Blank Spots in Vietnamese Athletics Injury Records Are Misread as Health

Blank Spots in Vietnamese Athletics Injury Records Are Misread as Health

At 5:40 in the morning, lane three of Mỹ Đình Stadium is still covered in mist. The young 400m squad enters its final rep: two hundred metres relaxed, three hundred metres braced, then everything released over the last eighty. I stand on the inside of the second bend, the spot every coach chooses because it is the only place where you can clearly see how the foot lands.

A male athlete, twenty-one years old, right leg. At the three hundred and twentieth metre, his right heel drives down and rolls inward, and his left hip drops about fourteen degrees. Nobody shouts. He still finishes, still stops the watch, still gets written down as "11.84 — good". In the coach's notebook, the remarks column stays empty.

Three months later, in the same bend, he sits down and holds his face.

I do not tell this story to frighten anyone. I tell it because the empty box in that notebook is the real protagonist. After fifteen seasons reading back through the injury records of Vietnamese athletics, I have come to one conclusion: we do not lack data about pain. We lack data about the days before the pain arrives. And that blank, again and again, gets read as health.

Context: an athletics culture running on feel

Vietnamese athletics runs on the coach's instinct and the athlete's memory. For decades, performance assessment has boiled down to three things: a stopwatch, a pair of eyes on running form, and the question "how do you feel today". All three are useful. None of them measures what I need to measure.

The problem is not human capability. It is the structure of the record. A provincial athlete trains four times a week, and each session the coach writes down distance and time. A national-team athlete trains nine to eleven sessions, sometimes twice a day. Both are recorded in the same kind of notebook: how far, how long, how it felt. There is no column for the load striking each foot. No column for knee-axis deviation. No column for hours slept, litres drunk, or how the Achilles feels first thing in the morning.

In other words: we record outcomes, not mechanisms.

I came into this work by accident. In January 2026, as Vietnamese sports media was rapidly expanding, the board of Sông Lam Nghệ An called me — a forty-five-year-old analyst — to assess the risk on a young defender about to move to Hà Nội FC for a fee of eight billion đồng. I built a homemade model I later called the "hip rotation coefficient", and it returned a 71 percent risk of anterior cruciate ligament rupture within ninety days. The transfer was postponed for two weeks. I was mocked across the forums. On day sixty-four, the player left the pitch in a friendly with exactly the injury I had predicted.

I do not retell that to praise myself. I retell it because it taught me how people read data. When a model says something the eye cannot see, the first reaction of the majority is denial. The second reaction, if the number turns out right, is to turn it into legend. Both reactions are useless. What is needed lies in between: treating the number as an ordinary column of data, recorded consistently, read the way you read a pulse.

In 2026, when the pandemic froze every competition, I sat at home and started building an open dataset called "Vietnamese Injury Code". It began with 547 profiles of players and athletes across fifteen seasons. Today the dataset is considerably thicker, but the proportion of blank space has barely shrunk. That is the most telling fact of all.

Core analysis: three layers of data and one missing tier

In my dataset, every injury case is recorded in three layers. The first is the event layer: where it hurt, when it hurt, what the diagnosis was, how long the layoff lasted. The second is the load layer: how much was run the previous week, the week before that, by what percentage intensity rose. The third is the biomechanical layer: landing axis, stride angle, hip rotation coefficient, pelvic tilt.

The first layer is the fullest, because it is what gets written into a medical report. The second thins out. The third is almost entirely absent at grassroots level and appears only sporadically at national-team level.

That is the architecture of a blind record.

The hip rotation coefficient: a measurable nobody measures

There is nothing mystical about the coefficient I use. It is the ratio between the rotation of the pelvis around the vertical axis and the rotation of the ribcage within the same running cycle. In a healthy runner, those two segments rotate in opposite directions with moderate amplitude, producing an elastic quality that offloads the lumbar spine and the anterior cruciate ligament. When the ratio drifts outside the safe band, shear force is dumped onto the knee and the hamstrings.

I measure it three ways, all of them cheap. The first is high-frame-rate video from behind, counting shoulder deviation against hip deviation at the moment of foot strike. The second is a pair of inertial sensors, one on the sacrum and one on the twelfth thoracic vertebra. The third — crudest and by far the most common in provincial programmes — is direct observation scored on a three-point scale.

The third way is the worst way, and it is still better than an empty box.

Among the athletes I track in the dataset, ACL ruptures are not random. Most of them showed a stable hip rotation deviation for at least six weeks before the tear. The word "stable" matters. A single deviation says nothing. A deviation that repeats across sessions, on the same leg, in the same stretch of track, is a signature.

The injury code is never in a single number. It is in the repetition.

What worries me is that the Vietnamese context amplifies the deviation. Most of our track athletes train on hard surfaces — concrete, asphalt, school playgrounds — while only a very small number of sessions happen on a properly specified synthetic track. Running on hard ground increases vertical ground reaction force and forces the body to compensate by rotating the hip further. That compensation, repeated daily, is exactly what I see in the video.

In other words, our training surfaces are quietly writing the verdict in advance.

The load diary: the fatal blank

The second data layer — load — is where I lose the most time building a case file. The basic principle is well known: the ratio between this week's volume and the average of the previous four weeks. When that ratio spikes sharply in a short window, injury risk rises measurably.

The trouble is that in Vietnam, training volume is usually logged by feel: "heavy week", "light week", "a bit tired this week". Those three phrases cannot be entered into any model.

When I interview athletes in the dataset who have already been injured, I ask one question: how much did you run the week before the pain? Most do not remember. Some remember vaguely. A few give a number but are unsure. Only the group whose coach kept a detailed log can answer.

That is why I keep telling young coaches one thing: your notebook does not need to be pretty, but it needs a kilometre column. That column costs less than any device.

A simple example shows its power. Suppose an athlete runs 60km a week for four straight weeks, then jumps to 95km in the fifth week to prepare for a meet. That is a rise of nearly 58 percent. For a novice, that is a high-risk zone. For a long-established athlete, it is a zone requiring close monitoring. But neither case can be noticed unless the kilometre column exists. Without it, the coach sees only one thing: "a heavy week".

And a heavy week is never considered wrong, until someone falls.

Surfaces, climate and the tropical trap

Vietnamese athletics trains and competes in conditions most international models barely account for. Average humidity is high, afternoon temperatures routinely exceed the thresholds international guidance treats as safe limits, and the rainy season changes track grip hour by hour.

Two injury mechanisms are specific to this environment.

The first is cramping and muscle spasm from dehydration and electrolyte loss. It is rarely serious anatomically, but it is a leading cause of sudden intensity reduction during a preparation block. A week cut short by cramp, then made up the following week, produces exactly the load spike I described above.

The second is injury from a slippery surface. When grip drops, athletes automatically increase knee flexion and lengthen ground contact time to stay balanced. Longer contact means the reaction force is stretched out, and soft tissue — Achilles, hamstring, plantar fascia — carries load for longer in every step. Multiply that by several thousand steps a session and you have the difference between fatigue and damage.

Here I have to say something plainly that many in the profession dislike hearing. Training and competing in tropical conditions does not automatically cause injury. Injury comes from failing to adjust volume and failing to track hydration. Weather is a multiplier, not a cause.

Shoes and carbon plates: a gift with an invoice

Over the past decade or so, racing shoes with carbon plates and supercritical foam midsoles have transformed elite athletics. The benefits are real and widely documented: energy savings, reduced muscle fatigue after competition, improved performance over longer distances.

But every gift carries an invoice, and that invoice is usually paid by the foot.

When the midsole thickens and stiffens in the forefoot, pressure shifts down onto the metatarsals and the plantar fascia. Athletes migrate from heel strike toward midfoot strike — good for speed, but it drags a new load onto structures that are not yet adapted. Among young athletes who move straight from training shoes into carbon-plated racers without an adaptation period, I have recorded a clear rise in plantar and metatarsal pain during the early phase.

The question is not whether to use good shoes. The question is how long the body needs to rewrite its own code.

I usually propose one simple rule: a new pair needs at least three weeks of familiarisation, starting with short, low-intensity runs, and during those three weeks the plantar sensation must be checked every morning. If the first step out of bed makes the plantar fascia feel tight, that is a signal.

Nobody needs expensive equipment to notice that signal. Only a question asked at the right moment.

The competition calendar: a system eroding itself

On the night of 12 June 2026, while commentating live on Denmark against Finland, I watched Christian Eriksen collapse in the middle of the pitch. The whole studio went silent. I said on air something that has been quoted many times since: we are killing the players with a packed calendar.

That was true of football. It is also true of athletics, except that in athletics it is less visible because athletes fall quietly.

Look at a four-year cycle for a reasonably strong Vietnamese track athlete. The year includes the national championships, the junior championships, one or two regional international meets. Add the regional multi-sport games every two years, the continental games every four years, and short overseas training camps. Add the demand for results on the domestic stage, where medal pressure flows down from the province, from the sector, from sponsors.

The result is that athletes live permanently in a peak-preparation state rather than a base-building state. Peaking twice a year is science. Peaking four times a year is slow self-destruction.

At the same time, the system has no mandatory rest mechanism. An athlete who wants two weeks off after a major meet has to negotiate with their own coach, and is often read as lacking commitment.

That is why I do not believe in promises of fast recovery. The problem is not individual willpower.

The counterintuitive angle: a blank is not evidence

Now to the part I consider most important, and also the most easily misread.

In medicine and in sport there is a classic reasoning error: failing to find evidence of risk gets read as evidence of safety. When an injury is not recorded, people assume there was no injury. When an athlete does not report pain, people assume the athlete is healthy. When the medical file is empty, people read it as a clean sheet.

That is a systematic error, and it is more common where record-keeping is weak. The paradox is that confidence is highest where data is thinnest. With nothing to contradict them, nobody has to correct themselves.

When the "Vietnamese Injury Code" dataset launched and I published the first cases with codes like ACL-07 and HAM-23, the reaction split in two. The first came from young coaches: they wanted to know how to measure. The second came from people with authority: they said I was overreaching.

I understand why. A coach who has given twenty years to the craft, hearing someone outside the track say his model is missing a signal, will naturally go on the defensive. I have no intention of dismantling anyone's work. But I cannot stay silent either.

I still remember Tokyo, six weeks after Eriksen went down. A gymnastics coach came to me for help with a nineteen-year-old athlete suffering a recurrent ankle injury. I proposed a counterintuitive approach I later named "reverse offloading": raise intensity by about fifteen percent for two weeks, then cut it by roughly forty percent abruptly. The national team doctor called it a con. I said: then let us bet. That athlete went to the Games without picking up any injury.

I do not tell that story to prove I am clever. I tell it because it illustrates something: what gets called a con is often just data read a different way. A short controlled load increase, followed by a sharp reduction, creates a mechanical shock that forces soft tissue to adapt while alert, then gives it rest before the second shock arrives. It is a game of timing, not of luck.

And here I have to remind myself of something I declared long ago: I do not prophesy. When I put out numbers like 71 percent or 62 percent, I am not reading the future. I am reading back what the body has already written, and comparing it with hundreds of similar files. If I have ever written in absolute terms, that was my error, and it was also what readers wanted to hear after a tournament.

An honest analyst must be able to say this sentence: if the file contains no data on an athlete, the only correct conclusion is "insufficient information", not "no risk". The silence of a record is not the voice of health. It is only silence.

Takeaway: one column to begin with

If I were allowed to add only one thing to the training notebook of a Vietnamese athletics squad, I would not ask for machinery. I would ask for a column.

That column records, every session, for every athlete, a score from one to three for the landing axis of each leg at race pace. One is stable. Two is slightly off. Three is clearly off. Thirty seconds a session. No cost. No specialist required.

A nineteen-year-old in a mountain district can do it after being shown once.

The value of that column is not in any single number. It is that three weeks before someone falls, the squad will have something to look back at. And in those three weeks, one decision might have gone differently.

Injury is the only thing on the track that never negotiates. But it is a superb negotiator with our time — and we are almost always the side that nods first.

The question I leave for those working in athletics: if next year one of your athletes has, for the first time, a notebook with that column in it, will you be patient enough to read it in silence before anyone calls the injury by name?

Data box

Codes in the "Vietnamese Injury Code" dataset: ACL is the anterior cruciate ligament, HAM the hamstring group, ACH the Achilles tendon, PF the plantar fascia, MT the metatarsals. Each case is tagged with a code and a season.

Landing-axis scale: 1 stable, 2 deviation under ten degrees, 3 deviation of ten degrees or more, or deviation persisting across three consecutive sessions.

Reading rule: one deviation proves nothing. Three deviations on the same leg in the same stretch of track are a signature.

Glossary

Hip rotation coefficient: the ratio between pelvic rotation and ribcage rotation within one running cycle.

Acute-to-chronic workload ratio: the ratio between the current week's training volume and the average volume of the previous four weeks.

Stride angle: the angle formed by the leading and trailing thighs at the moment the lead foot strikes the ground.

Running cycle: the interval from one foot strike to the next strike of the same foot.

Carbon-plated shoes: racing shoes containing a stiff carbon-fibre plate within the midsole to save energy during late propulsion.

Reverse offloading: a method of briefly and deliberately raising intensity, then cutting it sharply, to force soft tissue to adapt while alert and then rest it before the damage threshold.

Return to play after injury: the process of bringing an athlete from medical recovery back to competition at maximal intensity.