The Last Two Kilometres in Berlin: Assefa, 2:11:04 and the Price of a Race Without a Rival

The asphalt in Berlin on a late-September morning is colder than people expect. Temperatures hover around ten degrees, humidity is low, the sky is an even grey with no hard sunlight. For anyone who has ever stood at kilometre 35 of the Berlin Marathon, that is the weather organisers only dare dream of: near-perfect conditions for a fast race.

Tigst Assefa passed the 35-kilometre mark at exactly the moment the timekeepers began to believe in a record. She was roughly twenty seconds ahead of world-record schedule. Her pacemaking group was still around her. No rival could stay with her. Hold that rhythm and she would finish somewhere between 2:09:36 and 2:09:45.

In other words, for the final seven-plus kilometres of a major marathon, a woman was running on a trajectory that broke the two-hour-ten barrier.

Then at kilometre 40, the stride changed. The steps shortened, the full extension disappeared, the hips began to drift. By kilometre 41, the calf muscles had locked into hard knots. She kept running. She did not stop. But the pace fell to something a decent amateur could match for a few hundred metres.

She finished in 2:11:04.

On another morning, that number is a triumph. In Berlin, it was a world record dropped inside the final two kilometres. Immediately after the line, Assefa sat down on the asphalt, hands braced on the ground, head bowed, chest heaving. Medical staff ran over. She waved them off for a few seconds, then let them lift her up. She did not collapse. She only sat.

I have watched that footage many times, and what stays with me is not the clock. Numbers tell only half the story; the other half lives in legs trembling on the grass.

Two years ago I stood at kilometre 38 on this same course, the day Assefa ran 2:11:53 and set the women's world record. It was just as cold. I remember the strange sensation of a human passing in front of me at a speed the eye cannot process: she arrived, and two seconds later I could only see the back of her singlet. The crowd around me went silent for an instant, then broke into a roar.

This time I watched from Shanghai, on a screen set on the dining table with a live results page open in the next tab. No stadium noise, no smell of liniment in the warm-up area, no murmur of thousands of people holding their breath together. Only numbers jumping on a screen and a gap that kept widening behind them.

I once built a spreadsheet for a World Cup before I learned that the stands never appear in the formula.

That gap reminded me of 2026, when stadiums closed and I sat in Shanghai writing about sport without spectators. I told myself then that sport was still sport, that performance did not depend on cheering. I was badly wrong. An empty stadium does not lack a match – it lacks a soul borrowed from the roar. For a marathon that is doubly true. With nobody on the roadside and nobody beside you, an athlete has to generate all the energy alone, and that is a loan with no guarantor.

Berlin had crowds that morning. But out on the road, ahead of everyone, Assefa had no rival.

Berlin is a machine for manufacturing records, and that does not happen by accident.

The inner-city course sits at roughly thirty-four metres above sea level, low enough to be counted as sea level. It is flat enough that measurers place it among the fastest layouts in the World Marathon Majors system. Few turns, wide roads, and a finish line directly under the Brandenburg Gate – a symbolic detail so powerful that almost every photograph of this race is taken there.

Terrain is only half the story. The other half is culture.

Berlin is where pacemaking is organised at industrial scale. Organisers invite professional pacer groups, split by target time, running exact splits kilometre by kilometre and leaving the course when their job is done. For an athlete chasing a record, that means most of the race unfolds in a controlled environment: little wind, even rhythm, every variable pushed out of the equation.

The history of the race reads like a hall of fame for distance running. In 2026 Ronaldo da Costa broke the men's world record here. In 2026 Paul Tergat did the same. Haile Gebrselassie broke it twice, in 2026 and 2026. Patrick Makau, Wilson Kipsang and Dennis Kimetto passed through the same gate in succession. In 2026 Eliud Kipchoge ran 2:01:39; four years later he lowered his own record to 2:01:09. And in the women's race in 2026, Assefa herself ran 2:11:53 to set the world record.

One course, many generations, a single function: turning a limit into a measurable number.

So when Assefa planned her 2026 season, choosing Berlin carried a clear message. She already had Olympic silver from Paris in August 2026. She already had a world record once in her career. What was missing was a race faster than the 2:09:56 that Ruth Chepngetich ran in Chicago in October 2026 – the mark that took a woman under two hours ten for the first time in history.

That was the target. And that target shaped the entire organisation of the race.

One more thing about Assefa matters here. She came out of the 800 metres. That is a very different foundation from most Ethiopian marathoners, who usually climb up from 5,000 and 10,000 metres. An 800-metre athlete carries a higher speed reserve, better tolerance at maximal intensity, but also a different habit: dividing a race into short segments and deciding it at the end.

A 42.195-kilometre course does not forgive that habit.

She moved up to the marathon late and almost immediately reached the top tier: winning Berlin on her debut at the distance, then breaking the world record on her second attempt, then taking Olympic silver. This Berlin was her third victory here. A record like that is not luck. It is an athlete who has found the right formula and is trying to push it one step further.

That step measured exactly two kilometres.

Let us do the arithmetic nobody wants to do.

A world record of 2:09:56 equals 7,796 seconds over 42.195 kilometres, an average of 184.75 seconds per kilometre – roughly 3:04.8 per kilometre. At halfway, that pace demands passing 21.0975 kilometres in under 1:04:58. According to organisers' data, Assefa went through halfway fast enough to be inside record schedule.

By kilometre 35 she was about twenty seconds ahead of that schedule. Record pace at 35 kilometres is 1:47:46; she was there at roughly 1:47:26.

The rest is pure arithmetic. Hold the rhythm and she finishes around 2:09:36. In reality she finished after 7,864 seconds, that is 2:11:04. The final seven kilometres – precisely the 7.195 kilometres from 35 to the line – took 1,418 seconds, or 23:38, averaging 3:17.1 per kilometre.

Set side by side, those two figures give the real picture. At kilometre 35, Assefa was twenty seconds faster than the world record. At the finish line she was sixty-eight seconds slower than it. She surrendered eighty-eight seconds across the last seven kilometres.

Eighty-eight seconds.

To picture it: eighty-eight seconds is how long a marathon runner needs to cover four hundred metres at international racing speed. That is the entire distance between a world record and second place on the all-time list.

This race did not fail over the long haul – it failed inside the final two kilometres, precisely the stretch every marathon book calls the place where the truth shows up.

A reasonable estimate, based on the fact that her stride visibly changed only from kilometre 40 and broke down entirely in the last two kilometres, suggests most of the lost time was concentrated between 40 and 42.195. If the 35-to-40 segment cost only seven or eight seconds per kilometre against record pace – entirely normal once the fast pacers had gone – then roughly fifty of the eighty-eight lost seconds sit inside the last two kilometres.

This split structure has a name among analysts: the positive split. It signals a body that spent its budget before reaching the line, fundamentally different from the signature of a tactical error, which usually shows up as an early, even slowdown from kilometre 25 or 30 onward.

Assefa did not slow from kilometre 25. She accelerated. She led the record. She simply could not hold it for two more kilometres.

And here is what most headlines missed: 2:11:04 is faster than 2:11:53 – the very world record she once held. If the all-time list stands as it is, this result ranks second, behind Chepngetich alone. A morning described as a failure was, by historical standards, the second-fastest morning in the history of women's marathon running.

The distance between those two descriptions is exactly the distance between a number and a person.

Why would a body that had run 40 kilometres at record pace suddenly stop cooperating at kilometre 41?

There are two schools of explanation for cramping in endurance sport, and both have evidence.

The first, familiar to most amateur runners, blames dehydration and electrolyte imbalance. When the body loses sodium and magnesium through sweat, muscle cell membranes become unstable and fibres begin to contract spontaneously. The fix, in this view, is to drink more, add more salt, control core temperature more tightly.

The second, favoured by many sports physiologists over the past two decades, holds that cramping is primarily a neuromuscular phenomenon, arising from the central nervous system losing control of motor units inside the muscle. When a muscle is stimulated repeatedly at maximal intensity for a long time, receptors in the muscle send excessive signals to the spinal cord, and the cord responds by sustaining contraction. This is what is often called neuromuscular fatigue cramping.

The two theories are not mutually exclusive. In Assefa's case, both may be true.

Conditions in Berlin were relatively cold. In a race run in cool weather, the pure dehydration hypothesis loses some weight, because athletes sweat less and lose less sodium. The neuromuscular fatigue theory, by contrast, is unaffected by temperature. It depends on mechanical workload and contraction intensity.

A woman running 35 kilometres at an average of 3:04 per kilometre accumulates a vastly larger workload than someone covering the same distance at 3:30 per kilometre. The number of neuromuscular firings, the impact force per step, the stretch amplitude of each fibre – everything scales up.

At some point the nervous system can no longer distinguish between running fast and keeping the legs from breaking. It chooses the safe option: it locks the muscle.

There is another variable I consider the single most important in this case: the absence of a rival.

When you run beside someone, your brain has an external anchor to calibrate effort against. You see their back, you adjust to them, and you distribute energy against a target that physically exists in front of you. When you run alone against a watch, your brain has to construct that target itself and judge whether it is feasible. At kilometre 35, as the pacing group peeled away on schedule, Assefa's body lost its last external reference.

She still had the watch. But the watch could not run faster than her in the final two kilometres.

Behind Assefa, something else was happening that few reports treated with the importance it deserved.

The first five women across the line were all Ethiopian.

Hirpa finished second. Dida finished third. The entire remaining lead pack, down to fifth place, carried Ethiopian nationality. In the men's race, Adola won in 2:02:50, Gemechu Dida took second, and the only non-Ethiopian to break the top three was Geay of Tanzania.

Placed side by side, the two sets of results give a much clearer picture than the story of a single dropped record.

Ethiopia does not simply have one outstanding athlete. Ethiopia has a system, and that system operates in ways most other nations cannot copy.

Training centres cluster on the high plateau, where altitudes above two thousand metres turn every run into a hypoxia-adaptation session. Groups of several dozen athletes train, eat, race and push each other together. A young runner in Bekoji or Asella does not need a modern sports academy; they need a ten-kilometre dirt road and a group of people running faster than they do.

That model produces two effects.

The first is density. When a system produces hundreds of internationally qualified athletes, the probability of generating a world-record-calibre one is far higher than in a country with a few dozen.

The second is selection pressure. In Ethiopia, a place at a World Championships or the Olympics does not come from clearing a qualifying standard. It comes from topping a list of people who already clear it. With five athletes of top-five calibre in a single major marathon, that pressure is not small.

This places Assefa in a particular position. She is not Ethiopia's only athlete capable of running under 2:15. She is the fastest in a group of very fast people. That means her championship place is never automatic, and every major race carries a calculation about which events to spend energy on.

For an athlete who already has an Olympic medal and a world record, choosing Berlin – a race with no world title but the best possible record conditions – is a rational career strategy. But it also places the entire value of a season onto a single number: 2:09:56.

And when a season's value depends on one number, the final two kilometres can erase everything else.

The men's race confirmed something important about the course itself.

Adola finished in 2:02:50. Not a world record, but among the fastest performances ever recorded on a public marathon course. Gemechu Dida stayed close behind. Geay of Tanzania forced his way into third.

For analysts, the men's result works as a cross-check. When both races produce performances in the highest band of history on the same morning, the possibility that the women's race was a one-off can be set aside. The conditions were real and measurable.

It is also why Berlin holds a special position in the World Marathon Majors system. This race does not need a narrative. It needs a clock, a course, and people willing to run as fast as they possibly can.

One thing needs saying clearly, because it belongs to professional discipline.

The published data from this race is relatively sparse. We have two anchor points: halfway and 35 kilometres. We have the final result. But the per-five-kilometre rhythm – what analysts call the pace map – was not fully released.

That means any conclusion about exactly where Assefa began to fade carries a margin of error. The estimate that roughly fifty seconds were lost inside the last two kilometres is an inference drawn from observed movement patterns, not a subtraction from complete raw data.

I raise this not to dilute the conclusion but to say that anyone who reads numbers for a living has to be able to state what they do not know.

What we know with certainty is three facts: she was under record pace at kilometre 35, she lost more than a minute in the remaining stretch, and she finished with the second-fastest performance in history. Those three facts are enough to tell a story, provided we do not invent extra details to make it prettier.

There is one more part of that morning that numbers cannot reach.

The Brandenburg Gate sits at the end of Unter den Linden, and on the final grandstand the shouting does not sound like shouting at any other race. At big marathons, crowds usually chant to a rhythm, to names, in waves. In Berlin there is a particular kind of silence that appears when a race is running exactly on record pace. The crowd notices it before the organisers do. The noise shrinks, then stops, and only footsteps and the loudspeaker reading times remain.

Then comes the moment the clock shows the pace has gone. The noise returns, but different. It is no longer cheering for a record. It is cheering for a person trying not to fall.

For a reporter who has stood in finishing areas for years, this is the kind of sound I always note in my notebook: not the number, but the moment the grandstand changes its tone.

Where I grew up, marathon was once seen as a sport for people with something wrong in their heads. In Vietnam now, tens of thousands of people pay to run forty-two kilometres at five in the morning. In China, where I live, races appear so quickly that a person could enter one every weekend of an autumn. That movement and the story of Berlin are not on the same tier, but they share one thing: both revolve around a human being setting a limit and then trying to touch it.

The difference is distance from the limit. Amateur runners try to touch their own limit, usually somewhere around four or five hours, and for them touching it is a complete victory. Assefa tried to touch humanity's limit, and when she slipped, her performance became a failure in the newspapers.

That comparison is unfair, but I think it needs to be said, because it explains why we so often misread races like Berlin.

This is where I want to turn away from the usual telling.

After races like this, the press rushes in two directions. The first is tragedy: the record was within reach and it slipped away. The second is blame: she went out too fast, she made a mistake, she lacked nerve.

Both are wrong, and wrong for the same reason: we are judging a marathon by the standards of a short race.

In an 800 metres – Assefa's original event – the runner controls almost every variable. They know exactly how much energy remains, where their rivals are, and when to launch the final surge. A mistake over 800 metres is a mistake of judgement.

In a marathon at world-record pace, the athlete controls very little. They are borrowing energy from the future and repaying it with interest at kilometre 40. No training model predicts the breaking point precisely, because the breaking point depends on hundreds of small variables: road surface temperature, humidity, the number of missteps, one badly timed bottle, one shallow night of sleep three days earlier.

To say Assefa erred by running under record pace at kilometre 35 is to say she should have aimed at a smaller record. But her target was not a smaller record. The target was 2:09:56. And to touch that mark, she had to be under that pace at kilometre 35. There was no other route.

This is where data analysis and an understanding of human limits part company. An optimisation model would tell Assefa to run evenly and finish in 2:12:30, a safe result. But an athlete who has held the world record does not need a safe result. She needs a different one.

Tactics never die; they are merely misunderstood until someone is brave enough to try again.

There is another angle rarely explored, and it comes from where I sit.

Eight years in China and two decades around running courses in Vietnam give me a somewhat odd vantage point on how different sporting cultures organise fast races.

In East Asia, marathon is a mass movement. Hundreds of races appear each year, millions take part, and most social energy goes into finishing the distance rather than running it fast. In Vietnam, the number of people running a marathon under three hours remains small, and a national championship place does not generate the selection pressure seen in Ethiopia.

In East Africa, marathon is a profession. It is a means of changing economic circumstances, and so it is organised like an export industry of performance, with a supply chain running from villages to the major races.

The gap between these models is not innate talent. It is competitive density and how many people of equal standard an athlete must face each week.

But there is one point of overlap where I think both worlds are handling something poorly: the role of pacemakers.

In 2026, a half marathon in Beijing caused controversy when three Ethiopian runners ran ahead of a home athlete for most of the race, producing an image organisers had to apologise for. That incident shows something Berlin understood long ago: once pacers become a structural part of the race, the line between support and staging becomes fragile.

In Berlin everything is transparent. Pacers are announced, run in time groups, and leave the course when finished. For a marathon, this is how you keep a stopwatch mark meaningful.

But it also creates a psychological problem few discuss. When an athlete is placed inside such a perfect structure, every external variable has been removed, and the only remaining source of error is the athlete themselves. That is a mental trap. There is no rival to blame, no freak weather to cite, no collision to complain about. Only a clock counting down and a body asking to stop.

In Assefa's case the body won the negotiation. Perhaps it was right: had she forced the pace for two more kilometres on locking calves, the outcome might have been a long-term injury rather than a missed record.

A marathon runner who finishes with an intact body is still better off than one who trades a world record for six months of recovery.

There is one detail I saved for the end, because I want the story to close on it.

In the footage after the finish line, having sat down and been tormented by cramps, Assefa stood up again. She took a few more steps, even though her legs no longer obeyed. She did not lie back down on the road. She was not carried off on a stretcher. She walked out of the finish area herself.

No data table records that detail. No analytical model measures it. But it is the truest part of the whole morning.

At forty-two, I have learned that every athlete is a poet who never gets published. They write with their bodies, and the manuscript is erased the moment they stop. Assefa that day wrote a poem about touching a limit and failing to hold it. It was a lesser poem than the one she wrote two years earlier on the same course, but it was longer, more painful, and more honest about how human beings operate at the edge of their own capacity.

The world record of 2:09:56 still stands untouched. Someone will touch it. It may be Chepngetich, it may be Assefa, on another Berlin morning, with the final two kilometres better prepared. What is worth waiting for is not the number. What is worth waiting for is the moment when someone no longer has to sit down after the finish line.

And until then, the race continues. Runners in Hanoi, Shanghai and Addis Ababa still wake at four in the morning, still lace their shoes, still run toward a limit whose location they do not yet know. Berlin is simply the place where that limit briefly takes shape, in the final two kilometres, in front of all of us.