The Data Gap on the Track: The Discipline of Answering 'Insufficient Evidence'

The wind gauge at Berlin's Olympic Stadium on the evening of 16 August 2026 stopped at +0.9 metres per second. That figure sat comfortably inside the legal limit, which is why Usain Bolt's 9.58 seconds was ratified as the 100m world record. Fifteen years later, in Chorzów on 25 August 2026, Armand Duplantis cleared 6.26 metres and did not need a wind gauge at all. Both record performances shared something the coverage rarely mentions: the paperwork attached to them. Wind, shoe model, doping control, referee reports, date of ratification. Strip that paperwork away and the number still looks beautiful, but it is no longer data. It is a story. This week a dossier landed on my desk with nine sections, four tables and thirty-seven cells. All thirty-seven cells carried the same phrase: insufficient information, cannot assess. No competition name, no athlete name, no event, no source. The sender wanted a deep conclusion. The only thing I could honestly deliver was the exact same sentence the dossier had already written. During a transfer window, the flow of athletics information thickens in a way that differs from football. Athletes change training groups, change coaches, change shoe sponsors. Each time that happens, a batch of numbers gets recycled: personal bests, season bests, training volume indices, a few timed runs on the practice track. Those numbers travel through news cycles with no measurement conditions attached. In football, people are used to querying xG, PPDA, pressing counts. In athletics the units are seconds and centimetres, so everything looks more transparent. The reality runs the other way. A shot can be adjusted by a model; a 9.84-second mark cannot. It is either right or wrong, and it is only right inside a narrow set of conditions. Precisely because the track leaves no room for soft interpretation, all interpretation gets pushed to the front and the back of the number. Five categories of data separate a real mark from a narrated mark. First, environmental conditions. World Athletics rules state that a tailwind above 2.0 metres per second disqualifies a mark from record consideration. Bolt's 9.58 survived because the reading was +0.9. A 9.79 with a +3.4 wind is just a fine afternoon. Altitude is the second variable in the same family: thin air reduces drag, and no conversion table turns it into genuine ability. Second, the equipment dividend. Since January 2026, World Athletics has capped road shoe stack height at 40mm with a single rigid plate, and track spikes at 25mm. The rule exists because the data showed part of the performance gain came from the foam midsole. When an athlete runs 0.3 seconds faster in a single season, I always separate how much belongs to the legs and how much belongs to the sole. Third, sample size. One run under ten seconds does not create a sub-ten-second athlete. The 100m carries large variance: reaction to the gun, wind, track temperature, lane quality. You need at least three runs in the same season, under three different conditions, before saying anything about a base level. Fourth, ratification status. A record does not become a record automatically. It passes through a confirmation process involving doping control, equipment re-measurement and cross-checking of officials' reports. That waiting period is grey space, and grey space is where rumour breeds best. A training mark with no wind gauge, no officials and no test sample is only a claim. Fifth, split data. This is the most ignored category and the most articulate one. A 100m run with only a final time is like a financial report with only the bottom-line profit. A 60m split shows where an athlete builds speed; a 150m split in a 200m shows how well that speed is held. Without splits, I cannot distinguish a sprinter who wins with top-end speed from one who wins with maintenance. Based on my experience tracking competitions, the difference in data quality sits not with the athletes but with the organisers. In Japan, the school and university ekiden system publishes kilometre splits almost by default, because the entire tactical calculation of the race depends on them. Japan's national 100m record first went under ten seconds in 2026, and every sub-ten run since has been logged with its wind reading in the federation database. In Vietnam, most domestic results stop at the final standings, rarely with splits or wind notes attached. This is a process comparison, not a comparison of people: the same athlete running in a meet that records splits leaves behind many times more information than in a meet that does not. The conventional reading treats a data gap as a technical defect, a hole to be filled with belief. I read it the other way. The gap is itself information. When a report publishes a mark and omits the wind, the beneficiary is not the audience. When an analysis showcases a season best and does not name the shoe, the beneficiary is the party selling the story. The absence is structured, and that structure points in one direction: wherever unfavourable data was left outside the frame. But I have to stop myself here. Occam's razor applies to the analyst too. Most missing data is not concealment; it is loose bookkeeping. An organiser with nobody assigned to splits does not publish splits; a grassroots meet without a wind gauge does not record wind. Inferring conspiracy from administrative laziness is an error, and a more dangerous one than believing a pretty number, because it creates the sensation of standing on the clever side of the game. The counter-intuitive point lies elsewhere. Numbers never lie; the liar is the person choosing how to read them. And the most common reading in this industry is assigning causation to correlation. An athlete switches training groups and runs faster: nobody checks whether he had a hamstring injury the previous season. A national team improves after importing a foreign coach: nobody subtracts the part that comes from the group simply ageing two years, landing exactly on the peak of the physiological curve. When everyone looks in one direction, I start examining the gap behind their backs. What people call a generational leap is usually just the surface paint of a deeper order: a cohort born at the right moment, a training cycle designed at the right moment, and a generation of equipment released at the right moment. Three curves overlapping produce a story, not a cause. Back to the dossier with thirty-seven empty cells. The answer insufficient evidence is not a refusal; it is a result. In an environment where everyone is compelled to have an opinion, the ability to stay silent in front of missing data is a competitive advantage, in the narrowest sense and in the longest-term sense alike. Three signals go on my board for the coming cycle. Ratification time: the gap between the day a mark is set and the day it is approved is an indicator of how complex the file was. Approved shoe lists: when a new model appears, wait for the following season before reading the results of the group using it. And the presence of splits at regional meets: wherever splits start being recorded, domestic analysis starts having raw material. Recovery is never a miracle; it is only something you already saw in the numbers three months earlier. The same holds for stagnation. Both were already in the data, waiting to be recorded, or waiting to be forgotten alongside the wind gauge.