What do Tiger Woods and Roger Federer have in common with hedgehogs and foxes? Find out as we sit down with David Epstein, New York Times bestselling author of The Sports Gene and Range: Why Generalists Triumph in a Specialized World, who argues that in most fields, generalists—not specialists—are primed to outperform.
David's comments are edited excerpts from our podcast, which you can listen to in full below.
Why should we be generalists instead of specialists?
David: Research into the development of elite performers—athletes, musicians, artists, inventors—shows not the Tiger Woods pattern but the Roger Federer pattern.
Tiger Woods' story is one of early specialization. Roger Federer, on the other hand, played a number of different sports, and that is typical for elite athletes.
Science that tracks these athletes' development shows that during a “sampling period,” they learn about their interests, gain a breadth of general skills, and test their abilities. Thus, when they do specialize, they come to it with a much broader skill set.
That's not to say that specialists are unimportant, but that the Tiger path is actually the exception and we've been treating it as the norm. And I don't think we're in danger of undervaluing specialization, whereas we are in danger of missing out on a lot of the power of.
That's contrary to conventional wisdom, isn't it?
David: Stories we've accepted about specialization may not be what we think they are. Consider parents who insist their child play an instrument. With six hours of practice a day, the child becomes proficient. But you don't hear about the part where most children say, “You picked this instrument, not me,” and quit.
Isn't delayed specialization risky?
David: It's psychologically upsetting to delay specializing because we think of it as the exception, but it's actually not.
Harvard researchers are trying to figure out the habits of people who maximize their “match quality,” which is a term economists use to describe the fit between an individual and the work he or she does, because doing so is essential for motivation, growth, and overall success.
And people who maximize their match quality tend to zigzag their way to a place where they can feel fulfilled and succeed. Successful people, from master sommeliers to investment bankers, say they always viewed themselves as oddballs because they tried many things before they specialized.
So we have a neurosis around generalization, but it's not warranted.
Political science writer Philip Tetlock talks about foxes and hedgehogs. Are they similar to generalists and specialists?
David: Yes. Hedgehogs, as Tetlock says, know one big thing, so they're the specialists. Foxes know many little things, so they're the generalists.
Tetlock asked people to give specific probabilities of events happening with specific deadlines. His project, which ran 20 years, took about 82,000 different forecasts.
Tetlock found that credentials have nothing to do with how good people are at forecasting. What makes them good at forecasting is how broad their thinking is—their curiosity, their mental models, their information consumption.
Some of Tetlock's prediction questions were hard. One was whether Greece would leave the Eurozone during its debt negotiations. There was no precedent for this.
If the hedgehogs have to predict a scenario, they drill into all of the little specific details of that exact scenario. To answer the Greece question, they drilled into all the specifics of Greece.
Foxes take a much broader view and start looking for other similar scenarios, like in history. To answer the Greece question, foxes were more likely to go looking for analogies in other areas.
It turns out that the hedgehogs' inside view tends to be extremely inaccurate. Whatever scenario you start investigating, you just deem it more and more likely as you get more information.
Can you resolve the contradiction between Angela Duckworth's concept of grit, or dogged perseverance, and generalization before specialization?
David: Duckworth's 12-question grit scale, which includes a question about consistency of interests, is focused on the short term. And I think it's really problematic to extrapolate grit beyond a specific environment.
The best example may be a study of cadets going through “beast barracks” at the U.S. Military Academy at West Point—a six-week orientation during which they face physical and emotional challenges.
Duckworth and her colleges found that the grit survey was a better predictor of who would make it through than more traditional measures, like test scores or athleticism.
But it was a short-term program, and you shouldn't extrapolate such a narrow goal to the rest of life. If you follow those gritty cadets through the military academy, about half of them drop out of the army the day they're allowed to. And they do that because they've learned about themselves.
You make a distinction between kind and wicked environments. Could you explain?
David: Psychologist Robin Hogarth coined these terms based on a disagreement between two famous researchers of expertise and judgement.
Danny Conoman found that many experts don't get any better with experience; Gary Klein found that they do. They ultimately realized it depends on the environments in which those experts are learning.
In a kind learning environment, like golf, information is visible, the next steps are clear, people take turns, and feedback is immediate and perfectly accurate. In those environments, you improve by repetition, so specialization works.
In wicked environments, information is hidden, people don't take turns, the next steps are unclear, and you get poor or delayed feedback (if any). In those environments, people who are too narrowly specialized will not improve, and in some cases will get worse. That is often the case in today's complex working environment.
In the humans vs. machines debate, would you say that humans win because they can deal with wicked environments more easily than machines can?
David: When Deep Blue beat Garry Kasparov in chess in 1997, the cover in Newsweek for that match was, “The Brain's Last Stand.”
But chess is a perfect activity for computers. It's rule-bound and based on pattern recognition, and there's an enormous store of previous data.
If you go into a slightly less consistent environment like driving, challenges remain, and some of them are formidable.
And at the far end of the spectrum, like cancer research, IBM Watson, which was great on Jeopardy, has been a titanic flop. That's because we know the answers to Jeopardy, but in open-world problems, humans trounce machines.
How do these ideas tie into fast vs. slow learning?
David: The example I use is training naval commanders to respond to certain types of threats.
One group trains on one problem at a time. They repeat a single problem until they're good at it, then move on to another problem.
The other group gets different problems each time, and they get really frustrated because they don't appear to be improving.
But when you bring those people back and give them scenarios that neither of them have ever seen before, the frustrated group destroys the group that was practicing the same things over and over—because they're forced to match types of strategies to types of problems. That's how you build general conceptual frameworks.
So, the more broad your training, the more able you are to transfer your knowledge to scenarios you've never seen before.
Despite your views, you and Malcom Gladwell, whose 10,000-hour rule makes the case that any task can be accomplished just by practicing, are still friends (and running buddies), correct?
David: Yes, Gladwell and I were just on a panel at the MIT Sloan sports analytics conference, and at the end of the panel, he said he changed his mind about the 10,000-hour rule. He still believes a lot of deliberate practice is important to achieving success—that's uncontroversial. But he said, “I assumed that implied that early specialization was the way to go, and I've changed my mind.