The Epic Untold Story of Nike’s Two Hour Marathon Attempt

On the night of May 5, 2017, Eliud Kipchoge, the world’s best marathon runner, lay awake with his eyes open and his mind racing.

Under ordinary circumstances he is amiable and serene, with his furrowed, leonine features often lit with an ice-white smile. But that night, in his room in the Hotel de la Ville in Monza, Italy, he was more nervous than at any other time in his professional life. In the morning, on the 63rd anniversary of Roger Bannister’s historic sub-four-minute mile and at the culmination of Nike’s three-year-long Breaking2 project, he would attempt to do something nobody had ever come close to doing: run a marathon in less than two hours.

Kipchoge had asked Valentijn Trouw, one of his managers, to wake him at 2:45 am, exactly three hours before the start of the race. But when Trouw checked Kipchoge’s WhatsApp profile at 2:29, he found him online and awake. The pair decided to go to breakfast. In the hotel restaurant, Kipchoge betrayed no hint of tiredness as he greeted his two Breaking2 competitors—Zersenay Tadese, the current world-record holder in the half marathon, and Lelisa Desisa, a two-time winner of the Boston Marathon, neither of whom could sleep, either—or to the 30 pacemakers who had been recruited to guide these three contenders around the course. As Kipchoge ate his oatmeal, he smiled and shook hands with the battery of scientists and designers from Nike who circled the hotel, sleepless themselves.

At 4:15, Kipchoge was driven to the Autodromo Nazionale Monza, the Formula 1 racetrack whose 1 1/2-mile junior circuit had been chosen by Nike to host the two-hour attempt. It was starless and overcast: 53 degrees Fahrenheit and a little humid. In the home straightaway, where the race would end, giant screens showed gauzy highlight reels of the athletes in training, and the tarmac was lit in lurid pinks and blues as a crowd of 800 people waited for the show to begin.

On the back straightaway, where the race would start, there were no crowds. The atmosphere was tense and quiet. After changing into a red-orange singlet and arm sleeves, black half-tights, and the controversial racing shoes that Nike had engineered for him and his competitors, Kipchoge began a 30-minute warm-up: some easy jogging followed by a few “strides”—sprints to wake up the body. He spoke very little.

Kipchoge’s anxiety came not from the mere prospect of having to race, which he always welcomes, or from the expectations of Nike, which had spent millions of dollars applying the most advanced technology and sports science to get a marathon runner across the finish line in under two hours. Kipchoge was nervous because he simply didn’t know how his body would react to the stress of running so fast for so long. The fastest anyone, ever, had run a marathon was 2:02:57. Kipchoge wanted to run nearly three minutes faster, a 2.4 percent improvement, which might sound small but represents a giant leap in human performance. And when the body fails in the marathon, it can fail dramatically and painfully. Millions of people across the world were tuning in to watch livestreams of the event. Kipchoge, the marathon’s reigning Olympic champion, faced the real prospect of not just failure but mortification.

He was also aware of the skepticism, if not venom, that many obsessive running fans felt toward the Breaking2 project. Since Nike announced its effort to break the two-hour mark last December, many have called it a barely veiled marketing exercise for the shoe behemoth and a derogation of the sport’s spirit. Some decried Breaking2’s emphasis on record-breaking in a sport recently beset by a myriad of doping scandals, particularly from East Africa. A typical post on the influential LetsRun message board read simply, “What a stupid publicity stunt. I hate Nike even more after this.” Kipchoge believed all the complaints would dissolve if he could achieve what so many had thought impossible. But first, he had to do it.

A black Tesla Model S, with a digital clock mounted on its roof, was parked at the starting line in Monza. The electric car would lead the athletes around the track, driving at a constant 1:59:59 pace and showing them their split times on the display. The Tesla also shot out a green laser onto the ground, which would help the pacemakers know exactly how fast they needed to run to maintain the two-hour pace. The sight of the lead athletes warming up in their vividly colored uniforms among the black-clad pacers, and the green laser beams spilling across the tarmac, was strange and eerie—like a silent, illicit rave on a deserted freeway.

As the seconds counted down to the start of the race, Tadese bounced on his toes like a boxer and Desisa worried at his arm sleeves. Kipchoge, at 5'6" and 125 pounds, was strikingly composed. His eyes looked straight ahead. With 15 seconds to go, he dipped his torso and braced his body to run.

At 5:45 am, the starter’s air horn bleated and the three athletes shot off, behind the car and a group of six pacemakers, into the darkness.

The Nike Sports Research Laboratory is a secretive area within the company’s sprawling headquarters in Beaverton, Oregon, where scientists and designers work on new products, many of which never see the inside of a store. The whole purpose of this space, with its clandestine projects and NDA requirements, is to create interesting new things through repeated and creative failure.

Early in the summer of 2014, a handful of NSRL employees escaped from Beaverton to a resort in the town of Sisters to drum up new ideas. Matt Nurse, a tall, broad-shouldered former high school swimmer who leads the NSRL, was feeling bored and belligerent, and he told his colleagues that they weren’t taking enough risks. To inspire them, he played one of his favorite YouTube videos—a monologue by former Black Eyed Pea will.i.am: “Sometimes,” will.i.am says in the clip, “you have to put yourself in the situation to fuck up … Anybody else who’s not going out there and risking messing up … they never get any bigger.”

Throughout the previous year, shoe designers within the NSRL had been working on a project intended to help distance runners of all abilities improve their performance by up to 3 percent. To Nurse, this felt worthy but safe; he wanted his designers to devote themselves to something at which they could either succeed or fail definitively. Something like a sub-two-hour marathon.

It was a science fiction idea, long dismissed by aficionados of the sport as a waste of time and energy. In 1896, in the first Olympic marathon race, only one man broke three hours—the winner, Spiridon Louis, of Greece—and that race was less than 25 miles long, rather than the 26.2 that became standard in 1921. Over the past century, the world’s best time for the marathon dropped at an average rate of about five minutes per decade, driven by the professionalization of the sport as well as advances in shoe technology.

In 1991, when the world best for the marathon was 2:06:50, an American physician named Michael Joyner wrote a now famous paper, published in the Journal of Applied Physiology, estimating the best possible time for a marathon runner. By analyzing three main factors that limit a runner’s performance—VO₂ max (the maximum oxygen an athlete can consume while running), lactate threshold (the running speed above which lactic acid in the muscles accumulates prohibitively), and running economy (the efficiency with which a runner moves down the road)—Joyner argued that the perfect time for the perfect athlete in perfect conditions was 1:57:58. In other words, the sub-two was possible, but only just, and only in theory.

In the past 20 years, as the world record continued to inch downward, the debate over whether a sub-two might actually happen became more and more contentious. A parlor game arose among physiologists and statisticians, arguing when we might see the first sub-two: in 10 years, 25 years, 70 years, never. The game found its way to the retreat in Sisters, where Nurse tasked the NSRL team to imagine how they could make a two-hour marathon a reality. “We keep talking about the sub-two,” Nurse remembers saying. It was time to stop talking and actually do it.

Sandy Bodecker, a Nike employee for nearly four decades and the vice president of special projects, heard the call. He lobbied executives for funding and started a secret two-hour-marathon task force that was dubbed Project Able after one of the first monkeys to survive being sent into space. (In this context—a group of mostly white men aiming to propel an East African to a sub-two-hour marathon—that image has an unintentional but appalling resonance.) The name was changed to Breaking2 shortly before the project went public last December.

Bodecker and others on the senior innovation staff were adamant that the company not only design new shoes and performance clothing but a whole new kind of race, from the choice of venue to the selection of athletes to the training programs. Bodecker signaled his personal commitment to Breaking2 by having 1:59:59 tattooed across his wrist. The same number was also emblazoned, in digits 2 feet tall, at the entrance to the NSRL, even before the rest of the company was aware of the project’s existence.

These displays were a reminder not only of Nike’s mission but of its competition. In 2011 a young Scottish shoe designer at Adidas named Andy Barr had stood up in a meeting at the company’s German headquarters and declared that it should be thinking about a sub-two project. (Adidas brought out an Adizero Sub2 shoe this spring but has not announced when or if it will make an actual attempt.) Meanwhile, at roughly the time of Breaking2’s launch, in 2014, a physiologist from England’s University of Brighton, Yannis Pitsiladis, announced that he was seeking $30 million in sponsorship for a sub-two-hour marathon project of his own.

Tony Bignell, a jocular English designer who leads Nike’s footwear innovation team and works inside the NSRL, told me that Nike was never in any kind of “arms race” with Adidas or anyone else. Nurse, however, admitted to feeling some time pressure early in the project. “We believed that at some point within a few years, someone would get to a sub-two,” he says, and he wanted the team that succeeded to have a swoosh on its chest. To do that, Nike had to build its own perfect marathon from the ground up—starting, of course, with a shoe.

Though nobody at the company will admit it outright, it seems naive to think that selling sneakers was not a significant driver behind Breaking2. Geng Luo, a biomechanist and a lifelong sneakerhead from the outskirts of Beijing, was among the core group in the NSRL charged with designing a shoe specifically for Breaking2.

Starting out, Luo and his fellow NSRL designers had been thinking about creating a “track spike for the marathon”—an idea born of the prevailing wisdom that weight should be the primary concern in long-distance shoes. Weight costs runners. For every 100 grams on a sneaker, a runner’s energy expenditure increases by 1 percent. With this in mind, Luo and his team created a number of prototypes of stiff, light flats, shorn of any extraneous materials, based on a strong carbon-fiber plate. In one prototype, they even chopped the heel off the shoe, because many elite runners land on their midfoot or forefoot. The theory had some merit, but Nike’s test runners hated it. The ride was much too unforgiving.

Then, in early 2015, Luo says, they made the critical decision to sacrifice “lightweight for right weight”—to incorporate foam into their prototypes, despite the added weight, if it meant they could cushion and protect their runners’ feet over the length of a marathon. They began using a huge stack of what they called ZoomX, a superlight but responsive foam of the type used in aircraft insulation, underneath the foot. They then embedded a spoon-shaped carbon-fiber plate within the foam, which stiffened the shoe and rocked runners forward, as if they were running downhill.

The test runners embraced the new prototype, and the laboratory results were spectacular. In trials conducted by Rodger Kram, a biomechanist at the University of Colorado who consulted with Nike on the shoe, the meta­bolic cost to runners using the new footwear was found to be 4 percent less on average than for those using Nike’s previous best racing flat, the Streak 6. That means an athlete wearing the new footwear should, in theory, be able to run a marathon 4 percent faster (or, for an elite athlete with less room for improvement, 3.4 percent faster) than an athlete wearing a Streak 6.

The real-world results were promising, if slightly less astounding. Beginning in 2016, some of Nike’s athletes began wearing prototypes of the new shoes, the uppers disguised to make them look like existing Nike models. Not only did Kipchoge cruise to victory wearing a pair at the 2016 London Marathon, but all three medals in the men’s Olympic marathon, including Kipchoge’s gold, were won by athletes secretly wearing the shoe. As impressive as the results are, it’s worth noting that none of the runners made giant leaps of the type promised in the lab tests.

Nike finally unveiled the shoes, dubbed Vaporfly, in both elite and consumer versions in March of this year. The announcement caused controversy. Ross Tucker, an influential sports scientist and blogger in South Africa, called for the shoe to be prohibited, saying it gave runners who wore it an unfair advantage. “Any device inserted into the shoe, and which purports to add to energy return or elastic recoil or stiffness, should be banned,” he wrote. A ban was unlikely, though, because none of the basic components in the Vaporfly are new—all modern shoes use foam, and racing flats have incorporated plates in the past. (Haile Gebrselassie wore an Adidas Pro Plate when he broke the world record in 2007.) What was new about Nike’s shoe was how the plate sat within the foam. “The magic,” said Stefan Guest, an Englishman who finessed the design, “is in the geometries.”

Officials at the International Association of Athletics Federations, the running world’s governing body, have been mostly silent about the Vaporfly, suggesting that the shoes conform to their vague rules about fairness. Not that Nike executives would have minded a little controversy over a new product. “We always joked, ‘Wouldn’t it be great if [a shoe] were banned?’” Nurse says. “Just by opening an orange box you have an advantage? That’s the holy grail.”

Cait Oppermann

As the Tesla entered the bend at the top of the track at Monza for the first time, Kipchoge, Desisa, and Tadese were already settled into a steady, rapid rhythm, each of them about 12 meters behind the car and right behind the pacers.

Nike’s scientists had chosen the junior circuit of Monza’s historic racetrack because it lacked hills and corners, both of which hurt marathon times. It’s also looped, which meant a support team could regularly hand sports drinks to the Breaking2 runners to keep them hydrated and slow the depletion of glycogen in their muscles. The only drawback was the city’s slightly imperfect weather profile—Monza rarely gets much colder than 52 degrees Fahrenheit in early May, and studies suggest the optimum temperature for a world record is 50 degrees or colder.

The course at Monza also allowed the team to try a radical new approach to pacing. Most world records are set when runners spread their energy over the course of a whole marathon and move at an even, steady pace. But the most important function of pacers in Breaking2 wasn’t setting the speed—there was the Tesla, after all, which was driven with great skill at a near-constant 13.1 miles per hour—but rather blocking the wind. A number of studies have shown that the energy savings for a runner who is shielded from a headwind over the whole course of a fast marathon can be significant: enough to shave more than 60 seconds off the athlete’s final time.

This, however, is where science and the rules collide. In traditional marathons, the IAAF states that all pacers must start the race. Moreover, pacers typically drop out long before the race is completed, leaving the remaining competitors without any wind protection for the final 6 to 8 miles.

The Nike team began experimenting with having teams of pacers running in perfect formation, swapping in and out of the course, for the entire race. Breaking2’s lead physiologists, Brad Wilkins and Brett Kirby, who both have the laconic air of surfers, traveled to the University of New Hampshire’s Flow Physics Facility, the largest wind-tunnel testing site for runners and cyclists in the world.

In the middle of the Flow Physics lab, with twin 400-horsepower fans, Wilkins and Kirby placed an athlete on a treadmill and hooked him up to instruments measuring VO₂ max and heart rate. They then placed other runners in front of him in a simple wall formation, and measured the benefits. With that data, they then used computational flow dynamics software to test as many other formations as they could think of: two stacked walls of three pacers, three rows of two pacers, and many others¹.

Of all the potential formations, they found that a tight six-man arrowhead shape offered the most benefits. The savings the runners could make by hunkering behind the six-man arrowhead would be akin, Kirby said, to running the whole marathon downhill on a gradient of 2.5 percent. (As for the impact of the Tesla pace car and the large clock atop it, Nike ran tests both with CFD software and with test runners and determined that the car had no effect on air pressure 12 meters back, where the elite runners would be.)

In the end, a squad of 30 pacemakers was divided into six teams of three, with reserves ready in case of injury. On every circuit, the front three pacers would peel off and the rear three would take their place, while a new team of three pacers would fill in at the back row of the triangle. Many of the pacers were world-class athletes in their own right, mostly at shorter distances. (Speed, rather than endurance, was the most important prerequisite to pace Breaking2.)

Thirty minutes into the race, Desisa, wearing a white singlet, began to look profoundly uncomfortable. Even under optimal circumstances he has a ragged style, reminiscent of one journalist’s description of the great Czech runner of the 1950s, Emil Zátopek: “like a man wrestling an octopus on a conveyor belt.” But even so, it became obvious by the roll of Desisa’s shoulders that he was struggling. Near him, Tadese looked tough and busy—he is the Joe Frazier of running. He had twice run much faster than a two-hour-­marathon pace while breaking the world record for the half marathon; his problem has always been holding a fast pace for a whole marathon.

Kipchoge, meanwhile, ran with utter ease, his legs flicking along rhythmically and his torso as still as a rifle target.

To select the athletes who would compete in Breaking2, Wilkins and Kirby spent months analyzing the performances of professional runners, looking for key markers that reflected the necessary combination of speed and stamina. (Whether someone had run a sub-60-minute half marathon, for instance.) They were limited in their choices, because Nike does not have an exclusive hold over the world’s best distance athletes; indeed, most of the fastest marathoners of recent times, including the last four world-record holders, have been sponsored by Adidas. They also needed athletes who would forgo other marathons (and the attendant paychecks) to compete in Breaking2. Nike would not confirm what it paid its Breaking2 athletes, but one of Kipchoge’s managers told me that Nike offered two or three times the London Marathon rate. (London has offered star athletes appearance fees of around $250,000.) There were also time bonuses for the Breaking2 runners, including a rumored $1 million for a runner who broke two hours.

Wilkins and Kirby began bringing in athletes to test each runner’s “critical velocity”—the maximum speed they could maintain for a long period. (The very best marathoners can sustain up to 95 percent of their critical velocity for 26.2 miles.) Despite never having run a marathon in less than two hours and 10 minutes, Tadese had run the fastest-ever half marathon in history: 58:23. What’s more, his numbers from the tests, particularly his running economy, were extraordinary. Desisa too showed that despite not having the fastest of marathon personal records—a 2:04:45—he had a huge engine. Both his numbers and his relative youth told the team he had potential to improve his marathon time dramatically. Nike booked him for the project too.

When Kipchoge came in for testing in September 2016, his results were not as uniformly eye-popping as the others’ had been. (One reason may have been his utter loathing of treadmills. That trip to Beaverton was the first time he had used one, and he ran on the machine like a newborn foal.) But by then, he had won four major marathons and the Olympic gold medal, and what he lacked in raw physical potential he made up for with … something.

Kipchoge was born just outside the village of Kapsisiywa, in a verdant, peaceful spot in Nandi County, the youngest of four children. If you believe his passport, he was born in 1984, but both his family and fellow runners believe he is a few years older than that document suggests. (This discrepancy is common among Kenyan runners, most of whom are born in rural areas and lack birth certificates.)

Cait Oppermann

As a boy, Kipchoge ran 2 or 3 miles to school in his bare feet—often reversing and repeating the journey at lunchtime to eat at home. Even though he had won intramural running races at school, he never considered running as a career. About a year after he graduated from high school, he fell into conversation with a neighbor, Patrick Sang, a University of Texas graduate and Olympic silver medalist in the steeple­chase who had returned to his native Kenya to cultivate the next generation of runners. Sang wondered whether Kipchoge might have some aptitude for athletics. He gave Kipchoge running shoes and a training program, a combination of intense interval work, intermittent long, fast runs, and lots of easy running. Sang lives by the Kenyan running mantra: slowly by slowly. If an athlete has talent, he believes, it will blossom in its own time. Kipchoge’s flowered fast. Within three years of Sang’s guidance, Kipchoge was the 5,000-meter world champion.

Kipchoge spent the next few years winning a handful of medals at major championships at 3,000 and 5,000 meters, but he was never dominant in the way he is at the marathon. Since he started his marathon career at the 2013 Hamburg Marathon, he has won every race except his second marathon, in Berlin. Kipchoge’s two imperious wins in 2015 and 2016 at the London Marathon, which typically invites the strongest field, marked him as the best in the world, even before he won Olympic gold in Rio last year. Unlike such dominant athletes as Michael Phelps and Usain Bolt, Kipchoge has no real physiological point of difference from many other small, slight, super-fit Kenyans. His domination appears to result from his adamantine commitment, both in training and in races, and his self-belief. “It’s not your legs that run,” he explained to me one sunny morning at his training camp in the forest of Kaptagat, in Kenya. “It’s your heart and your mind.”

Nike booked Kipchoge for Breaking2—they would have been crazy not to—and with their three runners set, they assembled the squad for the first time in late November at what they called Camp 1². The three-day event in Beaverton was an opportunity for the three athletes to meet with Breaking2’s growing sports science team, which now included two more contractors: Phil Skiba, a burly, garrulous physician from Chicago whose data-driven coaching has led triathletes to world championships, and Andy Jones, a dapper professor from the University of Exeter in England who had formerly advised the women’s world record holder, Paula Radcliffe. In Beaverton, this team looked on as the athletes recorded new data on the treadmill. The runners wore masks to measure their oxygen uptake and were tested while running in climate-­controlled chambers.

These tests did not touch on performance-­enhancing drugs. Although all three athletes received frequent blood and urine tests because of their elite status, they did not undergo any additional testing because of Breaking2. This omission dismayed many in the international running community who felt that Breaking2 didn’t just need to be clean—it needed to be seen as clean. The optics should have been especially important to Nike, whose star running coach Alberto Salazar is at the center of a US Anti-Doping Agency investigation into the use of controlled supplements in his program. When I spoke to Travis Tygart, head of the USADA, he said that Nike could easily have asked the IAAF to institute a tougher antidoping regimen for the three athletes involved in the project. Nike declined to do so, for reasons the company never satisfactorily explained to me.

The three principals were sent home from Camp 1 with GPS watches and heart-rate monitors that allowed the scientists in Beaverton to track every run they did. In January, the scientists then traveled to Kenya, Ethiopia, and Spain, where Kipchoge, Desisa, and Tadese lived and where the Nike team could observe the training in person. Tadese, for one, was fast but had trouble with stamina. After the science team watched him run a 9 x 1,200-meter session in Madrid, Skiba said, “We saw a monster out there today!” The science team asked Tadese’s coach, a Spaniard named, unimprovably, Jerónimo Bravo, to introduce more runs in the “heavy zone,” just below his lactate threshold, as well as longer runs of between 20 and 25 miles.

Desisa, meanwhile, appeared to log extraordinary mileage in his training in Ethiopia, sometimes as many as 200 miles a week, but did very little fast running. The Nike team encouraged Desisa’s coach to shorten and intensify some of his workouts in an effort to increase his speed.

With Kipchoge, they found very little they wanted to change. Skiba churned Kipchoge’s numbers to find areas of potential weakness and thought that he had found one: that the Kenyan “tapered,” or slowed down his training, a little too late before a big race. Months later, with more data in Skiba’s algorithms, it turned out that Kipchoge’s taper time was more or less perfect. His training continued unchanged.

In March, two months before before the attempt, the three athletes reunited on a bright and windy day in Monza for a “rehearsal” half marathon. The idea of the rehearsal was that Nike would test some of the techniques it hoped to apply in the race itself. They recruited pacers to help them test the interchanging teams and asked the principal athletes to try to run an evenly paced 60-minute half marathon. (It was so casual, this request: 60 minutes dead would win almost every half marathon in the world.) Despite unusually high winds and some shambolic organization from the pacers, Kipchoge and Tadese both looked comfortable as they ran half marathons below 60 minutes. Desisa’s form, however, was worrisome—he finished in more than 62 minutes.

Lingering over the test event, and over the whole project, was a question the team never satisfactorily answered among themselves: How far are we willing to go to break two hours?

They could have run the race with a fan blowing the athletes, just as the sprinter Justin Gatlin ran a stunt 100 meters in Japan³. They could have run the race downhill all the way. There were any number of cheats Nike could have used. Matt Nurse, head of the NSRL, felt that Gatlin-esque stunts would be going too far. He was against anything “actively pushing or pulling” the athletes to their goal. And although the Nike sports science team had been working on methods to help the athletes run faster that breached the rules of the sport—such as having interchanging teams of pacemakers in front of them, or feeding them water bottles from moving bikes rather than stationary tables—the company felt those didn’t violate the spirit of the marathon as much as other potential transgressions.

This debate came to a head as Nike executives gathered at the March rehearsal. Some senior figures at the company, Nurse says, had become concerned that “certain parts of the media” had accused them of “cheating the sport,” and they were anxious about the negative publicity. And some wondered whether it would be better to scrap the two-hour attempt and simply go for an official world record, ratified by the IAAF, which would have precluded the use of the interchanging pacing groups. When word of these discussions reached Jos Hermens, who founded Global Sports, Kipchoge’s management company, he emailed Nike CEO Mark Parker to argue against watering down Breaking2.

“Nike should not be afraid,” said Hermens, paraphrasing his email. “Nike is meant to be … a rebel company! To make an unofficial record is fantastic! For the next five or 10 years everyone talks about Monza and Eliud and the shoes. You couldn’t have it any better.”

In the end, Nike listened to its athletes, and whatever anxiety the company was experiencing was overwhelmed by the enthusiasm of the runners themselves. Kipchoge’s explanation for why he wanted to attempt a sub-two-hour marathon would sound goofy were it not delivered so earnestly. He told me that he wanted to show ordinary people that they can overcome barriers in their own lives. It was a message he intended to carry not just to fans of the marathon, or to fans of sports in general, but to “8 billion”: every single person on the planet.

“After I break two hours in May,” he said, “the whole world will break.”

The night in Monza had given way to a milky dawn, and at around 6:35 am on race day, with 11 miles down and 50 minutes on the clock, Desisa could bear no more. As the pacers swapped in and out in the transition zone at the top of the home straightaway, he began to trail Kipchoge and Tadese. So much of distance running is psychological. It’s mentally exhausting to run a pace you think you cannot maintain. Once Desisa knew he could not hang on to the two-hour pace, his chance of success was over and he separated from the pack like he’d been thrown off the back of a speedboat.

By the end of the next lap, just before the halfway mark, Tadese too began to fall behind Kipchoge, albeit less dramatically than Desisa. Out of the three elite runners chosen by Nike to attempt the sub-two marathon, only one was going to make it to halfway under 60 minutes.

As Kipchoge entered the second half of the race with no discernible change in his rhythm, he looked increasingly assured. When Kipchoge is running fast and feeling good, there is a tell in his form: As his hands pass his chest, it looks as if he’s using his thumbs to brush lint from the lapels of a dinner jacket. Thirteen miles became 15 miles became 17 miles. Still, he was on schedule to break two hours, and still he was brushing lint.

With three laps and roughly 4.5 miles to go, Kipchoge had been alone for 40 minutes. Trouw and Kirby cycled alongside him, offering words of encouragement and information about his progress. For a while, the display on the pace car had failed, and the two cyclists had informed Kipchoge of his splits themselves. Meanwhile, the pacers had interchanged in the transfer zone at the top of the home straight seamlessly. He seemed almost at peace, his face relaxed and his breathing steady.

But now his split times began to drift, almost imperceptibly at first, by a second or so a mile. The sub-two was still on, but the prospect was on a knife edge. The Tesla car, driven mercilessly at the 1:59:59 pace, began to edge away from him. The turnover of Kipchoge’s legs appeared to slow, just a fraction. At times, he appeared to smile broadly—a conscious effort, he later said, to relax and work through the pain.

Cait Oppermann

Although Kipchoge had not worn any monitors during the race, Nike scientists knew roughly what was going on within his body at this moment: His core temperature had risen by at least a couple of degrees, the glycogen stores in his muscles were nearly depleted, and those same muscles were suffering mechanical damage from the repeated impact on the road.

With two laps to go, Kipchoge was a few seconds off the target pace. There were many—including his managers, Trouw and Hermens—who thought he could somehow still break two hours. In his track career, Kipchoge showed electrifying speed to close races. If he could summon the strength for two rapid final laps, he might yet reach 1:59:59. As he rounded the bend into the home straight for the penultimate time, Hermens danced beside Kipchoge and bellowed encouragement.

On the final lap, Kipchoge needed to close with an impossibly fast 4:17 mile. While simple math told you he could not do it, still there was belief around the track. Trouw, on the bicycle, shouted, “Come on Eliud, history!”

As he rounded the final bend, urged on by an emotional team of pacers, Kipchoge knew the sub-two was beyond him. He drove onward, breaking the tape in two hours and 25 seconds—more than two and a half minutes faster than the world record. Having walked a few paces beyond the finish line, he found some hidden store of energy to run to his coach, Patrick Sang, who embraced him like a son. Kipchoge then lay down on the tarmac for a few seconds, and a salty smile spread across his face.

By the terms of its mission, Breaking2 was a failure. Nike tried everything to help a runner break two hours, and they couldn’t do it—not quite. But nobody who knows the sport, and who was at Monza to witness the attempt, could have seen Kipchoge’s run in such terms. Kipchoge’s 2:00:25 was one of the most impressive displays of distance running in history.

What’s more, Breaking2 has proved that a sub-two-hour marathon is not only possible but within reach. It’s hard for anyone, even Nike, to isolate the factors that contributed most to Kipchoge’s final time. Many believe the innovation that made the biggest difference was the arrowhead formation of pacers and the resulting wind deflection. The shoes probably shaved off some time too; not 3 or 4 percent, but something. The most significant factor may have been Nike’s audacity in making a sub-two push in the first place. “I’m sure there are people who think we’re a big, greedy, corporate monster that only wants to sell shoes,” Bignell says. “But this really was about human potential. I hope people will be inspired.”

If there was a grander narrative to extract from Breaking2, it was not about the possibilities of science but rather its limits. Data can only get you so far. What good, ultimately, came from having all of Kipchoge’s data? Kipchoge didn’t change his approach to training, even if—as both he and his coach told me—they were gratified to have their methods validated by scientists. As for the information Nike gathered on Desisa and Tadese, which did lead to real changes in training for both men, those insights could and should have been made already, by a coach. (Nike will not release any of its Breaking2 data, presumably so it can use it to enhance the company’s understanding of the hows and whys of fast running.)

There may also be lessons about the limits of the human body. Kipchoge ran at a two-hour pace for nearly a whole marathon but could not summon the speed for the final laps, because his effort in the first 24 miles had cost him too much. Andy Jones believes that Kipchoge simply ran out of fuel. “That might have been the limiting factor,” Jones says. “If we could find some way to have more glycogen on the start line or to consume more carbohydrates during the event … he might sustain that speed a bit longer.”

The single biggest scientific misjudgment in Breaking2 might have been running the race in Italy, in May. The bodies of elite marathoners get extraordinarily hot. To maintain a steady state of exertion while running fast, it’s necessary to dissipate that heat. The hotter it is, the harder that is to do. Literature suggests that it was at least a couple of degrees too hot for the very fastest times on May 6 in Monza. (The Nike scientists disputed this with me after the race. They thought the temperature about perfect.) There was also another, less scientific problem with how the race was staged at Monza: For long stretches of the course, it was deserted. Kipchoge told me afterward that he “missed the crowd” he gets at normal marathons.

Kipchoge admits no disappointment at having missed his goal. He is now back at his training camp in the forest of Kaptagat, where he stays for six days of every week away from his wife and three young children, who live in a well-appointed modern house in the nearby town of Eldoret. He trains twice a day, once in the morning and once in the afternoon, along with 30 other athletes. Between sessions, the runners do household chores; when it’s his turn, Kipchoge, a millionaire in a country where the average income hovers round $1,300 a year, pitches in and cleans the “long-drop” toilets.

After the race, Kipchoge sent me a text message, which contained a series of questions about my long-standing interest in the two-hour marathon: “Do you feel your dreams have come true? Do you feel complete now?”

I didn’t have a good reply, but in the minutes after he crossed the finish line, Kipchoge had shown his own answer to the same questions. He embraced Sandy Bodecker, whose early enthusiasm had launched the project, and pointed to the 1:59:59 tattoo on his wrist. The gesture was gracious, but there was also a pinch of yearning in it, and it swelled the heart to see.

---

Ed Caesar (@edcaesar) is the author of Two Hours: The Quest to Run the Impossible Marathon.

This article appears in the July issue. Subscribe now.

¹Correction appended 3:15 PM ET 6/30/17: Updated to clarify the methods used to test Breaking2's pacer formations.

²Correction appended 3:15 PM ET 6/30/17: Updated to clarify the name of Nike's Camp 1 training event.

³Correction appended 2:45 PM ET 6/29/17: A previous version of this story misidentified the location of Justin Gatlin's 100-meter stunt run. It took place in Japan, not China.