Why sporting records keep falling

Nobody has been born with better legs since 1920. Human biology hasn't changed measurably in a century — not the muscle fibres, not the heart, not the tendons. The gene pool of 2026 is the gene pool of 1926 with a few more people in it.

And yet the record book keeps moving.

Every single explanation for that has to come from outside the body: from equipment, from surfaces, from what athletes eat and how they train, from who's allowed to compete, from how carefully the clock is read, and — honestly — from drugs. Sort those out and the record book stops looking like a story about human potential and starts looking like a story about everything surrounding the human.

Some of the improvement is just better clocks

Start with the least glamorous cause, because it's larger than people expect.

Sprint times used to be taken by hand. A timekeeper watches for the smoke or flash of the gun, thumbs a stopwatch, and stops it when the runner's torso crosses the line. Human reaction time at the finish is quick; human reaction at the start isn't, because you're anticipating rather than responding. The net effect is that hand times are systematically fast, and hand-timed marks were rounded to a tenth of a second on top of that.

World Athletics has required fully automatic timing for sprint world records since 1977, and when it converts an old hand time for comparison it adds 0.24 seconds for the short sprints. That's an enormous margin in an event decided by hundredths. A chunk of what looks like a century of progress is really the sport learning to measure itself.

Swimming went through the same reckoning, and there's a specific moment. At the 1972 Games the men's 400 metre individual medley was separated by two thousandths of a second, which is less than the tolerance in the length of a swimming pool. Timing to thousandths was abandoned shortly afterwards, and swimming has settled on hundredths with shared golds when it's closer than that. The lesson was that beyond a certain precision you're no longer measuring the swimmer, you're measuring the building.

Photo finishes now run at thousands of frames a second. Starting blocks contain force sensors, and a reaction under 0.1 seconds is ruled a false start on the grounds that nobody can respond to sound that fast. None of this makes anyone quicker. It changes what gets recorded.

Surfaces and equipment produce step changes, not curves

When a record improves smoothly over decades, that's usually training. When a cluster of records in one event all fall inside two seasons, something got invented.

Athletics ran on cinders — packed ash — well into the 1960s. Cinders absorb energy, drain badly, and cut up under spikes so the last runners in a heat get a worse track than the first. The 1968 Games in Mexico City used a synthetic surface, and the sport never went back. Later work at Harvard by Thomas McMahon and Peter Greene, published at the end of the seventies, showed that a track tuned to a particular stiffness — springy enough to return energy, not so soft that it swallows it — could improve running speed by a couple of percent without the runner doing anything differently. Modern tracks are engineered against that principle.

Pole vault is the clearest equipment story in sport. Bamboo, then steel, then fibreglass at the start of the 1960s. A glass pole bends and unbends, storing and returning energy, and it lets a vaulter convert speed into height in a way a rigid pole never could. The record climbed in steps for the next few years. The event essentially became a different event.

Speed skating got its own version in the late 1990s with the hinged skate, which keeps the blade on the ice longer through the push by letting the heel lift away. World records across almost every distance fell within about two seasons of it becoming standard. That wasn't a generation of unusually gifted skaters.

Swimming provides the cleanest natural experiment anyone could ask for. Polyurethane bodysuits, which added buoyancy and reduced drag, were legal for roughly two seasons. Forty-three world records were set at a single world championships in 2009. The suits were banned from the start of 2010, and a number of those marks stood for a decade or more afterwards while swimmers in ordinary costumes tried to reach them. Same pools, same sport, different fabric.

Distance running has been through this recently with carbon-plate shoes: a stiff plate embedded in a thick, highly resilient foam midsole, arriving from around 2016 and now universal at the top level. Road records tumbled. World Athletics responded by capping sole thickness and requiring that shoes be available to buy for four months before competition — a rule about fairness of access rather than about the physics.

Cycling took the most drastic option. By the mid-nineties the hour record had become a competition between bicycle designers, so the governing body simply reset it in 1997, restoring a 1972 mark as the benchmark and requiring broadly traditional equipment. That's an admission written into the rulebook: some records had stopped measuring riders.

Conditions are half the result

Athletics accepts a following wind of up to 2.0 metres per second for sprints and horizontal jumps. Above that the mark doesn't count. It's a sharp cliff for a continuous variable, and a 1.9 metre wind is worth a great deal.

Altitude is stranger, because it's allowed. Thin air means less aerodynamic drag and slightly less gravity, both of which help anything explosive and short. The 1968 Games were held at over 2,200 metres and the sprint and jump results were extraordinary — a long jump of 8.90 metres that stood for 23 years, and sprint marks well ahead of what sea-level competition had produced. The same thin air is a serious handicap over any distance requiring sustained oxygen delivery, which is why the distance results that year were slow.

Altitude-assisted marks get a note in the record book and are otherwise fully valid, which is a defensible decision and an odd one.

Then there's the race environment itself. Pacemakers, drafting, lighting strips set into the track kerb that move at world-record pace so the field can see the line they're chasing. In 2019 a specially organised marathon exhibition produced a time under two hours using rotating teams of pacers in formation and a vehicle projecting the optimal line. It wasn't ratified as a record, correctly, because those conditions aren't available in a normal race. What it demonstrated is roughly how much the environment is worth: a couple of minutes over the distance.

Athletes are professionals now, and they weren't

Through most of the twentieth century, the rules of amateurism meant that a large share of the best athletes in the world had jobs.

Tennis only opened to professionals in 1968. Athletics moved to trust funds in the early eighties and dropped the amateur distinction altogether soon after. Rugby union held out until 1995. Before those changes, training was something you fitted around employment, and elite competitors were often training a fraction of the hours a modern professional does.

What full-time athletes brought with them was a support industry: coaches with a scientific literature behind them, physiotherapists, sports psychologists, biomechanists with force plates and cameras, and surgeons who can rebuild a joint or a ligament well enough for a return to the top level. The elbow reconstruction first performed on a pitcher in 1974 is now routine enough to have a nickname.

Nutrition changed just as much and it's easy to underrate.

Scandinavian researchers in the 1960s, using a needle biopsy to sample muscle directly, established the link between stored muscle glycogen and endurance capacity, which is where carbohydrate loading comes from. Later work showed that the gut can absorb far more carbohydrate per hour if you supply glucose and fructose together, because they use separate transporters — which is why endurance athletes now take in 90 grams an hour or more, where the old advice was a third of that. Caffeine, sodium bicarbonate and dietary nitrate all have reasonable evidence behind them. Hydration, heat management and recovery are all managed rather than left to chance.

More people are trying, and that's a mathematical effect

A record is an extreme value. It's the best result from a very large sample, so the size and composition of the sample matter more than the average ability within it.

The population of the world has roughly quadrupled in a century. More significantly, the fraction of that population with any realistic route into elite sport has grown far faster — through desegregation, through the end of amateurism, through women's competition expanding from a handful of events to near parity, and through the sport becoming genuinely global rather than confined to a few wealthy countries.

Women's records improved fastest during the decades when women's participation was expanding fastest. That's a participation effect, not a biological one, and it's the single clearest demonstration that talent pools drive record books.

The catch is that extreme values scale badly. Doubling the number of people who try an event doesn't double the best performance; it shifts the extreme a little. That's why gains get harder and harder even as participation keeps rising, and it's part of why so many records now sit untouched for decades.

The part nobody enjoys writing

Doping is in the record book and pretending otherwise would be dishonest.

Anabolic steroids spread through strength and power events from the 1950s and 1960s onwards, at a time when there was no testing at all. Drug testing began at the 1968 Games and initially covered stimulants; steroid testing arrived in 1976. Blood doping — reinfusing your own red cells — was banned in 1986. Synthetic erythropoietin became available at the end of the 1980s and there was no direct test for it until 2000. The athlete biological passport, which looks for suspicious changes in an individual's own blood values over time rather than for a specific substance, only came in around 2008.

So for long stretches, in several sports, the deterrent was close to nothing.

Some of this is documented rather than alleged. The East German state programme, run under a numbered state plan, was reconstructed in detail from surviving archives after 1990 and led to criminal convictions of officials and doctors in German courts during the 1990s. That's a matter of court record, not rumour.

The effect on the record book is visible in the shape of the data. A number of marks set in the 1980s, particularly in women's track and in the throws, have stood for around four decades in a period when almost every other record in every other sport has moved. In 2017 European Athletics formally proposed resetting records that couldn't be verified under modern anti-doping standards, including the requirement that samples be stored for re-testing. The proposal wasn't adopted, largely because it would have punished athletes who'd done nothing wrong.

That's the sensible place to leave it. Individual athletes who were never sanctioned deserve no accusation here, and this article isn't making one. The structural point stands on its own: for a long time the testing was weak, some programmes were organised by states, and any record book covering that period carries the consequences.

Records are only comparable while the rules hold still

A world record measures human performance under a specific rulebook, and the rulebook changes.

The men's javelin was redesigned in 1986 by moving the centre of gravity forward, because throws were reaching the far end of stadiums and becoming dangerous. The new implement stalls and drops earlier. Every existing record was wiped and the event started again from zero. The women's javelin got the same treatment in 1999.

The hour record in cycling was reset for equipment reasons. Swimming's suits were banned. Shoe soles are capped. Every one of these is a decision that the sport is measuring the wrong thing, and each one puts a discontinuity in the data that makes cross-era comparison meaningless.

Which is the honest answer to "was X better than Y". You can't know. They competed under different rules, on different surfaces, with different equipment, different clocks and different pharmacology.

Where the records have stopped

Progress isn't uniform, and the pattern of where it has halted is informative.

• The men's 100 and 200 metre records have stood since 2009.
• The men's discus record dates from 1986 and the long jump from 1991.
• Several women's track marks from the 1980s remain untouched.
• Swimming sprints now improve by hundredths per decade rather than tenths.
• Road running, by contrast, has moved sharply in the last ten years — which is exactly where the equipment changed.

The rule of thumb that emerges is simple enough. Events where technique, surface and equipment have matured have gone flat. Events where something is still being invented keep moving. Nothing about the athletes themselves explains the difference.

There's also a warning here about prediction. A well-known analysis published in 2004 extrapolated the trend lines of men's and women's Olympic 100 metre times and found they would cross in the middle of this century. The mathematics was fine; the assumption that a trend line continues indefinitely wasn't. Both curves had been steepened by decades of expanding participation, and both flattened afterwards. Extrapolating a record is a good way to look foolish in twenty years.

So what would actually break the flat spots

If you wanted to guess where the next cluster of records comes from, you wouldn't look at physiology.

You'd look at whichever sport is about to have an equipment argument — a new material, a new shoe, a new blade, a new surface, a new way of measuring the finish. You'd look at events that have recently opened up to populations that weren't competing in them before. And you'd look at where the money has just arrived, because full-time training with a support team is still the single biggest difference between an athlete of 1970 and one of today.

The body is the fixed term in the equation. Everything else is negotiable, and it's the negotiable parts that keep moving the numbers.