, the football leagues froze. I lost my live data feed, lost my match-watching sessions, lost even the habit of sitting and counting the sideways passes of a team winning 1-0. Over four months I went back to what I had accumulated: 2,300 matches spanning five V.League seasons and three major European leagues. I reassembled them, recalculated them, and tried to build a new pressure index combining PPDA, defensive distance and pressing speed. The group of teams with an average PPDA below 8.5 collected 1.8 points per match, well above the rest. I published the "Pressure Index" model on my personal blog. A few domestic analysts left comments. That was the first time I understood that a data person in Vietnam can still create something, provided he is willing to sit down with old numbers instead of waiting for new ones.

But football is a dirty laboratory. Too many variables: eleven players per side, the pitch, the weather, the referee, psychology, the opponent's tactics. Athletics is different. Athletics is the cleanest laboratory mass sport can offer: one person, one distance, one clock. Precisely because it is clean, it exposes the trap of every measurement — and that trap is called condition.

From the Hải Phòng Pitch to the 400-Metre Track

In 2026, aged 31, I worked as a data consultant for Hải Phòng FC. During a review of the youth academy's metrics, I came across a name few people mentioned: Vũ Minh Hiếu, a midfielder, modest in build, unremarkable in any training session the coaching staff had watched. But his average PPDA was 6.8, the highest in the entire academy system. It meant he pressed extremely well, chose his positions well, and did so so quietly that nobody noticed.

I brought the spreadsheet into the meeting room. I presented it, and I remember the discomfort in the room clearly: an outsider to the pitch telling people who had lived by football that their eyes had missed something. Coach Trương Việt Hoàng listened to the end and gave Minh Hiếu his chance. In the match against Hà Nội FC on V.League round 17, he won the ball 14 times, provided one assist, and Hải Phòng won 2-1.

I do not tell this story to praise myself. I tell it because it shaped how I have worked ever since: for every young player, I put PPDA and ball-recovery counts first, and names and impressions second. Hải Phòng taught me: the star is not on the shirt, it is in the metric.

When I shifted my focus to athletics, I carried that same habit with me, and I discovered that Vietnamese athletics is missing something Vietnamese football also once lacked: a record-keeping system dense enough for the right question to be asked.

One thing about context needs saying clearly. Vietnamese athletics is not weak. At regional level we are strong in middle- and long-distance running, race walking, long jump, and relay events. Nguyễn Thị Oanh is the defining figure of this generation, with three gold medals at SEA Games 31 on home soil in Hà Nội in 2026, in three events demanding near-opposite physical qualities, contested within a very short window. Hoàng Nguyên Thanh won the men's marathon gold at SEA Games 32 in Phnom Penh in 2026. Bùi Thị Thu Thảo once topped the women's long jump at the 2026 Asian Games in Jakarta with 6.55 metres. These are real facts, and they deserve full recognition.

But behind those facts lies a gap. We have medals without split tables. We have performances without condition notes. We have names without comparison baselines. An athletics programme can advance for a few years on inspiration, but it advances durably only on method. And method begins with being willing to read the numbers correctly.

Wind: The First Variable the Scoreboard Never Prints

World Athletics technical rules state plainly: a sprint or long jump performance can only be ratified as a record when the tailwind does not exceed 2.0 metres per second. Above that threshold, the performance still exists, still enters the competition report, can still be ranked within a specific race, but it is no longer a record. It becomes data of a different class.

Very few spectators understand this. They see 10.47 seconds and think immediately of a national record. They do not think that the same athlete, in the same form, running with a +2.3 m/s tailwind and running into a -1.0 m/s headwind, can differ by three to four tenths of a second over 100 metres. Three tenths of a second, in Vietnamese men's sprinting, is the distance between a SEA Games berth and a seat at home.

The correct reading is not to reject wind-aided performances. The correct reading is to place them inside a window of conditions. When I review domestic meet data across years, I always split performances into three groups: wind-legal (from -2.0 to +2.0 m/s), strong tailwind (above +2.0 m/s), and no wind reading. The third group is the most worrying, because it accounts for a significant share of grassroots meets, and every performance in it is an unanswered question.

There is another common error: comparing two athletes' marks from two different meets without cross-checking wind. An athlete running 10.50 seconds with a +1.8 m/s wind and another running 10.60 with a -0.4 m/s wind are effectively at different levels, and the second is the faster. The ranking list records 10.50 and 10.60. The ranking list does not record that.

At major international meets, organisers record wind for every run and every jump, and that data is published openly. That is the minimum standard. At many domestic meets, the wind gauge is either absent, or present but unrecorded, or recorded but unpublished. The result is that we accumulate a performance archive we ourselves cannot verify.

I once proposed to a few meet organisers that they print a wind column in the official results, beside the performance column. The reply was: nobody needs it. Perhaps that is right, if the only goal is handing out medals. But if the goal is knowing where we stand, the wind column is the most important column in the table.

Altitude: A Gift Available Only in Certain Places

At altitudes around 1,000 metres and above, the air is thinner, drag is lower, and sprint and jump events benefit markedly. Endurance events, by contrast, suffer, because available oxygen drops. This is a simple law of physics, but its consequences for the record books are anything but simple.

Bob Beamon's 8.90-metre long jump on 18 October 2026 at the Mexico City Olympics, at a venue roughly 2,240 metres above sea level, stood for 23 years. It was a monumental performance, and nobody disputes that. But in reading it, one is obliged to note that the stadium sat at an assisting altitude. When Mike Powell jumped 8.95 metres in Tokyo in 2026, he did so near sea level. Two numbers sit side by side in the book, from two entirely different conditions. No footnote in the record book forces the reader to know this. The reader must know it.

For Vietnamese athletics, this story has a notable reverse side. Almost all domestic meets take place on flat terrain, near sea level. That means our national records are "clean" in altitude terms, receiving no assistance. Technically, that is something to be proud of. But it also means we have never had the chance to observe our athletes in an assisting environment, to know where their true ceiling lies.

Countries with altitude training conditions, or the money to send athletes to train high, gain two kinds of benefit: long-term physiological adaptation, and the opportunity to record marks in favourable conditions. We can pursue the first only if funded, and we almost never pursue the second at home.

So when comparing a Vietnamese athlete with one from a highland region or from a country with an altitude training system, treat the comparison with caution. The results list does not record the altitude of the venue. It records only the final outcome.

The Shoe: A Technology Race and the Cost of Silence

No part of modern athletics has changed as quickly and as quietly as the shoe.

In 2026, at the Monza track in Italy, an experimental project was staged with the goal of taking a human being under two hours over 42.195 kilometres. Eliud Kipchoge finished in 2 hours 00 minutes 25 seconds, missing the target, and the mark was not recognised because of the pacing car and rotating pacers. But the shoe he wore that day changed the sport forever.

In 2026, a study published in the journal Sports Medicine indicated that a new generation of racing shoe with a carbon plate and a super-foam midsole could improve running economy by roughly 4% compared with previous traditional racing shoes. Four per cent. For a marathoner running 2:10, four per cent equals about five minutes. For a world record, it is a decade of progress compressed into a single shoe.

In October 2026, Kipchoge ran 1:59:40 in Vienna. It was a perfectly staged event, not a race, and the mark was not recognised as a world record. But it proved that the two-hour barrier no longer lay at a purely biological limit. It lay in a combination of fitness, conditions, tactics and equipment.

In January 2026, World Athletics published new competition shoe rules: a maximum sole stack of 40 millimetres, a single rigid plate, and the model must have been available on the open market for at least four months before the competition. This was a reasonable attempt to prevent a situation where each athlete has a bespoke shoe nobody else can buy. But the rule does not answer the bigger question: how much of a performance belongs to the human being, and how much to the shoe?

In 2026, in Berlin, Kipchoge set the official world record of 2:01:09. In October 2026, in Chicago, Kelvin Kiptum ran 2:00:35, breaking that record in an official race. In February 2026, Kiptum died in a road accident. He was 24, and the clock stopped at a point where nobody can know where it would have stopped.

In October 2026, also in Chicago, Ruth Chepngetich ran 2:09:56, becoming the first woman to run a marathon under 2 hours 10 minutes. She broke the previous world record by more than two minutes, a gap that for decades almost nobody thought possible.

In Vietnam, the wave of high-technology racing shoes has arrived, but by a different route: the mass running movement, expanded marathon events, weekend running groups. This has been one of the most positive social changes of the past decade. But in data terms, we are forgetting a variable. When an amateur runner smashes a personal best by twenty minutes, the first question I ask is not how they trained. The first question is whether they changed shoes.

No mass-participation race in Vietnam records an athlete's shoe model alongside the result. That means our entire amateur performance archive is mixing data from two different technology eras into a single column. A problem that cannot be solved with the data currently available.

Splits: The Truth Lies Between Two Checkpoints

Finish time is the easiest thing to measure and the least informative.

A marathoner finishing in 2 hours 20 minutes may have run evenly throughout. He may also have run the first half in 1:08 and collapsed over the final ten kilometres. Two athletes arriving at 2:20 in those two ways are not the same class, do not share the same ceiling, and do not need the same strategy to improve.

The split table is what answers that question. Looking at five checkpoints every ten kilometres, you see the shape of the race. Looking at 5 km columns, you see the shape of each tactical segment. An athlete with a negative split, running the second half faster than the first, has a reserve base entirely different from one with a strong positive split. In coaching circles, a heavy positive split over the last ten kilometres is the clearest sign of depleted reserves, and it points to what needs fixing in the next programme.

Over middle distances, splits matter even more. In the 800 metres and 1,500 metres, 200-metre and 400-metre splits distinguish two athletes with opposite tactical personalities: the even pacer and the finisher. Those two types need different training programmes, different psychological handling, and different race tactics. If all you have is the finish time, you cannot know whom you are coaching.

I have spent many sessions rewatching domestic meet footage to time each lap by hand. It is manual, slow, and error-prone work, but it is the only method when official data does not exist. Once I timed out a split showing that an athlete ran the first 400 metres far faster than her capacity allowed, and that explained the entire collapse over the final 200. Looking only at the finish time, one would conclude she lacked speed. The truth was the reverse: she had speed, but lacked distribution.

In developed athletics nations, split data is published for every race at major meets, and coaches treat it as a mandatory input. Here, it remains a luxury. This is the cheapest technical gap to close, and the one with the fastest return. One timekeeper per lap, one spreadsheet, one email to the coach. No imported equipment required.

A season is a confession of tactics. But a confession can only be read if someone writes down every sentence.

Small Samples, Sudden Leaps and the Biological Passport

I have one professional rule I have kept for years: never judge any athlete on a single competition. With one performance, I have a point. With three to five performances in a season, I have a trend. With three consecutive seasons, I have an assessment.

The reason is simple. One good competition can be the product of a nice day, a rival setting the pace, a psychological state that will not recur, or simply probability. In athletics, where error margins race against human limits, a single sudden jump is statistically weak data, however impressive it looks.

This matters more than it appears. In long-term performance analysis, one simple quantitative check is used: if an athlete's improvement within a season exceeds roughly three times their own historical annual gain, that is a signal requiring cross-validation, not a conclusion. The signal accuses no one. It merely demands more data.

Where does that data come from? From two systems international athletics has built over more than a decade.

The first is the Athlete Biological Passport, introduced in 2026 under the World Anti-Doping Agency. Rather than relying on a single test, the system monitors an athlete's biological markers over time and detects anomalies a single test cannot reveal. It is a trend-analysis tool, not an instant detector.

The second is the whereabouts system. Elite athletes must file information on their location to enable no-notice out-of-competition testing. Three missed tests or three filing failures within twelve months constitute a violation, even with no positive test result.

Both systems fall into the category that smaller athletics nations often lack or hold only thinly. And this is where we must speak plainly: in a sport system short on verification data, an athlete's cleanliness cannot be proven by silence. Nor can it be suspected by silence. Both directions are reasoning errors. The only honest move is to state clearly that the data is missing.

Data is a mirror. Most of the market looks into it and sees only itself.

Entry Berths: Performance Is Never the Only Condition

There is a very common misunderstanding among fans, and it is not their fault: that running fast enough automatically earns a place at the championships. The actual mechanism is far more complex.

At major championships such as the Olympics or the World Championships, there are two routes in. The first is meeting a qualifying standard published by World Athletics within a defined window. The second is accumulating World Ranking points across scoring competitions. Each event then has a maximum number of athletes per country.

That cap produces a familiar paradox: in an event where one country has four athletes meeting the standard, the fourth-place athlete stays home, even though their mark was good enough to compete. In a weaker event, an athlete with a far lower mark may travel. This is not injustice. It is structure. But it means that when reading news about a berth, you need to know which route it came through, and whether a national quota shaped it.

There are more complex cases still. Neutral athlete status applies to countries suspended from membership, allowing athletes to compete without representing a national flag. Nationality-change waiting periods also create berth gaps that can be predicted in advance but are usually ignored in the reporting.

And there is a body of rules that anyone analysing athletics must know without ever speculating: eligibility regulations for certain women's events between 400 metres and 1,500 metres, concerning testosterone thresholds for athletes with differences of sex development. The Court of Arbitration for Sport ruled on these regulations in 2026, clearing the way for World Athletics to apply a threshold of 5 nanomoles per litre to that group of events. This is a field with genuine scientific and legal debate, and it is not the kind of subject for a sports writer to adjudicate on instinct.

For Vietnamese athletics, most regional berths are determined through the national selection system. There, the selection model decides a great deal. A model resting on a single selection race creates enormous peak pressure and often fails to pick the athlete in the best form across the whole cycle. A composite model based on multiple competitions, training data and physical metrics would be fairer, but it demands a record-keeping system we do not yet have. Until we do, we will keep choosing on a single afternoon.

Peaking and the Calendar

In athletics, peaking is not a natural state. It is a designed product.

A good training cycle pushes fitness up, holds it there for a short window, and the peak of that window must coincide with the most important competition date. If the peak arrives two weeks early, the athlete competes in a declining state. If it arrives two weeks late, the athlete has not yet converted fitness into performance.

This makes the competition calendar an analytical variable, not an administrative fact. The number of races before the target meet, the spacing between them, the number of events contested within one championship — all of it consumes finite physical resources.

Take the competition structure at a recent SEA Games. A Vietnamese athlete contested multiple endurance events at the same championship, with heats, semi-finals and finals interleaved over a few days. That is a colossal physical and recovery challenge, and it makes the performance in the last event hard to read as a pure measure of capability. When I analyse such cases, I always separate two layers: the capability layer (the best mark achieved under peak conditions) and the management layer (how resources were allocated across the whole meet). The two layers can diverge widely, and collapsing them into one number is an analytical error.

Applied to Vietnam, there is a structural distortion worth naming: because regional berths are limited and because some of our events are thin in depth, we tend to enter a strong athlete in multiple events to maximise the medal count. That is a reasonable choice for short-term results. But it generates a data record in which that athlete's performances no longer reflect true capability, making all long-term analysis harder. Injuries in sport are random. But the number of times an athlete must start within four days is not random at all.

The Vietnamese Track: What Exactly Are We Measuring?

At this point I want to raise the issue one level deeper: the data system of Vietnamese athletics itself.

At the athlete level, we have stopwatches and coaches keeping training logs in notebooks. That is a decent foundation, but it is not digitised and not standardised, so it cannot be queried.

At the competition level, we have official results from national championships and youth meets, but they lack condition notes, lack splits, and are often not archived in an analysable format.

At the international level, we have public World Athletics data for athletes competing abroad, but that is someone else's data, measured to someone else's standard.

These three layers cannot talk to each other. That is the core problem.

I once tried something small: reconstructing the performance progression curves of about thirty young Vietnamese athletes across several years, using public results from domestic meets. The work took weeks, and the final product came with far lower reliability than I had hoped, because the data lacked condition notes. But one finding stayed with me: most young athletes show a very steep progression curve in their first two years, then flatline or decline. That does not necessarily reflect a problem with them. It may reflect the fact that we have no tool for detecting and correcting deviation before it becomes a trend.

In a sport system with good data, a flat progression curve in year three triggers a review: check the programme, check training volume, check accumulated micro-injuries, check recovery quality. Here, it usually triggers only one question: should we change direction.

Correlation Is Not Causation, and I Have Been Wrong

This is the part I want to write most frankly, because if I only presented the times I was right, this whole article would be worthless.

In 2026, I predicted Germany would be eliminated in the World Cup group stage. My basis was qualifying data: Germany's average PPDA sat at 9.2, far too high against the pressing standard of a champion, compounded by slow attacking speed and only mid-range final-third xG. I published it before the tournament. Social media mocked me, saying I only knew how to look at numbers. On the night of 27 June 2026, Germany lost 0-2 to South Korea despite 26 shots and around 1.5 xG, and went out. My old articles were shared thousands of times.

I did not see Germany lose. I saw numbers that do not know how to lie.

But the story did not end there. Months later I asked myself a harder question: if Germany had beaten South Korea 1-0 with an 89th-minute goal on exactly those same metrics, would I have sat down to re-examine my model? Honestly, I might have logged the wrong result with the right reason, and I would never have been forced to look again.

That is the biggest trap in data work: a correct result does not mean correct reasoning, and a wrong result does not mean wrong reasoning. Only the process deserves to be judged.

I have also been wrong in the opposite direction. Once I took a pressure model built on football data and applied it to assess a young athlete in another sport, on far too small a sample. The model produced a beautiful result, and I believed it. When that athlete moved up to a higher level, the model predicted nothing. My mistake was not in the formula. My mistake was forgetting that a model built on a small sample is a hypothesis, not evidence.

Since then I apply one principle in every athletics analysis: before comparing two athletes, I must check whether the two datasets share the same window of conditions. If not, I state clearly that the comparison carries undetermined error. That is what I learned from "flexible standardisation": standards are necessary, but a standard applied in the wrong place manufactures false certainty.

People call me a data monk. A monk needs no cathedral, only the truth.

What Data Cannot Measure

The most important counter-argument I have to write, and it is against myself.

This entire article is building an argument: read the conditions, read the splits, read the denominators, read the calendar. But there is a trap at the end of that road, and it is more dangerous than believing in superstition: believing that what can be measured matters more than what cannot.

Data has a natural tendency to pull attention toward what is easy to count. Run time is easy to count. Ball recoveries are easy to count. Possession share is easy to count, and that is why it is the most deceptive metric in football: a team can farm 60% of the ball through meaningless sideways passes between two centre-backs, and the number on the board still looks good. But a young athlete's sleep is hard to count. The quality of a recovery session is hard to count. Psychological state before a key start is hard to count. The quality of a coach is almost impossible to count.

That does not mean those hard-to-count things do not matter. It means we must clearly note that our models leave them blank.

In the context of Vietnamese athletics, that gap has a specific shape. We can know how long an athlete takes to run 1,500 metres. We usually do not know how many hours she sleeps each night during a peak week, how her menstrual cycle interacts with the training programme, or how many minor injuries she has carried without telling anyone. Those variables have far greater explanatory power than a tenth of a second, but they never appear on the scoreboard, so they vanish from the story.

And here is where I must be careful about my own profession. If I write about athletics using only publicly available metrics, I create an illusion of completeness. Readers will believe that what does not appear in the table does not exist. That is a methodological error, and it is worse than not analysing at all, because it dresses ignorance in scientific clothing.

The only way to avoid that trap is to state the limits clearly. I do not know this athlete's true progression curve, because I have only three data points and no condition notes. I do not know whether she has championship potential, because I have no physiological data. I do not know how much her shoes contribute, because nobody recorded them.

An analysis without a limitations section is propaganda with tables.

Signals to Track in the Next Cycle

If I had to pick the smallest changes capable of producing the biggest difference for Vietnamese athletics in the coming years, I would order them by cost, from cheapest to most expensive.

First, record wind in the official results of every meet with sprint and jump events. The cost is near zero. The long-term value is enormous.

Second, record splits for at least the middle- and long-distance events at national championships. One timekeeper per lap, one table, one public file. This is the data that turns a finish time into a behavioural record of a race.

Third, record athletes' shoe models at major mass-participation races. It sounds trivial, but it is the condition that keeps our amateur performance archive analysable five years from now.

Fourth, build a national database storing young athletes' performance progression year by year, alongside injury notes and training-load notes. This is the most expensive item, but also the one that makes the biggest difference between a sport driven by inspiration and a sport driven by method.

While waiting for those to take shape, there is one thing every fan can do right now, for free: when you see a performance, ask two more questions. Under what conditions was it achieved? And how many similar performances has this athlete produced before?

Those two questions will not make you an expert. But they take you out of a world of numbers that tell only half the truth.

The ball rolls in only one direction, but data can see in every direction.

And the 400-metre track, like everything else in sport, never has only one direction to be read. It has conditions, context, history, and gaps we have not been humble enough to admit. The final question is not who runs fastest. The question is whether we have enough data to know that — and if not, whether we are willing to say so.