Range: Why Generalists Triumph in a Specialized World

by David Epstein

The bestseller Talent Is Overrated used the Polgar sisters and Tiger Woods as proof that a head start in deliberate practice is the key to success in “virtually any activity that matters to you.” The powerful lesson is that anything in the world can be conquered in the same way. It relies on one very important, and very unspoken, assumption: that chess and golf are representative examples of all the activities that matter to you.

Just how much of the world, and how many of the things humans want to learn and do, are really like chess and golf? Psychologist Gary Klein is a pioneer of the “naturalistic decision making” (NDM) model of expertise; NDM researchers observe expert performers in their natural course of work to learn how they make high-stakes decisions under time pressure. Klein has shown that experts in an array of fields are remarkably similar to chess masters in that they instinctively recognize familiar patterns.

One of Klein’s colleagues, psychologist Daniel Kahneman, studied human decision making from the “heuristics and biases” model of human judgment. His findings could hardly have been more different from Klein’s. When Kahneman probed the judgments of highly trained experts, he often found that experience had not helped at all. Even worse, it frequently bred confidence but not skill.

In 2009, Kahneman and Klein took the unusual step of coauthoring a paper in which they laid out their views and sought common ground. And they found it. Whether or not experience inevitably led to expertise, they agreed, depended entirely on the domain in question. Narrow experience made for better chess and poker players and firefighters, but not for better predictors of financial or political trends, or of how employees or patients would perform. The domains Klein studied, in which instinctive pattern recognition worked powerfully, are what psychologist Robin Hogarth termed “kind” learning environments. Patterns repeat over and over, and feedback is extremely accurate and usually very rapid.

Moravec’s paradox: machines and humans frequently have opposite strengths and weaknesses.

There is a saying that “chess is 99 percent tactics.” Tactics are short combinations of moves that players use to get an immediate advantage on the board. When players study all those patterns, they are mastering tactics. Bigger-picture planning in chess—how to manage the little battles to win the war—is called strategy. As Susan Polgar has written, “you can get a lot further by being very good in tactics”—that is, knowing a lot of patterns—“and have only a basic understanding of strategy.” Thanks to their calculation power, computers are tactically flawless compared to humans. Grandmasters predict the near future, but computers do it better. What if, Kasparov wondered, computer tactical prowess were combined with human big-picture, strategic thinking?

A few years later, the first “freestyle chess” tournament was held. Teams could be made up of multiple humans and computers. The lifetime-of-specialized-practice advantage that had been diluted in advanced chess was obliterated in freestyle. A duo of amateur players with three normal computers not only destroyed Hydra, the best chess supercomputer, they also crushed teams of grandmasters using computers. Kasparov concluded that the humans on the winning team were the best at “coaching” multiple computers on what to examine, and then synthesizing that information for an overall strategy. Human/Computer combo teams—known as “centaurs”—were playing the highest level of chess ever seen. If Deep Blue’s victory over Kasparov signaled the transfer of chess power from humans to computers, the victory of centaurs over Hydra symbolized something more interesting still: humans empowered to do what they do best without the prerequisite of years of specialized pattern recognition.

That test reenacted an experiment from 1973, in which two Carnegie Mellon University psychologists, William G. Chase and soon-to-be Nobel laureate Herbert A. Simon, repeated the De Groot exercise, but added a wrinkle. This time, the chess players were also given boards with the pieces in an arrangement that would never actually occur in a game. Suddenly, the experts performed just like the lesser players. The grandmasters never had photographic memories after all. Through repetitive study of game patterns, they had learned to do what Chase and Simon called “chunking.” Rather than struggling to remember the location of every individual pawn, bishop, and rook, the brains of elite players grouped pieces into a smaller number of meaningful chunks based on familiar patterns.

Studying an enormous number of repetitive patterns is so important in chess that early specialization in technical practice is critical. Psychologists Fernand Gobet (an international master) and Guillermo Campitelli (coach to future grandmasters) found that the chances of a competitive chess player reaching international master status (a level down from grandmaster) dropped from one in four to one in fifty-five if rigorous training had not begun by age twelve. Chunking can seem like magic, but it comes from extensive, repetitive practice.

Our greatest strength is the exact opposite of narrow specialization. It is the ability to integrate broadly. According to Gary Marcus, a psychology and neural science professor who sold his machine learning company to Uber, “In narrow enough worlds, humans may not have much to contribute much longer. In more open-ended games, I think they certainly will. Not just games, in open ended real-world problems we’re still crushing the machines.”

The progress of AI in the closed and orderly world of chess, with instant feedback and bottomless data, has been exponential. In the rule-bound but messier world of driving, AI has made tremendous progress, but challenges remain. In a truly open-world problem devoid of rigid rules and reams of perfect historical data, AI has been disastrous. IBM’s Watson destroyed at Jeopardy! and was subsequently pitched as a revolution in cancer care, where it flopped so spectacularly that several AI experts told me they worried its reputation would taint AI research in health-related fields. As one oncologist put it, “The difference between winning at Jeopardy! and curing all cancer is that we know the answer to Jeopardy! questions.” With cancer, we’re still working on posing the right questions in the first place.

Marcus gave me this analogy for the current limits of expert machines: “AI systems are like savants.” They need stable structures and narrow worlds.

Psychologist Barry Schwartz demonstrated a similar, learned inflexibility among experienced practitioners when he gave college students a logic puzzle that involved hitting switches to turn light bulbs on and off in sequence, and that they could play over and over. It could be solved in seventy different ways, with a tiny money reward for each success. The students were not given any rules, and so had to proceed by trial and error.* If a student found a solution, they repeated it over and over to get more money, even if they had no idea why it worked. Later on, new students were added, and all were now asked to discover the general rule of all solutions. Incredibly, every student who was brand-new to the puzzle discovered the rule for all seventy solutions, while only one of the students who had been getting rewarded for a single solution did. The subtitle of Schwartz’s paper: “How Not to Teach People to Discover Rules”—that is, by providing rewards for repetitive short-term success with a narrow range of solutions.

There are domains beyond chess in which massive amounts of narrow practice make for grandmaster-like intuition. Like golfers, surgeons improve with repetition of the same procedure. Accountants and bridge and poker players develop accurate intuition through repetitive experience. Kahneman pointed to those domains’ “robust statistical regularities.” But when the rules are altered just slightly, it makes experts appear to have traded flexibility for narrow skill.

Erik Dane, a Rice University professor who studies organizational behavior, calls this phenomenon “cognitive entrenchment.” His suggestions for avoiding it are about the polar opposite of the strict version of the ten-thousand-hours school of thought: vary challenges within a domain drastically, and, as a fellow researcher put it, insist on “having one foot outside your world.”

The most successful experts also belong to the wider world. “To him who observes them from afar,” said Spanish Nobel laureate Santiago Ramón y Cajal, the father of modern neuroscience, “it appears as though they are scattering and dissipating their energies, while in reality they are channeling and strengthening them.”

As psychologist and prominent creativity researcher Dean Keith Simonton observed, “rather than obsessively focus[ing] on a narrow topic,” creative achievers tend to have broad interests. “This breadth often supports insights that cannot be attributed to domain-specific expertise alone.”

No amount of cajoling, explanation, or examples could get remote villagers to use reasoning based on any concept that was not a concrete part of their daily lives. The farmers and students who had begun to join the modern world were able to practice a kind of thinking called “eduction,” to work out guiding principles when given facts or materials, even in the absence of instructions, and even when they had never seen the material before. This, it turns out, is precisely what Raven’s Progressive Matrices tests.

Some of the changes wrought by modernity and collective culture seem almost magical. Luria found that most remote villagers were not subject to the same optical illusions as citizens of the industrialized world, like the Ebbinghaus illusion.

Those findings have been repeated in other traditional societies, and scientists have suggested it may reflect the fact that premodern people are not as drawn to the holistic context—the relationship of the various circles to one another—so their perception is not changed by the presence of extra circles. To use a common metaphor, premodern people miss the forest for the trees; modern people miss the trees for the forest.

The more they had moved toward modernity, the more powerful their abstract thinking, and the less they had to rely on their concrete experience of the world as a reference point.

In Flynn’s terms, we now see the world through “scientific spectacles.” He means that rather than relying on our own direct experiences, we make sense of reality through classification schemes, using layers of abstract concepts to understand how pieces of information relate to one another.

Exposure to the modern world has made us better adapted for complexity, and that has manifested as flexibility, with profound implications for the breadth of our intellectual world.

Jeannette Wing, a computer science professor at Columbia University and former corporate vice president of Microsoft Research, has pushed broad “computational thinking” as the mental Swiss Army knife. She advocated that it become as fundamental as reading, even for those who will have nothing to do with computer science or programming. “Computational thinking is using abstraction and decomposition when attacking a large complex task,” she wrote. “It is choosing an appropriate representation for a problem.”

Perhaps the greatest improv master of all could not read, period—words or music. Django Reinhardt was born in Belgium in 1910, in a Romani caravan. His early childhood talents were chicken stealing and trout tickling—feeling along a riverbank for fish and rubbing their bellies until they relaxed and could be tossed ashore.

The strict deliberate practice school describes useful training as focused consciously on error correction. But the most comprehensive examination of development in improvisational forms, by Duke University professor Paul Berliner, described the childhoods of professionals as “one of osmosis,” not formal instruction. “Most explored the band room’s diverse options as a prelude to selecting an instrument of specialization,” he wrote. “It was not uncommon for youngsters to develop skills on a variety of instruments.” Berliner added that aspiring improvisational musicians “whose educational background has fostered a fundamental dependence on [formal] teachers must adopt new approaches to learning.” A number of musicians recounted Brubeck-like scenarios to Berliner, the time a teacher found out that they could not read music but had become adept enough at imitation and improvisation that “they had simply pretended to follow the notation.” Berliner relayed the advice of professional musicians to a young improvisational learner as “not to think about playing—just play.”

Charles Limb, a musician, hearing specialist, and auditory surgeon at the University of California, San Francisco, designed an iron-free keyboard so that jazz musicians could improvise while inside an MRI scanner. Limb saw that brain areas associated with focused attention, inhibition, and self-censoring turned down when the musicians were creating. “It’s almost as if the brain turned off its own ability to criticize itself,” he told National Geographic. While improvising, musicians do pretty much the opposite of consciously identifying errors and stopping to correct them.

Cecchini has taught musicians who played professionally for the Chicago Symphony, which in 2015 was rated as the top orchestra in the country and fifth in the world by a panel of critics. “It’s easier for a jazz musician to learn to play classical literature than for a classical player to learn how to play jazz,” he said. “The jazz musician is a creative artist, the classical musician is a re-creative artist.”

After Django Reinhardt lit the nightclub music scene on fire, classically trained musicians began trying to transition to jazz. According to Michael Dregni, who wrote multiple books on that period, improvisation was “a concept that went against conservatory training. . . . After years of rigorous conservatory training, it was an impossible transition for some.” Leon Fleisher, regarded as one of the great classical pianists of the twentieth century, told the coauthor of his 2010 memoir that his “greatest wish” was to be able to improvise. But despite a lifetime of masterful interpretation of notes on the page, he said, “I can’t improvise at all.”

Cecchini’s analogy to language learning is hardly unique. Even the Suzuki Method of music instruction, synonymous in the public consciousness with early drilling, was designed by Shinichi Suzuki to mimic natural language acquisition.

In totality, the picture is in line with a classic research finding that is not specific to music: breadth of training predicts breadth of transfer. That is, the more contexts in which something is learned, the more the learner creates abstract models, and the less they rely on any particular example. Learners become better at applying their knowledge to a situation they’ve never seen before, which is the essence of creativity.

Compared to the Tiger Mother’s tome, a parenting manual oriented toward creative achievement would have to open with a much shorter list of rules. In offering advice to parents, psychologist Adam Grant noted that creativity may be difficult to nurture, but it is easy to thwart. He pointed to a study that found an average of six household rules for typical children, compared to one in households with extremely creative children. The parents with creative children made their opinions known after their kids did something they didn’t like, they just did not proscribe it beforehand. Their households were low on prior restraint.

Rather than letting students grapple with some confusion, teachers often responded to their solicitations with hint-giving that morphed a making-connections problem into a using-procedures one. That is exactly what the charismatic teacher in the American classroom was doing. Lindsey Richland, a University of Chicago professor who studies learning, watched that video with me, and told me that when the students were playing multiple choice with the teacher, “what they’re actually doing is seeking rules.” They were trying to turn a conceptual problem they didn’t understand into a procedural one they could just execute. “We’re very good, humans are, at trying to do the least amount of work that we have to in order to accomplish a task,” Richland told me. Soliciting hints toward a solution is both clever and expedient. The problem is that when it comes to learning concepts that can be broadly wielded, expedience can backfire.

In 2007, the U.S. Department of Education published a report by six scientists and an accomplished teacher who were asked to identify learning strategies that truly have scientific backing. Spacing, testing, and using making-connections questions were on the extremely short list. All three impair performance in the short term. As with the making-connections questions Richland studied, it is difficult to accept that the best learning road is slow, and that doing poorly now is essential for better performance later. It is so deeply counterintuitive that it fools the learners themselves, both about their own progress and their teachers’ skill.

Psychologist Robert Bjork first used the phrase “desirable difficulties” in 1994. Twenty years later, he and a coauthor concluded a book chapter on applying the science of learning like this: “Above all, the most basic message is that teachers and students must avoid interpreting current performance as learning. Good performance on a test during the learning process can indicate mastery, but learners and teachers need to be aware that such performance will often index, instead, fast but fleeting progress.”

Most problems, of course, are not new, so we can rely on what Gentner calls “surface” analogies from our own experience. “Most of the time, if you’re reminded of things that are similar on the surface, they’re going to be relationally similar as well,” she explained. Remember how you fixed the clogged bathtub drain in the old apartment? That will probably come to mind when the kitchen sink is clogged in the new one. But the idea that surface analogies that pop to mind work for novel problems is a “kind world” hypothesis, Gentner told me. Like kind learning environments, a kind world is based on repeating patterns. “It’s perfectly fine,” she said, “if you stay in the same village or the same savannah all your life.” The current world is not so kind; it requires thinking that cannot fall back on previous experience.

The trouble with using no more than a single analogy, particularly one from a very similar situation, is that it does not help battle the natural impulse to employ the “inside view,” a term coined by psychologists Daniel Kahneman and Amos Tversky. We take the inside view when we make judgments based narrowly on the details of a particular project that are right in front of us.

Our natural inclination to take the inside view can be defeated by following analogies to the “outside view.” The outside view probes for deep structural similarities to the current problem in different ones. The outside view is deeply counterintuitive because it requires a decision maker to ignore unique surface features of the current project, on which they are the expert, and instead look outside for structurally similar analogies. It requires a mindset switch from narrow to broad.

Following their private-equity-investor experiment, the outside-view researchers turned to the movie business, a notoriously uncertain realm with high risk, high reward, and a huge store of data on actual outcomes. They wondered if forcing analogical thinking on moviegoers could lead to accurate forecasts of film success. They started by giving hundreds of movie fans basic film information—lead actor names, the promotional poster, and a synopsis—for an upcoming release. At the time, those included Wedding Crashers, Fantastic Four, Deuce Bigalow: European Gigolo, and others. The moviegoers were also given a list of forty older movies, and asked to score how well each one probably served as an analogy to each upcoming release. The researchers used those similarity scores (and a little basic film information, like whether it was a sequel) to predict the eventual revenue of the upcoming releases. They pitted those predictions against a mathematical model stuffed with information about seventeen hundred past movies and each upcoming film, including genre, budget, star actors, release year, and whether it was a holiday release. Even without all that detailed information, the revenue predictions that used moviegoer analogy scores were vastly better. The moviegoer-analogies forecast performed better on fifteen of nineteen upcoming releases. Using the moviegoers’ analogies gave revenue projections that were less than 4 percent off for War of the Worlds, Bewitched, and Red Eye, and ...

Netflix came to a similar conclusion for improving its recommendation algorithm. Decoding movies’ traits to figure out what you like was very complex and less accurate than simply analogizing you to many other customers with similar viewing histories. Instead of predicting what you might like, they examine who you are like, and the complexity is captured therein.

Interestingly, if the researchers used only the single film that the movie fans ranked as most analogous to the new release, predictive power collapsed. What seemed like the single best analogy did not do well on its own. Using a full “reference class” of a...

Evaluating an array of options before letting intuition reign is a trick for the wicked world.

In one of the most cited studies of expert problem solving ever conducted, an interdisciplinary team of scientists came to a pretty simple conclusion: successful problem solvers are more able to determine the deep structure of a problem before they proceed to match a strategy to it. Less successful problem solvers are more like most students in the Ambiguous Sorting Task: they mentally classify problems only by superficial, overtly stated features, like the domain context. For the best performers, they wrote, problem solving “begins with the typing of the problem.” As education pioneer John Dewey put it in Logic, The Theory of Inquiry, “a problem well put is half-solved.”

Dunbar witnessed important breakthroughs live, and saw that the labs most likely to turn unexpected findings into new knowledge for humanity made a lot of analogies, and made them from a variety of base domains. The labs in which scientists had more diverse professional backgrounds were the ones where more and more varied analogies were offered, and where breakthroughs were more reliably produced when the unexpected arose.

Hesselbein decided to form a foundation for nonprofit management, to help bring best business practices to social entrepreneurship.

She never did graduate from college, but her office is festooned with twenty-three honorary doctorates, plus a glistening saber given to her by the U.S. Military Academy for teaching leadership courses—as well as the Presidential Medal of Freedom, the highest civilian award in the United States. When I visited just after her 101st birthday, I brought her a cup of steamed milk, as I had been advised, and right away asked what training had prepared her for leadership. Wrong question. “Oh, don’t ask me what my training was,” she replied with a dismissing hand wave. She explained that she just did whatever seemed like it would teach her something and allow her to be of service at each moment, and somehow that added up to training. As Steven Naifeh said regarding Van Gogh’s life, some “undefinable process of digestion” occurred as diverse experiences accumulated. “I was unaware that I was being prepared,” she told me. “I did not intend to become a leader, I just learned by doing what was needed at the time.”

At the first ever Girl Scout training event Hesselbein attended, she heard another new troop leader complain that she was getting nothing from the session. Hesselbein mentioned it to a dress-factory worker who was also volunteering, and the woman told her, “You have to carry a big basket to bring something home.” She repeats that phrase today, to mean that a mind kept wide open will take something from every new experience.

Because personality changes more than we expect with time, experience, and different contexts, we are ill-equipped to make ironclad long-term goals when our past consists of little time, few experiences, and a narrow range of contexts.

Bingham had noticed that established companies tended to approach problems with so-called local search, that is, using specialists from a single domain, and trying solutions that worked before. Meanwhile, his invitation to outsiders worked so well that it was spun off as an entirely separate company. Named InnoCentive, it facilitates entities in any field acting as “seekers,” paying to post “challenges” and rewards for outside “solvers.” A little more than one-third of challenges were completely solved, a remarkable portion given that InnoCentive selected for problems that had stumped the specialists who posted them. Along the way, InnoCentive realized it could help seekers tailor their posts to make a solution more likely. The trick: to frame the challenge so that it attracted a diverse array of solvers. The more likely a challenge was to appeal not just to scientists but also to attorneys and dentists and mechanics, the more likely it was to be solved.

“It took me three evenings to write it up,” an outside solver told the journal Science after he answered Johnson & Johnson’s request for help with a production problem in the manufacture of tuberculosis medication. “I think it’s strange that a major pharma company cannot solve this kind of problem.” Karim Lakhani, codirector of the Laboratory for Innovation Science at Harvard, had InnoCentive solvers rate problems on how relevant they were to their own field of specialization, and found that “the further the problem was from the solver’s expertise, the more likely they were to solve it.”