Four Subtractions Before Trusting an Athletics Mark

Slow by one beat, and I saw the race begin at the twelfth frame.

London, the night of August 5, 2026. Usain Bolt was on the track, for the last time in a World Championships 100 metres. The board lit up: Justin Gatlin 9.92 seconds, Christian Coleman 9.94, Bolt 9.95. Most of the stadium remembers the moment Bolt sank to the ground past the finish line. I remember a different figure in the left-hand column of the results: reaction time. Gatlin 0.138 seconds, Bolt 0.183. A gap of 0.045 seconds, exactly the final margin between them.

I rewound the footage at quarter speed and stopped at the twelfth frame after the gun. By then the race was already decided. The remaining eighty-eight metres were simply Gatlin holding a gap he had won before Bolt left the blocks. The final mark does not lie, but it tells you nothing either. Since that night, every time an athletics result appears on social media, I force myself through four subtractions before I believe it.

Context

At this point in the year, athletics results arrive faster than records can be ratified. An athlete can run 9.9x at a national meet with a strong tailwind, the clip gets trimmed, shared tens of thousands of times, and becomes a phenomenon within twenty-four hours while the official file does not yet exist. The World Athletics legal wind limit is +2.0 metres per second for sprint and jump events. Above that threshold, the time is still recorded, but it is not considered for the record book. The wind column sits right beside the time column in every official result sheet, and most fans never look to the right.

The four subtractions I use require no special equipment. They are four questions: was there wind, at what altitude was the track, what shoe model was on the feet, and was the result electronically timed with split data? When all four answers hold up, I start talking about talent. When one is missing, the result is only data to be tracked.

Subtraction one: wind

A 100 metres run with a maximum tailwind can be about 0.10 seconds faster than in still air. Over 200 metres the margin is larger, because the wind acts for nearly twenty seconds instead of nearly ten. A 9.95 with +1.9 metres per second is not equivalent to a 9.95 in motionless air, and it is certainly not equivalent to 9.85. Bolt's 9.58 second record set in Berlin on August 16, 2026 was measured with a wind of +0.9 metres per second. Favourable, but comfortably legal, and part of the greatness of that record lies in the modest figure beside it.

Subtraction two: altitude

Mexico City sits at 2,240 metres above sea level. Thinner air means less drag, and on October 18, 2026, Bob Beamon long jumped 8.90 metres there. That record stood for twenty-three years. Analysts still argue over what share of the 8.90 belonged to Beamon's legs and what share to the Mexican atmosphere. Nobody can fully separate the two. All I can do is make a note: when a sprinter produces an extraordinary mark above 1,500 metres of altitude, wait for a repeat at sea level before ranking it.

Subtraction three: shoes and surface

Since 2026, World Athletics has capped road shoe stack height at 40 millimetres, allowed at most one rigid plate, and required that a model used in elite competition be made available for sale to all athletes. Those rules followed a string of questioned marathon records. Surfaces have changed too: the synthetic track in Tokyo in 2026 was measurably faster than earlier generations. Kelvin Kiptum ran 2:00:35 in Chicago on October 8, 2026 in a newest-generation shoe, and the result was ratified as a world record in February 2026. Naming the technological advantage correctly does not diminish the performance. It makes the comparison with the 1980s honest.

Subtraction four: split data and sample size

A result with reaction time, 30 metre and 60 metre splits, and a finish-line photo is a result you can interrogate. A result with only a final time is a photograph with no focal length. This is why I never judge an athlete on a single run. Sydney McLaughlin-Levrone took the women's 400 metres hurdles record down to 50.37 seconds in Paris on August 8, 2026, having already broken it several times before. A repeated series. Karsten Warholm ran 45.94 seconds in Tokyo on August 3, 2026, taking nearly eight tenths off the previous record. Both are real leaps, and the only way to tell them apart from a nice afternoon is repetition plus split data.

The contrarian angle

In 2026 I was wrong about Modrić. It remains the most honest analysis of my life. The lesson I drew is that scepticism has to be aimed correctly. The four subtractions above turn very easily into a cheap habit: dismiss anything large, find an excuse for any record. That reflex is laziness dressed in terminology.

The harder part is subtracting and then adding back. After removing wind, altitude, shoes and sample size, I still have to answer the remaining question: under standardised conditions, what did this athlete do that contemporaries did not? Skip that addition and I will miss the very moments that put me in this job.

When the stadium empties, I finally hear the numbers roll across every metre of grass. In 2026 I tracked the first 62 Bundesliga matches after the league returned and compared them with pre-pandemic data: the home win rate fell from 43 percent to 35 percent, and goals from counterattacks rose 12 percent. With no crowd, home advantage evaporated. Isolating a variable and measuring what remains is the same method I use on a 100 metre lane.

Takeaway

There will always be someone faster on a windy afternoon, and always someone slower on a perfect one. The reader's job is not to pick a side but to know what is being subtracted and what is being added back. Next time a 9.8x clip appears with no wind column on the results sheet, will you believe it straight away, or wait until the twelfth frame?