To run a fast marathon, you need to be good at two things: not spending energy and spending energy. For the first 20 miles or so, you need to go slower than your body wants. Conserve. Relax. Draft. And then, for the last six miles or so, as your body painfully shuts down—as your quads ache, your carbohydrate stores vanish, and your toenails rip—you need to go faster than your body wants. Endure. Hammer. Sprint. If you feel like you're jogging in the beginning and blazing at the end, then you're probably running evenly throughout the whole race.
The ideal way to run a marathon, in fact, is likely through what's called a negative split: running the second half slightly faster than the first, which turns out to be rather hard. Almost every experienced runner starts the race with a desire to run the second half faster, but only a very few actually do. The problem besets newcomers and it besets elites. Strava data shows that roughly one in 30 runners from the 2015 New York City Marathon ran negative splits. In the 2016 US men's Olympic marathon trials, only three of the 108 entrants ran the second half faster: the men who came in first, second, and third.
Running a negative split requires planning and discipline. Most people who succeed in doing so probably passed the marshmallow test as children. But it also really helps to have something surprisingly good happen during the race. Perhaps there's a tailwind. Or perhaps pain from an old injury didn't show up as expected. Or perhaps you had a pair of shoes that made you run faster than you'd run before.
This year, as WIRED reported thoroughly, Nike introduced a new shoe called the Zoom Vaporfly 4%. The shoe uses a special kind of very light foam, found in airplane insulation, and is designed to optimize the classic engineering trade-off between performance and weight. The essential motion in running is lifting your foot off the ground, so you want the shoes to be maximally light. But then your foot crashes down and you need to cushion the blow. Meanwhile, you also need to spring yourself down the road. And so Nike's engineers sandwiched in a small carbon-fiber plate, shaped like a spoon and curved in such a way that it theoretically propels you forward.
Nike sent me a pair of the shoes to test this summer. They felt strange and wobbly at first. My initial steps reminded me of putting on ice skates and then walking on a rubber mat to the rink. But once I started running, the strangeness disappeared, and I forgot I was wearing them—which is exactly what you want from a pair of shoes. They're unambiguously light, and they do feel cushioned.
According to the company, independent physiological tests show that the shoes make you 4 percent more efficient. Someone running 6-minute miles in Nike's previous fastest racers should, theoretically at least, expend the same amount of energy running miles 15 seconds faster in the new ones. Runner's World tested the shoes and scored the foam as having "the highest values ever recorded in our lab." Those results, if true, would be extraordinary. In the USATF 5K Championship, held this past Saturday, 22 men competed. The difference between first place and 20th was roughly 2 percent.
So far, Nike's results on the open road have been impressive. Eliud Kipchoge wore them to almost crack the 2-hour barrier in the Breaking Two spectacle this summer, and then again when he won the Berlin Marathon in September. Galen Rupp wore them when he won the Chicago Marathon in October. On Sunday in New York, Geoffrey Kamworor, who won the men's race, wore them. Shalane Flanagan, the first American woman to win in 40 years, had on a pair. Overall, three of the top four men and two of the top three women in Sunday's race wore the shoes.
But the data isn't unambiguous. In the Breaking Two test, Nike didn't just send Kipchoge onto their special track to trail behind perfect pacers and a special car in near-perfect weather. They sent out two other runners in the same shoes—both of whom underperformed, one of whom by quite a bit more than 4 percent. Furthermore, Nike could just be better at identifying, and paying, the kind of people who win marathons. Kipchoge would win his fair share of marathons wearing Uggs. And It should also be noted that the last five men's marathon records have been set in Adidas shoes.
The real question for Nike, which has invested an awful lot of money in these shoes, is whether they really work for regular runners—or at least regular runners willing to spend $250 for the Vaporfly 4%, or $150 for a slightly less advanced version, the Zoom Fly, that doesn't have the carbon-fiber plate. And testing shoes on the feet of regular runners in an open race is hard. All sorts of factors can slow you down or speed you up, from an upset stomach to a twisted ankle to slightly over- or under-doing it in the final week of training. With the New York City Marathon, the course is designed to be as psychologically and metabolically confusing as possible. Most people take a subway to a ferry to a bus in order to stand in the cold for 45 minutes at the start and then head straight up a steep bridge.
This year, though, at the start of the race, people in my corral were buzzing about the shoes that I and about a dozen other people had on. Friends wanted to hold them, bend them, and look for the carbon-fiber plate. One even reached to take one out of my hand as I scraped dog poop from the bottom with a stick. After the race, a friend suggested that wearing them was like a legal form of doping. And so I decided to try to design a test to see if they did actually help.
The methodology was defensible, if not entirely scientific. A friend of mine, David Greenberg, took photographs of roughly 150 runners as they passed through the Bronx late in the race. He had no idea that I would use these photos, but he stood at an angle that usually made the shoes visible, and all the pictures had been taken of people running reasonably fast. They're in a group of people who, I assume, all knew enough about marathoning to want negative splits. My research assistant and I eliminated the elites and just focused on the civilians. The fastest finished in around 2:30; the slowest in about 3:15. ( The data is all here.)
Of the 92 photos Greenberg took, 138 runners have both shoes and bibs visible. Of those, 21 are wearing a version of the Nike Zoom Vaporfly—either the $250 4% or the $150 Zoom Fly—and 117 are not.1 We looked up the data for each of these 138 runners, and, interestingly, the Vaporfly runners finished much better. Of that group, seven, or 33 percent, ran negative splits. Of the others, 17, or only 15 percent, ran negative splits. The average Vaporfly runner ran the second half of the marathon a minute and forty seconds slower than the first. But the average non-Vaporfly runner ran the second half five and a half minutes slower. Both groups of runners had faded, but the people in the new shoes had faded less.
These results couldn't remotely be published in a scientific paper. The sample size is too small, and the standard deviations are much too high. There also could be alternate explanations. It's possible that runners who pay exorbitant prices for awkwardly shaped shoes are the kind of people who plan their races well. The shoes might, for some reason, work particularly well on wet roads or in humid weather or on a course with punishing descents down a series of bridges. There could be a placebo effect. Still, the numbers are suggestive: Nike may in fact have designed something that limits the agony of a marathon's final 10K. Certainly, the theory behind the shoes makes more sense for a race through the five boroughs than the last time there was a boom in marathon technology: when, about eight years, ago everyone started running barefoot to be more like the Tarahumara.
Did the new shoes help me? It's impossible to know. I ran negative splits, and I closed faster than expected. It wasn't my quickest marathon, but I did do better than last year, which is good for someone getting definitely older and generally slower. All in all, I improved by about 2 percent. I'll wear the shoes again, and I suspect that next year, at the base of the Verrazano Bridge, the foam, and maybe versions of the spoon-like plate, will be in all sorts of elite racing shoes.
Still, one can never really know how much any one factor matters. There are too many things going on. When I got home from the race, I took off my snazzy sneakers and noticed something strange: a little red plastic disc from one of my kids' board games had somehow gotten wedged into the side of one of them. I hadn't noticed it while running or while putting the shoes on. But it probably added a few extra grams that a team of Nike engineers had spent weeks trying to remove. Maybe it slowed me; maybe by countering some unknown muscular imbalance it sped me up. I'll have to take it out and figure out, next year, if I can drop the remaining 2 percent.
Update: After publishing, George Wu, a researcher at the University of Chicago Booth School of Business who has written several academic studies of runners, wrote to say that he had run a regression analysis on that data posted to determine the relationship between second- and first-half times. It turns out, in fact, that the Vaporflys appear to decrease finishing times by 4.43 minutes, controlling for first half speed. That result, he writes, is statistically significant by normal standards. You can see his data here.
Update: 1 The piece has been updated to reflect the fact that one runner's shoes were misidentified in the original analysis.
WIRED's Graham Starr contributed to this report.
