How Athletes Get Great
Just train for 10,000 hours, right? Not quite. In his new book, author David Epstein argues that top-shelf athletic performance may be a more complicated formula than we’ve recently come to believe.
Heading out the door? Read this article on the new Outside+ app available now on iOS devices for members! Download the app.
Long before he co-authored a damning investigative article on Lance Armstrong or introduced the phrase “deer antler spray” into the lexicon with his feature on Ray Lewis’s dealings with some hucksters, Sports Illustrated senior writer David Epstein had questions about the biology behind elite athletic performance.
In high school he wondered why the mini-diaspora of Jamaican runners that populated his team would blow away the competition despite, “some of these guys not showing up to practice that much,” he says. As a member of the track team at Columbia University (Epstein ran the 800m), more questions came. Epstein observed that despite starting the season slower than most runners, he responded better to training, even though he and his teammates trained stride-for-stride. Then, when he arrived at meets, he learned that the competition in distance events wasn’t merely from Kenyans, they were mostly from the same rural tribe, the Kalenjin.
His observations of nature’s influence on athleticism seemed to contradict the nurture-based 10,000-Hour rule, popularized by Malcolm Gladwell. Epstein saw the rule influence not only popular culture, but academia as well. He wanted to investigate the long-standing nature versus nurture debate among elite competitors, so he travelled the world, interviewing athletes, scientists, and coaches in search of some answers.
Outside recently caught up with him to talk about his new book, The Sports Gene: Inside the Science Of Extraordinary Athletic Performance ($26.95; Current Hardcover):
Do you find there’s a misconception about genetics in the general public?
The problem is that, when a gene comes out, the media will say, “Oh, you have the fat gene or the angry gene or the promiscuous gene.” First of all, sometimes the results are not repeated and they get blown out of the water later. Other times, they have a tiny effect. It’s like trying to say you have the whole puzzle when you have one of a thousand pieces, with some of those other pieces being not genes but environment. So I could see why people would say, “you have the brain damage gene, you’re going to have dementia.” It just tells you that you’re at increased risk.
Some genes are deterministic, the ones that makes us like other humans. Brain chemicals, our organs, 10 fingers—it takes a pretty major mutation to change those. Then there are others like for Huntington’s where if you have the gene, you have the disease, but most genes are predisposition and not destiny, and that nuance gets lost.
It seems like the problems come when we graft narrative onto a subject we don’t fully understand, but want to delude ourselves into thinking we do.
That definitely happens. We fit narrative to what we can see. We can’t see our genes. All we can see are the things we can come up with so we fit narratives whether we can see all the evidence or not. For my training partner at Columbia and me, I now know that I have genes that make me a higher responder to training. I could see that my teammate—who unlike me was pegged as naturally talented—would start the season in way better shape, but with training I would surpass him. People would tell me how tough I am. And he was told he had a lot of talent, but had he had psychological problems or something. No, we were doing the exact same training. But you fit a narrative to what you can see. There’s this twins study I found where two twins were separated at birth and found out about each other as adults and they were both obsessive about being neat and clean. One of them said in an interview that his adoptive mother was really neat, so he learned from her. The other said his adoptive mother was a slob and he never wanted to be like her. Ok, so maybe it’s actually that there’s a genetic inclination to behave that way and they fit it to a narrative that works that way, and it happens a lot in sports.
The popular narrative you found while researching the book was the 10,000-hour rule made famous by Malcolm Gladwell’s book Outliers. Was the idea that pervasive?
Totally. And it motivated me to do the book, actually, and this was before I even knew how I felt about the so-called rule. I didn’t know about the science, but I wanted to evaluate it. And it took a long time and a lot of self-doubt before I became confident that I was interpreting it correctly. It was popping up everywhere. When I would go to the American College of Sports Medicine conference, there wasn’t a day that went by that people weren’t using it in their talks.
I remember writing a story about brain trauma at Purdue and watching video of a hit LaDainian Tomlinson took. He fell on his head and these scientists said, “Wow, he should have broken his neck. That’s your 10,000 hours there. That guy has put in his time of taking hits and a normal person would have broken their neck.” And I thought, “Seriously, we’re going to use this everywhere?” I talked to Olympic scientists a lot and they didn’t seem to think about the science or have even read they underlying paper. They knew nothing about it. The study is based on the practice hours of 10 people who are already in a world-famous music academy, so they’re already prescreened. When you take a study and you already screen out most of the gene pool, it’s not a very representative sample.
And people would call it “Gladwell’s 10,000 hours” as if he had done the research for it. It started to bug me. People were using it just to mean that practice was important, that’s it. That’s not what the theory says. The researcher behind it, Anders Ericsson, has said that he thinks all people have the necessary genes to be elite performers. Just saying that practice is important is totally uncontroversial. From a scientific standpoint it’s useless. Scientists have to say how important it is, what else is important? I found it to be troubling from a scientific standpoint and the more I evaluated it, the more it seemed to unravel. And ultimately, Ericsson read Outliers and said Gladwell misconstrued his work. His words, not mine.
How did Gladwell misconstrue it?
Aside from not having copied the numbers from the actual paper correctly for his book? He says that there is a perfect correspondence between practice and the level of expertise a person attains. And you can’t tell that from the paper. The 10,000 hours is an average of differences. You could have two people in any endeavor and one person took 0 hours and another took 20,000 hours, which is something like what happened with two high jumpers I discuss in the book. One guy put in 20,000 and one put in 0, so there’s your average of 10,000 hours, but that tells you nothing about an individual.
Now, Gladwell doesn’t say there’s no such thing as genetic talent. I think other writers are stricter than him. [Matthew Syed’s] Bounce is a book that minimizes talent. Gladwell does say elite performers are more talented. One of the things that Ericsson criticizes Gladwell about is to say that 10,000 hours is some kind of rule. The paper just says that these performers by the age of 20, these performers have accumulated 10,000 hours but there’s no where that says it’s a magical number where that’s when they become elite or anything like that. These people, by the time they go into their professional careers, have way more than that. That’s just where they were when they’re 20 as an average, not even to mention their individual differences.
Tracking chess masters helped you dismantle the rule, because it wasn’t so restrictive in its sample size, right?
Even with chess masters you’re talking about a population that’s pretty trained. So you’re still eliminating a large swath of humanity before you even start. Researchers found it takes 11,053 hours on average to achieve international master status. But the range there is what’s important. One guy takes 3,000 hours to become a master and another takes 25,000 and he’s still not there. So you can average those and come up with some rule, but it doesn’t tell you anything. You can always average individual differences and come up with some sort of a rule. Gladwell does leave more room for elite talent. He says, “Michael Jordan has more talent than me, but he also put in the work.” And that’s uncontroversial, that someone is talented and put in work. Beyond the chess players, genetics is continually finding now that one person’s hour of practice isn’t as good as the next person’s hour. Talent isn’t something preceding you trying something, but your biological setup that allows you to benefit more than the next guy.
That’s one of the most fascinating and unexpected parts of the book, where you discuss the Heritage study’s findings on trainability. Explain its implications.
That’s the most famous exercise-genetics study ever done. It’s the collaboration of five colleges in the U.S. and Canada. They took sedentary, two-generation families, which didn’t have a training history, and put them through stationary-bike exercise plans that were totally controlled. Families had to go into the lab and exercise over five months. The goal was to see how people would improve, and they were split into four different university centers to do the training and every center saw the exact same pattern. About 15% of people improved their aerobic capacity very little or not at all. And 15% improved 50% or more doing identical training. Families tended to stick together in the improvement curve, so about half of any person’s improvement was determined by their parents. I remember the editorial that ran in the journal of applied physiology “some people’s alphabet soup—meaning their DNA—didn’t spell ‘runner.’” One person training the exact same as another person can have completely different outcomes.
What does that mean for the athlete who has plateaued?
No cookie-cutter training plan is ever going to work. I’m a great example. Before my senior year of high school, I got up to 85 miles per week of training, which isn’t a lot for a pro, but was a lot for someone my age. When I came to college, I really got interested in physiology and took a scientific approach to my training. I found I was better at cross-country by training 35 miles per week with hill intervals instead of doing 85 miles per week. People need to pay attention to their training plans, because if something is not working for you as well as the next guy, it may be your biology, so you should try another plan. If you’re not taking a trial-and-error approach to training where you’re measuring something your time, you’re way less likely to find a plan that works for you. The cookie cutter approach to training is purely a facet of having a large group of people to train. If you’re writing a training book, then you have to be more broad. In the book, there’s a Danish scientist who biopsies his athletes and he’s found guys with huge fast twitch muscles and he tells them, “You’re working out too much because you’re causing your fast twitch muscles to take on the properties of more endurance muscle fibers.”
That can be tough for coaches, especially of big teams, because making everyone train the same way is seen as being fair to all the players.
I think my high-school coach noticed this about me, because when I graduated he said, “Get them to train you like a sprinter when you go to college.” And I did, and I got better at every distance by doing short training. I think part of the genius of Usain Bolt is that if you read his biography, 9.58, he talks about how lazy he is and he likes his coach because his coach realizes he won’t show up for practice some days. Who knows what his proportion of fast-twitch muscle fibers, but probably it’s huge. Those guys get hurt if they train too much, or they convert their super-fast-twitch muscle fibers into normal fast twitch. They take on the properties of endurance muscle fibers. Bolt will ramp up to peak when he needs too. For some guys, less training is the best medicine.
It’s like the anti-10,000-Hour Rule—working out too much can negatively affect some athletes?
It’s great that the 10,000-Hour Rule emphasizes the importance of practice because some people underestimate what practice can do. I think I can take anyone with two working legs and in six months can get them to run a marathon. I think most people don’t believe they can do that, but they can. It’s cool that they emphasize how much practice can do, but kids are getting burned out and injured from overtraining. It’s especially true for guys that have a lot of fast twitch, who are shown to be more prone to getting hurt. Usain Bolt has figured it out. He trusts himself to take time off.
A lack of nuance can really make it difficult to talk about genetics, especially with race.
Writing about race almost scared me out of writing the book altogether, which is why I wrote a section that had nothing to do sports on whether race had a genetic meaning in the first place. I was hoping that would be a non-hysterical way to start the discussion of genetics. But I wasn’t alone. Some scientists told me that they had data on physical differences in some ethnic traits that they were not going to publish. They were worried about reinforcing stereotypes or that people would take this to mean they’d also be somehow implying that there are innate intellectual differences between ethnicities; never mind that their work had nothing to do with that. But that’s the fear. So there’s clearly a political aspect to science that people are angling for depending on what social message they want to convey, but that social message has no bearing on the truth. The best way to get the best outcome for all people is to figure out what ethnic differences are real and what are not. Once I heard scientists tell me—it wasn’t often, but it did happen—that they were holding back data, I decided I didn’t what to hold back with things that I found.
Black athletes have been so negatively impacted by pseudoscience stereotypes about what’s biologically innate to them that it’s easy to distrust any discussion of what is innate ability. Writers like William C. Rhoden of the New York Times want scrap the idea of athletic prowess—good or bad—being innate, chalking up differences in race and ethnicities as social constructs. You cite an example of that in the book.
Rhoden says that white cornerbacks are shuffled off to safety instead of playing corner because whites are stereotyped as slow. I didn’t know if he was right or wrong. The only way I thought I could evaluate was to look at combine times to see if anyone who had the speed to be a corner but were shuffled off to safety. What I found was that there weren’t safeties of any ethnicity that were running fast enough, most of the time, to be cornerbacks. There certainly are social constructs and bigotry, but people attack those ideas thinking that will negate the bigotry. I think those pseudoscience beliefs are the result, not the cause, of bigotry. People aren’t looking at innate differences and decide, “Well, I guess I’m going to be racist.” Patrick Cooper addresses that in his research in his book Black Superman, when he dismantles the incorrect idea that physical prowess and intellectual prowess are on some sort of teeter-totter. That was never even an idea until physical prowess became associated with African Americans in the 1930s. So, I understand why it’s important to be critical of those ideas about innate ability.
But you believe there are innate differences between ethnicities and that we need to be up front about them?
In medicine, this is a non-issue. There was a study this month that come out showing tuberculosis measurements should be tailored by ethnicity because people with African ancestries, their immune systems respond differently to treatment. So you monitor the disease differently. We know that people with African ancestry have lower hemoglobin levels so sometimes they get turned away inappropriately from blood donation because they get measured against European standards. It’s really important to acknowledge ethnic differences in those cases. What the problem is when you take generalities and you apply them to an individual. A stereotype is a way to evaluate someone indirectly. When you can have someone at the NFL Combine, it makes no sense to evaluate him indirectly with a stereotype when you can evaluate him directly and decide whether he’s good or not.
Michael Johnson believes that the slave trade bred exceptional black athletes, especially sprinters from Jamaica. Is that a theory where pseudoscience is rearing its head?
You would definitely want to see more work to be conclusive about it. Yannis Pitsiladis [researcher from the University of Glasgow] is the only guy killing himself to do that work to figure it out and there’s only so much of that that he’s gotten done. There have been theories on other traits beside athleticism that the “unnatural selection” of slavery selected for certain traits and some of those have really fallen by the wayside, even though they seemed intuitively right. The slave theory might make sense, but the science doesn’t support it right now. Maybe when genetic testing changes, we’ll see something different, but I want to highlight where the science is now. So far the genetic data is suggesting that while every man who has been in the Olympic 100 meter final is of sub-Saharan West African descent, they come from a variety of countries and variety ethnic groups. There isn’t some genetic monolith for sprinting in Jamaica or the Caribbean. I think we’d see a lot more really good sprinters from West Africa if there were some sports infrastructure in those countries.
You mention the Jamaican coach in the book that’s wary of sending sprinters to the US for college because of the risk of overtraining.
When I went to Champs, the national high-school track championships in Jamaica, and I talked to coaches about their training plans, their kids who are the equivalent of our freshmen and sophomores only practice two to three days a week. They take it easy on them until they become upper classmen. When I was a freshman, we started to lift weights right away. They don’t let them lift weights until they’re at least 16. The underclassmen were training way lighter than what you’d find at decent high school in the U.S. Their approach seems to be one of the keys to developing to sprinters on the island. That coach still thinks most sprinters should go to college in the US, because they’re not going to make money off of running, so they should get a scholarship and college degree out of it. But the guys at the top, they should stay on the island because they don’t over race them there.
But when you see a cluster of high-level athletes from one location, does that raise red flags that the phenomenon we’re witnessing is just the result of PEDs?
I’ve been involved in reporting on doping. You’d be naïve not to wonder about it and Jamaica’s testing is getting more rigid and we’re going to find out. But when I look at their high-school times, like what they do at Champs and the number of good runners they send to the U.S., who are then subject to the same kind of testing that the U.S. has—I don’t think PEDs are driving the phenomenon overall, but I’m quite sure there are athletes are doping.
Which brings us to a really unexpected theme to this book: It’s secretly an economics and social science book as well as a genetic science one. The 10,000-hours faction believes in nurture over nature. You show the intersection of genes, training, economic incentives, and cultural institutions that create athletes. Nature and nurture together.
Usain Bolt is a great example. He was 6’4” when he was 15 years old and blazing fast. He wanted to play soccer or cricket. What are the chances anyone lets him run track in the U.S.? To me, it’s zero. There’s no way he’s not playing basketball or football. Nowhere but Trinidad, the Bahamas, Barbados, and Jamaica would a guy that’s 6’4”, with blinding speed, be allowed to run track instead of something else. People have asked me, “Should we do genetic screening for the best athletes or at least some sort of measurements?” Yes, measuring kids and trying to fit them into the right sport for their body type absolutely works. That’s why you saw Australia and Great Britain up their medal haul with their talent search programs when they had their Olympics. However, when there’s a sport that’s most popular in an area, you don’t have to do that because you already have the natural sifting program. You don’t have to go hunt for the best football players in America because they’re already going to go play football and then we select them.
Sprinting in Jamaica is like our system here for college football. What are the chances that a really good high school football player will fall through the cracks and not go to college? Pretty small, because people are looking for them and they earn adoration and accolades from performing well. That’s the way it is for youth track in Jamaica. They have shady boosters and everything! I went to a warm up track at Champs and started to ask the coaches about recruiting and how it works and they kept telling me, “We’re not allowed to give refrigerators to their parents.” I’m like, “What?!?” Apparently there was a rash of bribing kids with fridges to get kids to come to their track high schools. In Kenya, there’s no joggers. There’s only people who are running for transportation, people who are absolutely killing themselves in training to be Olympians and pros and people who aren’t running at all. There are no opportunity costs.
There was a guy named Brian Sell, he was [a marathoner] on the US Olympic team in 2008. He put in a lot of work, he got really good, he made the team, and he was putting off dental school. He was putting off making a living to chase being a pro runner. In rural Kenya, where the Kalenjin are from, there is no opportunity costs for attempting to be a runner. You’re not putting off any other opportunity, so you might as well try, so you get this huge input in the talent funnel. All of them try to train like Olympians. They go down to the local dirt track and guys who have a gold medal or world championship are already there and they literally try to run right alongside them right away. Most of them fall by the wayside, but the ones who survive are world-beaters. It’s a pretty good talent system.
But it’s not all economics and social systems for Kenyans. I mean, can a person with cankles win the NYC marathon?
It’s interesting. Americans think that Kenyans are good runners. Kenyans aren’t good runners; the subset of people from the Kalenjin tribe are the amazing runners. They run to school and I think that primes them for training and serves as a talent selection mechanism. But millions of kids run to school all over Africa and in India, and most of the great runners come out of the Kalenjin, so that environment is not unique to the Kalenjin. The Kalenjin have this incredibly narrow build, with a very narrow pelvic girdle and long, thin limbs. That’s a result of having your ancestry in a hot, dry climate. The more surface area you have relative to volume, the more heat you unload through the surface. Also, the less weight you have further from your center of gravity, the easier it is to swing your legs. So your running economy is better. Oscar Pistorius’ running economy is as good as an elite marathoner, which is unheard of for a sprinter, because he has these artificially light lower limbs. The lighter the lower your limbs, the better pace you can go for a given amount of oxygen. Some of the cool studies that have confirmed this is they take runners and put 8 lbs on their waist and it increases how much oxygen they have to use a little bit when they run at a certain pace. But if they take that same 8 lbs and put it around each ankle, so it’s 4 lbs each ankle, it’s a 20% difference in the amount of energy they have to use to go the same pace. Weight at the end of your limb makes it hard to swing your legs, which makes your running economy much worse. So you want as long and as thin a leg as humanly possible. That’s the build Kalenjin have. A study showed that even untrained Kalenjin have better economy than untrained Danish kids. Some of the Danish kids had better aerobic capacity, meaning they’re in better shape, but they still had worse running economy because their lower legs are thicker, so if you have thick lower legs, you’re not winning the NYC Marathon, unfortunately.
DAMN YOU FOR CRUSHING MY DREAMS.
What you’re saying does give some credence to the minimalist shoe craze, right?
Dan Lieberman, a scientist who did some of the famous work on barefoot running featured in Born To Run has done some work where he will have people run in shoes and then have them run with an equivalent weight strapped to the top of their foot. They’re running barefoot, but they still have the weight of a shoe and their running economy is worse. So you definitely want to stay light down there.
What you explained about the Kalenjin, is part of a larger trend in sports spurred by economics, which you describe as the Big Bang of Body Types.
Athletes in sports where they have to be big have gotten even bigger and where they have to be small have gotten smaller. For instance female gymnasts have gone from an average height of 5’3” to 4’9”. Two things happened. For a while there Germans were leaders in athletic science and there was the idea that the average body type would be best for every athletic endeavor—Vitruvian man. That turned out to be woefully wrong. So once that science receded and scientists realized specialized body types were better. But also as the reward pyramid became really narrow. There used to be vibrant club systems that would support semi-pro and lower level pros, especially in Europe, because if you wanted to see sports, you had to go in person or participate. Then you get TV contracts and revenue sharing with athletes where they start to be able to make a huge living and so many more people want to be pro athletes. Now everyone has a ticket to the Super Bowl, basically, by watching on TV. You expand the consumer population enormously and make the rewards tremendous, but concentrated at the tiny pinnacle of performance. As those trends happened over the 20th century it caused people who wanted to be professional athletes to try, which put more people in the talent funnel and the most specialized and best bodies for those sports were coming out the bottom. General managers and scouts started searching farther and wider as the money became bigger. One of the places that showed up was as soon as you have revenue sharing with players in the NBA, the percent of the league that’s made up of 7-footers doubled almost over night. Everyone wants to be in the NBA and teams started going abroad to find guys with height when they ran out of guys at home.
That explains why the high jump record isn’t being challenged anymore. The economic incentive to attract a big talent pool has gone away.
For the most part no one gets close. Even in the 100-meters, you’ve got Bolt, basically. There are guys in the finals of the 100 of the Olympics who probably aren’t making that good of a living. Trindon Holliday, now an NFL player, he may have been the best young sprinter in America at one point. He beat Walter Dix, who medaled in the 100m in Beijing, in the US Championships in 2007, but declined to go to the World Championships because he didn’t want to miss any LSU preseason football. There’s a guy who could become America’s best sprinter, but doesn’t even want to miss a day of preseason football.
Are you afraid that genetic breakthroughs will be abused by athletes?
One of the interesting things was any researchers I’d talk to who had discovered a gene that had an impact on muscle growth, when they published their findings, they’d get deluged by weightlifters and athletes who volunteer themselves. So that possibility exists. There was a trial of a German track coach and in the trial it became clear he was trying to get his hands on a gene therapy to increase the body’s production of red blood cells.
There’s a guy who I write about in the book who has a mutation on his EPO receptor gene that causes him to overproduce red blood cells. Look at what EPO did to cycling. You think if they could tinker with their EPO receptor gene they wouldn’t do it? However, I think if I were an athlete who was hell-bent on cheating, current forms of doping like microdosing testosterone and re-infusing their own blood are so effective that I wouldn’t even bother with something that could kill me. But I think that time will come where things will be made safer and then will people pick the traits of their children and that’s a little scary because we haven’t had discussions about how to regulate and deal with that. I fear it’s going to come faster than we can think about it.
So, what’s the next big breakthrough?
One of the coolest things is genetically tailored diet and training. It won’t be perfect because we still don’t know what most genes do. But exercise genetics will potentially produce some of the most widely used and effective medicine. You may be able to tell people how they can train to get a certain health benefit instead of taking a drug, or maybe that they can’t get that benefit with training so they do need a medication. Also, injury predisposition genes are coming online right now. It would be nice to know more than just through straight trial and error what’s the best training for your and how you can avoid injury. For example, what if you have a predisposition to low collagen production, so now we know to keep you healthy we have -to strengthen your support muscles more than the next guy. Of course, with the concussion crisis, there’s clearly a gene that predisposes people to sustaining more damage from getting hit in the head than other people and I think it’s crazy that testing for it is not in wider use. I understand the fears of it going into wider use, but this isn’t a place where you should be hiding from any information. Even though it’s just statistical info, it’s not destiny to have a gene.