Wind, Shoes and Track: The Three Variables That Decide What an Athletics Mark Is Worth

On the stands at Nagai Stadium in Osaka, I wrote a single line in my notebook: 10.21, wind +3.4 m/s. A few seats away, a young coach raised his phone and called it his athlete's personal best. He was not wrong about the number. But the World Athletics technical rules are explicit: results in sprint and jump events are only recognised when the tailwind does not exceed +2.0 m/s. That race was 1.4 m/s over the limit, so it does not exist in any official database. I filed it in a column of my own: zero value. Nine years ago I thought that column was pure paperwork. Now I understand it is the hardest part of the job.

I work in Osaka as an athletics data analyst for the Japanese market. The daily task is to read a results sheet, reconstruct the competitive context, and answer one question: how much is this mark actually worth under real conditions?

To do that I use a nine-dimension matrix: competition conditions and mark value; athlete condition and career curve; competition structure and qualification mechanisms; event landscape and national strength; technical rules and anti-doping; coaching systems; the risk map; media pressure; and the durability of a mark over time.

Last month an editor sent me a spreadsheet to analyse. It had all nine columns. Every data cell was empty: no competition name, no athlete name, no event, no wind reading, no date. I said it could not be analysed. He asked whether I could simply guess. I declined.

That is the hardest rule of the trade to keep: when the input is empty, the only honest conclusion is that there is not enough information. An empty sheet in athletics is not a neutral sheet. It is a trap. An impatient analyst will fill it with assumptions that sound reasonable — this athlete is peaking, that meet matters, this record is about to fall. Those assumptions then sit in the article as facts.

On that night in Russia in 2026, I watched data fall apart in front of me. I was 17, logging every phase of Japan's play and believing statistics could explain everything. My first real lesson did not come from a model. It came from a gap in the data.

The +2.0 m/s threshold is the line between a recognised record and a footnote that gets discarded. It applies only to the 100m, 200m, long jump and triple jump; events from 400m upwards are not wind-measured, even though weather still shapes the result. Above 1,000 m of altitude, thinner air cuts drag, which is why meets in Bogota and Mexico City have long been destinations for anyone hoping to change their life with a single race. An analysis that ignores those three variables will misread everything that follows.

Road-shoe stack height has been capped at 40 mm since January 2026, with a single rigid plate permitted inside. It was the first time the governing body publicly conceded that equipment can create a measurable advantage. Since then, every road record — including Kelvin Kiptum's 2:00:35 at the Chicago Marathon on 8 October 2026 — has to be read with one question attached: how much belongs to the legs, and how much to the foam. I do not have a precise answer. I only have an obligation to raise the question.

The year-by-year personal-best curve is the variable I consider the most important and the most neglected. Peak windows shift by event group: sprints sit around ages 24 to 29, middle and long distance around 26 to 31, and throws stretch to 28 to 33. A 21-year-old is on the upward slope. A 31-year-old sprinter is almost certainly on the far side of it. Two athletes can share a personal best and face two completely different futures.

When I plot that curve, I look for one very specific thing: an abnormal jump. If a 100m athlete has a history of gaining roughly 0.15 seconds a year and then gains 0.5 seconds in one season, that is a point to flag. Flag, not conclude. The check is arithmetic and implies no accusation. Over short sprints, it is the only tool I have.

In parallel, I always check competition structure. World Athletics offers two routes into a major championship: hitting the qualifying standard, or accumulating world ranking points. Each country may enter a maximum of three athletes per event, which makes fourth place at a national trial one of the cruellest outcomes in the sport. The United States selection model, where a single meet decides selection, pushes that risk to its maximum: a world champion can still stay home.

Here I have to state plainly what I consider the biggest blind spot in athletics analysis.

When a dataset contains nothing about doping, the correct analytical result is not assessed, not no risk. Silence in the data is not evidence of cleanliness; it is only silence. Athlete biological passports, ten-year sample storage and the medal reallocation process that follows re-tested samples exist for one reason: gaps in data are where cheating lives longest.

Empty stadiums, but the numbers are still full of noise. In 2026, when competition was suspended, I rebuilt Cerezo Osaka's pressing data from old footage and predicted a slide down the table because they had lost home advantage. They finished fourth, below my forecast. The error was mine: I treated no crowd as no pressure. Pressure does not disappear. It changes shape. I collect mistakes, sort them, and then I know where an athlete is heading.

One note for Vietnamese athletics. Applying Japanese track-and-field yardsticks directly to the Vietnamese system is a methodological error. A Vietnamese athlete races four to six major meets a year, clustered around the SEA Games and Asian championships, under different training, recovery and sports-medicine conditions. The same 0.3-second drop can signal overload in one system and normal loading in another. Nguyen Thi Oanh has won multiple SEA Games titles, and the way she distributes effort across distances is a story about load management, not about a single line of results.

Two more cautions. Correlation is not causation: road records appearing at the same time as carbon-plated shoes does not prove the shoes created the records. And one mark does not make a stable level. Su Bingtian's 9.83 seconds in the 100m semi-final at the Tokyo Olympics on 1 August 2026 is the Asian record, but it is one data point within his career, not the whole of it.

The 2026 season is opening, and for me it starts with a single move: rebuild the wind column, the shoe column and the personal-best curve before reading a single headline. Every probability carries a shock inside it — I just make sure it does not repeat. What I still cannot answer is this: when a mark is struck out because of wind, are we ignoring the truth, or protecting the meaning of truth?