Wind, Shoes and Small Samples: How to Read a Track and Field Mark When the Data Is Not Enough

On 7 July 2026, at the Meeting de Paris, Faith Kipyegon ran 1500m in 3:49.04 and became the first woman to break 3:50 for the distance. The moment the replay ended, the group chat of sports people I belong to split in two within seconds. One side talked about the limits of the human body. The other side asked how many millimetres thick her shoe sole was. Both sides had a case, and both concluded before measuring.

I stayed with my sheet of paper. For years, when I assess an athlete's case, I work through nine groups of questions: the performance and the conditions that produced it; the state of the body; the qualification mechanism; the shape of the event itself; rules and anti-doping; the training system behind the athlete; the risk landscape; the media narrative in circulation; and the way the effect travels through the industry. Those nine groups are like nine windows. What interests me is not which ones are lit, but which ones are empty.

Some files arrive with all nine windows empty. No competition date, no wind reading, no splits, no source. The most honest conclusion is also the shortest: there is not enough data to assess. In this trade, writing that sentence is not evasion. It is a finished product, and usually the lowest-paid one.

A season without a big stage

This year is an annual season. No Olympics, no outdoor World Championships, no single evening in which an entire career changes colour. From the outside it looks light. From the inside it is heavy in a different way: it is the season of accumulation, of races nobody remembers by name, and of the ranking system that decides who competes on the big stage next year.

Professional athletics runs on two currencies: entry standards and ranking points. The entry standard is a fixed number, published in advance, non-negotiable. Ranking points work like a sliding window: the system takes a set of a competitor's best results inside a defined period, adds placing points, and averages them. That mechanism has a consequence few people notice. A mark produced in perfect conditions and a mark produced three weeks after illness are worth the same on the board. The system does not ask how the number was made. It only asks what the number is.

For Vietnamese athletics, the gap between the number and the way the number was born is even wider. The domestic calendar is thin: a national championship, a few junior meets, a few invitation meets, and selection trials. Electronic timing appears at major meets. Wind gauges are not always present. Split data is almost never published. So when a Vietnamese athlete steps onto an international stage, the mark enters the system stripped bare: a number without a shape.

Watching domestic meets over many years, I have sent the same question again and again: what was the wind reading for that jump, and what was the first 400m of that race. Most answers are silence. That silence is neither good data nor bad data. It is a gap, and a gap always means something.

Alongside that sits the pressure of the calendar. Vietnam's leading athletes routinely contest several events across consecutive days, sometimes two finals in one session. Nguyễn Thị Oanh is the example domestic audiences know best: an athlete who can appear in the 1500m, the 3000m steeplechase, the 5000m and the 10,000m at the same Games. A mark produced under that schedule is worth more athletically than the number suggests, because it was paid for out of reserves that had already been drawn down. The record book does not record that.

The annual season is therefore a season of missed signals. With no medal to frame the story, people read a results board like a verdict. What deserves reading most usually sits in the footnotes: when the race took place, after how many days of rest, in what conditions, against whom, and in which shoe.

When the wind is at your back

World Athletics enforces a hard threshold for wind speed: 2.0 m/s. In the 100m, 200m, hurdles, long jump and triple jump, if a favourable wind exceeds that limit, the mark is still recorded as a competition result but cannot be ratified as a record. The threshold is not bureaucratic ritual. It is a way of saying that part of the number does not belong to the athlete.

Over 100m, a strong tailwind can be worth several tenths of a second. To outsiders, several tenths is a small number. To people inside the sport, it is the distance between a heat and a final, between a ticket to a major meet and a blank season.

Altitude produces a similar effect, more subtle because no machine raises an alarm. At around 2,240m, the altitude of Mexico City, the air is thinner, drag is lower, and every sprint and jump benefits. On 18 October 2026, at the Mexico City Olympics, Bob Beamon long jumped 8.90m. It was a legal mark, not wind-illegal, and it stood for 23 years. It was not a neutral measurement of human ability. It was a measurement of human ability multiplied by an unusual atmosphere.

The professional handling is straightforward, but only in the sense that it demands discipline. For every mark, I record the wind reading for each attempt, the venue's altitude, the temperature, and I compare against the same athlete in ordinary conditions. When an adjustment has to be estimated, I present it as an estimate, never as fact.

From a long jump result at a grassroots meet, an expert can say very little if the photograph contains no wind gauge. A mark may be written into the book, but it is not yet eligible for comparison. That is the sentence I write most often, and the one most often misread.

When the sole becomes a variable

On 31 January 2026, World Athletics announced its competition shoe rules, effective from 30 April 2026: a maximum stack height of 40mm for road events, no more than one rigid plate or blade, and a requirement that the shoe be made available to the general public for a period before elite competition use. Those thresholds and deadlines were later adjusted again. They were not written to ban technology. They were written to set a marker for measurement.

From Tokyo 2026 to Paris 2026, a wave of stiff-plated, high-rebound foam shoes produced a run of national records in middle-distance events. The public's first reaction was to suspect the shoe. The analyst's first reaction should be to read the shape of the improvement.

Three scenarios recur. If the entire race gets uniformly faster, from the first 200m to the last, the prime suspect is equipment or a fast track. If the final lap gets faster while the early laps hold steady, the suspect is fitness or tactics. If speed looks unchanged but the finishing time improves, check whether the course was properly measured. None of these scenarios needs a laboratory. They need a sheet of paper and patience.

I hold a professional bias against the "shoe versus athlete" argument. It puts the emphasis in the wrong place. Every era has its equipment, from synthetic surfaces to special rubber tracks. The real issue is distribution. The equipment dividend does not erase the athlete; it redistributes the advantage to those who can access it. In a thin-budget market like Vietnam, the gap between the athlete who can buy the newest model and the one who must reuse an old pair is wider than the gap between two training programmes.

For the athletes I follow, I keep a small ledger: shoe model, spike type, track, surface, weather. It sounds trivial. But when someone asks why a mark jumped in a given season, that ledger answers faster than any speculation.

It is also why I distrust composite indices built from a handful of samples. A composite index can hide an athlete's real role inside the competition structure. It looks precise, and because it looks precise, it is believed too quickly.

When the sample size is one

A single mark is not a level. A level is a distribution: mean, spread, conditions, opponents, tactics. To speak about a level, I need at least five competitions within about twelve months under comparable conditions, and only then do I look at the spread.

But "small sample, therefore suspect" is a lazy conclusion. Kelvin Kiptum is this decade's clearest counter-example. He ran his first marathon in Valencia on 4 December 2026 in 2:01:53. He ran his second in London in 2026 in 2:01:25. He ran his third in Chicago on 8 October 2026 in 2:00:35, a world record. Three races, a spread of roughly 78 seconds across 42.195km, each in a different race, against different rivals, in different conditions. That is not three lucky days. That is a stable level expressed in three data points. Kiptum died on 11 February 2026, and the hardest professional loss is that we will never know what his true spread looked like.

A small sample is only suspicious when the variance is large. When the variance is small and the conditions vary, a small sample is strong evidence, not weak evidence.

In the opposite direction, an athlete with a flat ten-year career who suddenly runs one very fast mark in perfect conditions is a small sample with large variance. That is noise. Calling noise talent is the most common professional error in sports journalism, because noise is easier to write than a distribution.

For domestic athletics, this rule has a direct consequence. A national title is one data point. It is not a career. A writer needs the other four points before using the word "dominant".

When a number has no paperwork

Every season I receive dozens of "training marks": a 1500m time trial in camp, a 400m segment in practice, a hand-timed 100m. These numbers travel faster than any official result, because they carry expectation.

On 12 October 2026, in Vienna, Eliud Kipchoge ran a marathon inside two hours, in 1:59:40. It was an extraordinary performance and, at the same time, a mark never ratified as a world record: a staged exhibition with rotating pacemakers, a laser-pacing car and a controlled environment. The lesson is not in the number. It is in how the event was labelled: clearly, from the start, as an exhibition. That is why it did not damage the measurement system.

A ratifiable mark requires a chain of conditions: certified timing, a measured course, a wind gauge, doping control, officials, and a governing body with authority to approve. Without that chain, you have a story. Stories have value. They must be called by their right name.

A mark without paperwork is a story, not data. The problem with training numbers is not that they are shared, but that they enter the expectation market without a label. A young athlete will eventually have to live with a number they never produced in public.

When families or coaches send me a training mark, I ask three questions: who timed it, on what surface, and what was the wind. Most answers arrive late, or not at all. That silence is itself a data point.

When the splits are missing

A finishing time is the end point of a curve. Splits are the shape of that curve. Two athletes can finish in the same time while owning completely different structures of ability.

The reading is basic, and I want to state it plainly in one paragraph without metaphor. Over 800m, take the first 400m split and subtract half the finishing time. If the difference is negative, the athlete ran the first half faster than their own average, meaning they spent reserves early. Over 1500m, compare the first 400m with the last 400m. If the last 400m is faster than the first, the athlete had room left and tends to finish well. If the last 400m is clearly slower, that signals a speed-endurance limit, however attractive the finishing time looks. Those three calculations need no software. They need a second stopwatch in the hand of someone sitting in the stands.

In 2026, when I was sixteen, I wrote my first piece about a 1500m final at the national youth championships. The athlete I chose to analyse conserved energy over the first 800m, sitting only sixth, then kicked over the final 300m to win. To verify the stride-by-stride structure I rewatched twenty-four recordings. The male editor told me plainly that girls do not understand tactics. They said girls don't understand tactics, so I wrote until they had to read it again. The piece reached 12,000 reads, six times the other articles in the same section, and I learned something unrelated to gender: readers do not need a chart, they need to see the shape of the race.

In Vietnam, split data is almost never published. The consequence is very concrete. A domestic coach knows their athlete's split structure in detail, but the file sent abroad contains one number. A selector elsewhere reads that number and has no way of knowing whether it belongs to a late-race sprinter or an even-pace distributor. That is a technical disadvantage, not a psychological one.

The good news is that closing this gap costs almost nothing. Two stopwatches and a phone filming from a fixed angle are enough to reconstruct the splits of nearly every race at a domestic meet. The work is tedious, nobody hands out medals for it, and it produces what the official timing system does not.

There is another stretch of time that is full of data and almost never recorded: the period without competition. Training load, sleep quality, the feel of recovery, the date of return from injury. Two hundred and fourteen days without a competition taught me to listen for the pulse of persistence. Those days produce no marks, but they decide the next mark.

Whoever owns the data owns the story

The sports industry pays for certainty. Ranking systems, sponsors, selection panels and newsrooms all consume a number, not a distribution. Fast conclusions are therefore better paid than correct ones. That is a structural incentive, not anyone's personal moral failing.

My contrarian position is this: in a data-poor environment, the sentence "there is not enough data to conclude" is a competitive advantage. The market is selling certainty it does not own. Whoever sells calibrated uncertainty outlasts it, because over time readers can verify who was right.

There is another temptation worth guarding against: turning analysis into decorative data. Heat maps, composite indices, performance scores create a scientific feeling while erasing the real role of the person inside the system. A beautiful composite number can hide that an athlete ran for a team-mate, or was mis-positioned all season.

What I propose is not to stop comparing. It is to record more and conclude more slowly. A shortage of data is no reason to hold weaker opinions. It is a reason to build a different kind of opinion, one supported by a notebook.

A results board ends a race, but most of the story sits underneath it. Underneath sit wind speed, altitude, sole thickness, sample size, splits, rest days, and the number of times people stayed silent. Whoever owns those things owns the story. Whoever does not is simply reading the board out loud.

Who will be the record keeper

Vietnamese athletics does not lack talent. It lacks record keepers. One season with a few hundred marks logged alongside wind, shoes, splits and rest days would produce a domestic dataset nobody could ignore, including the people assessing Vietnamese athletes from far away. It requires no budget. It requires a habit.

If one generation of coaches, reporters and sports science students records this season to the same standard, in ten years we will no longer argue by feel about who is improving and who simply got lucky. Who will open the first notebook?