Practice Doesn’t Make Perfect: Scott Barry Kaufman and David Epstein Reconsider the Science of “10,000 Hours” to Greatness

“There are actually some drawbacks to suggesting a one-size-fits-all model of development.”

Scott Barry Kaufman is the scientific director of the Imagination Institute and a researcher at the Positive Psychology Center at the University of Pennsylvania. He is the author of two books on intelligence and creativity, Ungifted and Wired to Create, and his essays have appeared in The Atlantic, Scientific American, Psychology Today, and more. He recently joined David Epstein, an investigative reporter for ProPublica and the New York Times bestselling author of The Sports Gene, for a Heleo Conversation about the complex science of expertise. They argue for why we need to move beyond just “deliberate practice” when we talk about achievement, consider why certain theories reach peak popularity, and discuss why imagination matters immensely to achievement.

Scott: I’d like to talk to you today about the complex interplay of nature and nurture and what the latest science shows. You’ve done some research looking within the sports domain, so I thought we’d start there and build up to other fields.

Within sports, there’s this model of deliberate practice that Anders Ericsson has been studying for many years. I think that Anders’ research program has been tremendous in showing us that there are things we can do deliberately to increase our skill acquisition, to become better at almost virtually anything. It also shows quite clearly that despite our genes, or despite our innate endowments, the environment has a huge impact on our efforts.

Nevertheless, there is accumulating research showing that you could still practice in exactly the right way for 10,000 hours, or as many hours as someone else, and yet there would still be variability in outcomes. Is that correct?

David: That’s correct. In fact, the more complex the skill, people typically get more different rather than more the same as they practice. In some domains—not in all, but in some—and especially at the highest level skills, practice actually pulls people apart rather than bringing them together. There was a recent review that looked at evidence that would test Ericsson’s model, and the model didn’t hold up. There hasn’t been a perfect test, but that’s what the available work found.

Importantly, we’re learning that there are actually some drawbacks to suggesting a one-size-fits-all model of development. I know it seems like a great thing to tell everyone that the same kind of practice gets everyone to the same place, but it turns out that personalizing people’s development can often help them realize their strengths or formulate their practice in a particularly useful way. This one size fits all or almost-all deliberate practice model actually doesn’t fit any of the demonstrated skill development pathways in almost anything, save maybe very constrained sorts of activities, like golf, or activities that are more static and often formulaic perceptual tasks.

One of the most surprising elements in Ericsson’s new book was how he says that many baseball players don’t have better vision than normal. I now know that he based that on a flawed interpretation of some cricket studies, because I see the visual acuity data from baseball players most seasons and hitters’ average visual acuity is usually around 20/11, meaning they see standing from 20 feet away what a normal person with “perfect vision” has to stand at 11 feet to see. Some of them test near 20/8, around the theoretical limit of human vision.

So, I admire much of his work, but disregarding data like that is hard to explain. I can understand from the perspective of maybe when he was doing some of this work early on, that there wasn’t enough respect for practice and quality of practice and to get that he felt he had to take a strong stance about how important practice is. I talked to Janet Starkes, who’s also a legend in the skill acquisition field, and she told me that a couple decades ago she had to take this extremely hard stance of only practice matters in order to try to sway people’s opinions, even though reality is more “centrist,” as she put it.

Scott: Yes, this emerging understanding of the nature-nurture interactions is pointing us to a more nuanced view than the practice model.

“Science doesn’t care about a message. It cares about the truth.”

Indeed, It’s a very complicated issue. The popularity of Malcolm Gladwell’s book, Outliers, reflected society’s desire to soak up a particular message, which is that we can pretty much accomplish whatever we want by pulling ourselves up by our bootstraps and putting in a magical number of practice hours. That there’s a formula for success. Since then, I would say that other bestselling books like The Talent Code and Talent is Overrated— further indicate that society is really hungry for a particular message. Unfortunately, science doesn’t care about a message. It cares about the truth.

Trust me—as someone who grew up being told that I was less intelligent based on my test scores, when I entered the field, I had the same level of desperately wanting that message to be true. After over a decade of research on all the nuanced areas of this, it has become clear to me, however, that the nuanced truth is far more interesting—and I would even say more hopeful—than what the general public realizes, and what I even wanted to believe when I first went into this field.

The truth is that we all vary on a multitude of dimensions and, for better or worse, the development of those dimensions are influenced by a mix of nature and nurture. There are a wide range of cognitive and physical variables, as is becoming quite clear through various methods like twin studies, as well as directly looking at the genes and mutations and a range of genes. Our innate endowment does influence our development and the environment plays an essential role in sculpting and determining what pathways those genes go down. Ignoring those individual differences, I think, ignores what it means for you to be you. A purely deliberate practice model to me is very mechanistic; it treats us as machines: if we put in a certain input, we’ll get a certain output. I think it strips away our humanity.

David: You touched on a lot of really interesting things. Let me go through a couple of them. With respect to the deliberate practice model, or the so-called 10,000 hour rule, again that’s something proposed as a theory and, as you know in the scientific method, you propose a theory or a framework that makes predictions and then you test those predictions. And they must be falsifiable, or else it isn’t really science.

Many, or most of those predictions of this famous model were left untested for years. But a recent meta-analysis looked through over 9,000 relevant publications and identified studies that included measures of deliberate practice and measures of skill. What it found was that for certain types of skill, accumulated hours of deliberate practice mattered more and for other types it mattered less, but it never accounted for nothing and it certainly never accounted for everything. It usually accounted for somewhere between a medium-low and medium-high amount of skill difference, to put it colloquially.

I think this gets to one of the essential differences between scientists and normal people. Scientists can’t just say “it’s not nature or nurture.” Numbers matter. It has to be “how much does each matter and what’s the interplay of each in every different situation?” It’s not just, does this matter? Scientists have to care about magnitude and numbers as they test their theories.

Scott: Anders has responded to those meta-analyses by saying that a lot of the included studies don’t fit his definition of deliberate practice. He wants a controlled situation where the only way we learn something is by having a mentor who gives us step-by-step directions on how to achieve greatness; allows us to incorporate feedback, constantly, all the time; where we are constantly challenged, outside our comfort zone incrementally, et cetera. He’s got a very specific definition of deliberate practice, and he would say that a lot of those studies just look at quantity of hours as opposed to quality of hours.

David: They did attempt to control for deliberate practice in those meta-analyses. So let me make two points about that. One, he also has not tested his own framework. When I wrote about Dan McLaughlin, the golfer who decided to quit his job and do 10,000 hours of golf practice, one of the interesting things was that he consulted with Ericsson—

Scott: And me. I had the pleasure of talking to Dan as well…

“If you’re positing a comprehensive framework of skill development and you can only account for people on the gold medal podium, looking backwards and saying they practiced a lot, then you don’t have a model of skill development.”

David: One of the interesting things Dan said was, “It was great. I got to meet with Professor Ericsson and he said, ‘This is amazing. You’re the first person to try this,’” which I thought was very telling.

Two, it seems that the definition of deliberate practice has evolved to the point where it’s now practice that, in retrospect, did in fact make somebody better. Everything that didn’t isn’t deliberate practice, so it’s an unfalsifiable theory. Again, when I arranged a panel with Ericsson at the American College of Sports Medicine—he was really kind and gracious to come, and it was a great learning opportunity, and I loved and respected that about him—he was beside one of the top skill acquisitions scientists in the physical realm who said, “Look at this, when we take huge numbers of people and put them on identical training programs, there is enormous variation in how much they improve, and we controlled all of their practice in the lab over five university centers over four years. Families tended to stick together. Identical twins really stuck together. This is the variation.” And he responded, “These aren’t experts.”

If you’re positing a comprehensive framework of skill development and you can’t account for square zero, you can only account for people on the gold medal podium, looking backwards and saying they practiced a lot, then you don’t have a model of skill development.

Scott: You also don’t have science. Science is all about prediction.

David: I also think this relates to some of your own work. When you look at creativity, there are things that are clearly not deliberate practice that are really important. How far could daydreaming—which you’ve written about in relation to creativity—be from what qualifies as deliberate practice? I’d say quite far.

Where does your work with creativity fit into what would be considered deliberate practice under that model of one-on-one, hyper-rigorous feedback?

Scott: I’m glad you brought that up because I saw an interview where Anders said, “Deliberate practice is very different from play… Play is not a contributor to expertise, skill development.” I couldn’t disagree more. I think that people who really have fallen in love with a domain, or a method, to them what may appear like deliberate practice often feels like play. It really depends on how you define play; I define play as something deeply, intrinsically enjoyable, where you have a spontaneity to it, where there isn’t necessarily a direct goal. Deliberate practice is so strict, it’s all about the goal, and if it’s the only tool in your toolbox, you are really limiting the chances that you’re going to be creative.

David: It’s one of the reasons why we see this interesting pattern in the sports realm—in non-golf sports—where kids who get highly technical instruction early in life in a single sport don’t go on to become elite. It’s completely the opposite of what you expect from a deliberate practice framework. It’s the Roger Federer model, the kids who play a bunch of different sports, learn a whole variety of skills, a lot of improv, who delay focusing, actually go on to become elite more often. Of course, there are a million different pathways. Steve Nash didn’t play basketball until he was 13. They’re behind in technical skills early on, but they get this broad exposure and range of skills so the thinking is they tend to be much more creative and able to transfer their skills.

“Openness to experience is the best predictor of publicly recognized creative achievement, even better than conscientiousness.”

Scott: Moving to creativity, not just expertise, let me mention one of my favorite longitudinal studies. The E. Paul Torrance studies followed kids starting in elementary school and they’re still following them 50 years later. It found quite clearly that there are a wide range of characteristics that predicted life-long creative achievement—a lot more factors than just persistence or practice.

In fact, they found one of the most important characteristics was the extent to which kids fell in love with a future image of themselves. That has passion, but it also has an imagination component to it. Openness to experience, for instance, we’ve found is the best predictor of publicly recognized creative achievement, even better than conscientiousness.

David: I think we should note before moving on that the emphasis on the quality of practice is a tremendously important contribution. You don’t get better at golf by just going to the driving range and swatting a bunch of balls without being cognitively engaged, you have to work on stuff. But golf isn’t a good model of most things that humans want to excel at. Certainly not the creative domains.

I have a friend who just became the first curator at a major U.S. museum of what’s called “outsider art”, which are basically people who retired from their profession, took up art as a hobby, and it turned out they had a lot of ability and unique expression and they end up having things that are being auctioned for enormous amounts and hung in museums. I would hate to think that people like that would be influenced by a model that says it’s too late for you to start this, don’t do it.

I spent time out running in the countryside of Kenya, and I saw, unbeknownst to me at the time, a guy who had walked off a farm at age 26 and just started training, and two years later was a world record holder, Dennis Kimetto. Good thing he didn’t think that he had started too late to do that.

Scott: Great point. There’s a lot of variability around the mean and, in fact, Dean Simonson’s found those who make the history books in various fields like leadership and music composition, even psychology, actually took the least amount of time to get the prerequisite expertise in the domain. I wrote this article, “Creativity Is Much More Than 10,000 Hours of Deliberate Practice” for Scientific American, and I listed 12 things in favor of that argument that I’m making. One thing that is clear is that all of these traits—physical as well as cognitive—have a heritability coefficient.

It doesn’t mean something that’s innate is immutable, that’s not what it means, but there is a genetic influence. And we know a single gene is very rarely going to explain that much proportion of variance in an outcome. It almost seems like there’s some domains where we, as a society, are totally okay accepting that there’s a genetic component. For instance, high cholesterol, when you go to see a doctor they ask your family history, no one says, “Well the theory of deliberate practice says that we deliberately practiced our way to high cholesterol.”

With our emerging understanding of the genetics involved in mental disorders like schizophrenia and autism, it’s becoming quite clear there’s biological components to all these things. And yet, as a society, we seem blocked when it comes to opening our minds to a nuanced, complex understanding of the deterrents of greatness, which I don’t think are outside the realm of biology, and not altogether different than the development of any other complex outcome.

David: Writing The Sports Gene was a little bit tough for me because I consider myself to be preaching the gospel of practice. I went from a walk-on to university record holder in track and field. I think I can get any human being who doesn’t have serious cardiovascular disease to run a marathon within six months. Not in 2:10, but to run a marathon, so it was weird to feel like I was being put in the place of being a spokesperson for talent.

I think you’ve done a good job of showing some of the cases why aspects of a one-size-fits-all model isn’t as happy of a message as you might think. It isn’t as productive for everyone as you might think, if you look at how it plays out in realms of creativity or the reverse corollaries of it, which is, well, if something didn’t work out for you that’s because you’re too lazy or you didn’t practice right. In many cases we know that the same kind of practice isn’t necessarily right for everybody, and the same path isn’t right for everybody. I think that’s an important message, and not one that should be deemed negative.

Scott: I couldn’t agree more. On a related note, Angela Duckworth’s research—she’s a friend and colleague of mine and I respect her work immensely—showing that stick-to-it-ness and not giving up or having the right mindset are immensely important. I think that a full, comprehensive understanding of the development of greatness is going to require more of a “yes and” approach than a, “It’s this or it’s that” approach. That’s really what I’m calling for, more of “yes it’s grit, but it’s also imagination, it’s also curiosity, it’s also all these things that make you unique as a human being in this world.” I think that way of looking at things is actually more hopeful and exciting because it recognizes the importance of individuality.

David: I agree, and it also happens to be the way that the evidence is pointing. I think it’s great that Anders Ericsson is engaging the public and causing people to have conversations like this. I think a real contribution that science can make is to help people realize that we have to think about how does this and that interact in this particular situation? The influence of one aspect doesn’t crowd out the other. That just doesn’t look like it’s how reality works.