Hidden Games

The Surprising Power of Game Theory to Explain Irrational Human Behavior


By Erez Yoeli

By Moshe Hoffman

Formats and Prices




$24.99 CAD

This item is a preorder. Your payment method will be charged immediately, and the product is expected to ship on or around April 5, 2022. This date is subject to change due to shipping delays beyond our control.

Two  MIT economists  show  how  game theory—the ultimate theory of rationality—explains irrational behavior    

We like to think of ourselves as rational. This idea is the foundation for classical economic analysis of human behavior, including the awesome achievements of game theory. But as behavioral economics shows, most behavior doesn’t seem rational at all—which, unfortunately, to cast doubt on game theory’s real-world credibility.  

In Hidden Games, Moshe Hoffman and Erez Yoeli find a surprising middle ground between the hyperrationality of classical economics and the hyper-irrationality of behavioral economics. They call it hidden games. Reviving game theory, Hoffman and Yoeli use it to explain our most puzzling behavior, from the mechanics of Stockholm syndrome and internalized misogyny to why we help strangers and have a sense of fairness.  

Fun and powerfully insightful, Hidden Games is an eye-opening argument for using game theory to explain all the irrational things we think, feel, and do.




FOR MANY OF US, IT MIGHT BE HARD TO IMAGINE THAT THE COVER OF Sports Illustrated, which we are used to seeing adorned by athletes and models sporting bats, balls, helmets, and bikinis, once featured the laughing, square-jawed face of Bobby Fischer. The 1972 cover, now a collectible on auction sites like eBay, celebrated Fischer’s unprecedented twenty-game winning streak, not in baseball, basketball, or football—but in chess. Fischer’s performance in the game, and his dominance over Soviet rivals, had catapulted him into the public eye like no other chess player before or since. Many of the game’s greatest players considered him, and still do, to be the greatest of all time—the GOAT, to use the term usually applied to the likes of Michael Jordan, LeBron James, Simone Biles, Katie Ledecky, and Tom Brady.

How did Fischer, who grew up in Brooklyn in a cash-strapped, single-parent household, become so great? The answer is no different from the one that might be given for the likes of Jordan, James, Biles, Ledecky, and Brady: a touch of luck and a ton of practice, motivated by an obsessive passion. Frank Brady, Fischer’s longtime biographer, reports that by the time he was nine, if Bobby wasn’t playing chess then he was studying it, bent over his board or book in such a rapture that he wouldn’t pause to turn on the apartment lights when it grew dark. To coax him into the bath, his mother would lay a cabinet door over the tub and place his chessboard on top of it. (Getting him to then leave the tub was another challenge.) He certainly couldn’t be bothered with school, which he quit as soon as it was legal to do so.

Indeed, an obsessive passion is a key ingredient in the success of so many greats.

When Itzhak Perlman, the virtuoso violinist, first asked to play after hearing classical music on the radio at the age of three, he was denied admission to a local conservatory on the grounds that he was too small to hold a violin. So, the sickly boy, who soon contracted polio and today is still bound to crutches and a wheelchair, taught himself to play on a toy fiddle. By the age of ten, he was giving critically acclaimed recitals, and at nineteen, he appeared on The Ed Sullivan Show—for the second time—alongside the Rolling Stones.1

The mathematician Srinivasa Ramanujan was, in his short thirty-three years, so prolific that today an entire peer-reviewed journal is devoted to publishing results that derive or otherwise relate to those that he had claimed or proved. Like Fischer and Perlman, Ramanujan developed his passion at an early age, absorbing all he could from the college-aged lodgers who stayed at his mother’s home and reading math textbooks cover to cover. One, a collection of five thousand theorems that is particularly credited with elevating his genius, would be impossibly tedious even to the most enthusiastic participant in a high school Olympiad. As an adult, his work engrossed him so fully that he neglected his wife and even his own health, dying from complications of dysentery that doctors now think could have been cured had Ramanujan bothered to take a break from work to receive care.

Or what about Marie Curie? She’s still the only person to ever win two Nobel prizes in science. As a student in Paris, she was so engrossed in her studies that she often forgot to eat. She would remain so engrossed until her dying day, eschewing prizes and awards because they took her away from scientific pursuits. She even failed to acquire the funds from her (second) Nobel Prize out of what she termed “sheer laziness.” (She finally picked up the award money during World War I so she could contribute it to the war effort.)

For Picasso, it wasn’t chess, violin, math, or science but art, which he continued to produce obsessively for the entirety of his life. It is estimated that he produced over fifty thousand (!) works of art,2 regularly reinventing himself, when any other successful artist likely would have rested on his laurels.

As mere mortals, the rest of us marvel at these obsessive passions. If only we could be drawn to shooting free throws as much as LeBron James, grinding through endgames as much as Fischer, or flipping through theorems as much as Ramanujan, we might be so much more successful! Each New Year, when we assemble our resolutions, we hope that we, too, magically develop the elusive passion that will transform our working hours from a chore into a labor of love. Yet, inevitably, by early February, when no such fire has been sparked, aimless Netflix and Instagram sessions overtake our aspirations. Why can’t we be more like Bobby and Itzhak? How on Earth could Ramanujan have possibly been enthralled with a textbook of five thousand theorems?! Hand me the remote. Why were they, alongside Einstein and Picasso, anointed to receive the magic fire of passion, as though from Cupid casting arrows to a lucky few?

And why did they develop their particular passions—for basketball, chess, math, physics, or whatever? Why didn’t Picasso, who was sympathetic to Catalan rebels and antifascists, devote his prodigious energies to war instead of art? Why didn’t Einstein obsess over chess? Why would Fischer, who would sit still for hours as he poured over chess books and whose IQ was certainly no impediment, immediately grow restless as soon as a homework assignment of any kind was put in front of him (it inevitably went unfinished, to his mother’s great consternation)?

In short, how does passion—this elusive maker of greatness— work?

There are entire fields devoted to judgment, decison-making, and positive psychology, and sections of bookstores devoted to self-help, so you’d think someone would have a ready answer to these relatively fundamental questions by now. How can we understand the decisions we make, or what makes us happy, if we don’t understand what makes us passionate and gives our life meaning? Indeed, there are some things we know. We know, for instance, that passion goes hand in hand with a sense of meaning, purpose, and satisfaction and that it increases with praise or decreases if someone offers to pay us for our labors. But why? Why do passions work these ways? Do passions simply defy explanation?



Of course, some aspects of aesthetics have ready or well-understood explanations. We know why the wealthy and powerful pay for images of themselves, and why the church has long paid for art that teaches its myths and histories, which was especially useful back in the day when most parishioners were illiterate. We also know that art sometimes takes the things we already enjoy looking at, like symmetric faces, fecund women, or lush lakeside scenery, and exaggerates them. When it comes to music, some of it sounds like water or is otherwise soothing. And some of it keeps a steady beat, so that soldiers can march or townsfolk can dance. As for our food, we know it tastes better when it’s more nutritionally dense. Hence bacon. And that in places where foodborne illnesses are a constant risk, we develop a taste for the spicy concoctions that help to inhibit the growth of bacteria.

But these leave unexplained so much of what’s going on. What’s with the complicated rhyming schemes of Renaissance bards like Shakespeare and modern ones like MF Doom, Chance the Rapper, and Eminem? Or with the highly tannic wines of Bordeaux’s celebrated Left Bank? These aren’t inherently more pleasant. Nor are they simply exaggerating what we already find pleasant. This is made clear by the fact that to the novice Shakespeare is incomprehensible and tannic wines taste too bitter and astringent (while Three Buck Chuck is perfectly quaffable). We don’t mean to disparage these great works of art and culture. They’re great, just not because they’re inherently pleasant. So, what makes them great?

Also unexplained are the ubiquitous Easter eggs that artists of all genres litter in their artwork for critics and enthusiasts to discover by poring over the work for decades or sometimes centuries—while the rest of us to try to glean those hidden meanings from CliffsNotes or, for the exceptionally tenacious, from the ponderous writings of critics.

To find the Easter eggs, identify the best vintages, and understand Shakespeare, we can turn to art historians and critics. But to learn why—why we get excited by these things in the first place—we’ll need some of the tools developed in this book.

ALTRUISM IS ANOTHER DOMAIN WE’LL PUZZLE OVER. NOT JUST OVER why people are altruistic in the first place but also over the bizarre forms altruism takes.

For starters, it’s clear that though we are quite caring and giving, we are not driven by impact and don’t give in the most effective ways. We are moved to donate to GoFundMe campaigns for needy pets, instead of earmarking the funds for high-impact charities that can, by operating leanly and tackling some of humanity’s most pressing problems, save a human life for less than $5,000.3 When presented with a matching fund, we hardly respond, even though our money is going twice as far. When asked how much we’d donate toward safety nets that save migrating birds from being killed by wind turbines, we give the same response irrespective of whether we’re told the nets would save two thousand birds or two hundred thousand.4 We volunteer for Habitat for Humanity even though our airfare could have been better spent hiring local labor that is both more skilled and desperate for work. We care enough to turn off the lights when we leave the room but mindlessly leave the AC on, more than wiping out the gains from hitting the light switch.

We’re not just ineffective, we’re flat out ignorant. Most of us have, at best, a faint idea of what our donations are used for, and virtually none of us puts the same care into selecting a charity that we do into selecting a restaurant or vacation destination. We’re just as bad when it comes to conserving energy and recycling. Did you know, for instance, that recycling metal has roughly nine times the impact of recycling paper or plastic? And that recycling paper or plastic is far more impactful than recycling glass? Don’t believe us? Google it. But notice that this is probably the first time you’ve bothered to google it.

We’re not just ignorant, we’re strategically ignorant. We would never knowingly infect a sexual partner with an STI but are content with not getting tested, even if we know we are at high risk and testing is freely available at nearby clinics.

We don’t just avoid the information; we also avoid the ask. We might give to Planned Parenthood if asked but pull out our phones and try to look really busy when we see the nonprofit’s volunteers asking for donations on city sidewalks. And we’ll, of course, always do a friend a favor but might avoid calling if we suspect said friend needs said favor.

There’s more. Most of us have no qualms paying four dollars for a cortado rather than donating those funds to the poor. But we would never dream of taking four dollars from the poor to pay for a cortado. Why this distinction between action and inaction when the effect is the same? Why does altruism work this way?

THESE ARE THE KINDS OF QUESTIONS WE’LL TACKLE IN THIS BOOK. What tricks do cable news networks use to misinform? Why does motivated reasoning work the way it does? Internalized racism? Why is modesty a virtue? Where does our sense of right come from? Why couldn’t the Hatfields and McCoys bury their hatchets?

In short, we’ll ask: Why are human preferences and ideologies the way they are? Why do they work the way they do?

People tend to respond to questions like these with proximate answers, for example: We love tannic wines because they are more interesting. We love crafts because it is satisfying to work on discrete projects with a finite timeline, where we can quickly see the end results, or we develop a passion for research because we like the freedom to engage in long, detailed explorations of a particular topic and really become experts in it. We give out of empathy for the recipient and do so ineffectively because empathy itself is a blunt tool that’s not so sensitive to efficacy.

While such responses are often interesting, helpful, and valid, they aren’t really answers, at least not in the sense we will be looking for in this book. Sure, tannic wines are more interesting, but what counts as interesting? And why do we even care if they’re interesting? Sure, some people develop a passion only when they quickly see the results of their handiwork while others only get excited by longer, in-depth projects, but we’re still left wondering why some are drawn in one direction while others are drawn in the opposite one, as well as why anyone develops any passion at all. Sure, empathy is a blunt tool, but why? Each of these answers raises at least as many questions as we started with!

We will, instead, attempt to give answers that are, in some sense, more ultimate. In doing so, the key tool we will be using is, of course, game theory.

Game theory is a mathematical tool kit designed to help us figure out how people, firms, countries, and so on will behave in interactive settings—when it matters not only what they do but also what others do. The tool kit has been successfully deployed to help firms design and bid in auctions (where how each bidder should bid depends on others’ bids). It is also a cornerstone of federal antitrust regulation. At the Federal Trade Commission and US Department of Justice, armies of economists spend their days evaluating proposed mergers and acquisitions with the help of a game theory model called Cournot competition (which helps them predict how prices will change, taking into account that all firms in the market will react to what the merged firm does and vice versa). A few blocks away, at the US Department of State, game theory has influenced the thinking of generations of diplomats. For instance, the United States’ cold war strategy of mutual destruction and nuclear brinksmanship was reinforced by the game-theoretic analyses of Thomas Schelling (which took into account that the number of nukes the US should make depended on the number that the USSR had and vice versa).

You might be thinking to yourself that this has absolutely nothing to do with the kinds of behaviors that the book opened with. People aren’t even trying to optimize when they become passionate about playing chess, develop new art movements, or give to charity. They do these things based on intuition or feel or… it just kinda happens without them even realizing it at all. That sounds nothing like the cold-hearted calculus involved in boardroom and situation-room decision-making.

Moreover, you might also be thinking that game theory traditionally rests on a key assumption that’s, well, let’s say questionable: the assumption that people behave optimally. That we are rational. That we have all the relevant information and use it as a computer might to maximize its benefits—doing complex calculations in the process. Maybe this assumption is decent for the crew in the boardroom, strategizing over their radio spectrum bid, but for the rest of us going about our day-to-day lives? There have been not one but two Nobel Prizes in economics for emphatically knocking that assumption down (Daniel Kahneman’s in 2002 and Richard Thaler’s in 2017).5 Even some of our motivating puzzles—willingly dying for a cause, giving to ineffective charities when effective ones stand at the ready—seem to be strong evidence in Danny and Dick’s favor.

We are going to use these two arguments to cancel each other out. Yes, people are quite often quite bad at optimizing when they are relying on their conscious minds to do the optimization. But when they are not consciously optimizing, and it is learning and evolution doing the optimization—as we will argue is often the case for tastes and beliefs—things start to look a lot more promising.

When it comes to evolution, the logic is likely already familiar. People’s tastes evolved to motivate us to act in ways that benefit us. We evolved a taste for fatty, salty, and sweet foods because that motivated us to seek out foods high in fat, salt, and calories in an environment where these were rare. We evolved an attraction to symmetrical faces, chiseled jaws, and broad hips because this motivated us to seek partners who were more healthy, successful, and fertile.6

But it’s not like rap fans evolved from caveman ancestors who sat around the fire trading rhymes while modern art fans’ ancestors devoted their leisure time to abstract cave painting (“Ceci n’est pas une mammouth laineux”). Most of the tastes and beliefs we’re interested in aren’t biologically ingrained in us. They’re learned. So, in the next chapter, we’ll make the same argument for learning that we just made for biological evolution and show that learning (aka cultural evolution) does the job just as well (and quite a bit faster). We’ll see how cultural practices end up being highly tuned to our environment and needs: how, for instance, the igloo was tweaked and tuned over generations until it could keep the Inuit warm in the icy tundra or how traditional methods of preparing corn eked out extra nutritional value from this nutritionally sparse food staple—and without anyone consciously thinking about thermal dynamics or chemistry. We’ll also see how the spices people come to savor reflect their culture’s need to combat foodborne illness and how superstitions and taboos surrounding foods reduce the risk of dangerous illnesses during pregnancy.

After the chapter on learning, we will cover a few distinctions (primary versus secondary rewards, ultimate versus proximate, emic versus etic) that will help us interpret game theory when it is hidden—working its magic on our beliefs and preferences behind the scenes, with the help of evolutionary and learning processes.

After that, we will finally get to some game theory. But it won’t yet be focused on humans. Instead, we will have a chapter devoted to animal sex ratios—the ratio of males to females in any given species—which is a well-known application of game theory from biology. This chapter will introduce some of the key concepts in game theory and show off just how powerful it can be. It will also help us to see how a game is interpreted and applied when evolution is doing the optimizing.

Then, we’re off to the races. From that point out, each chapter will present a handful of seemingly irrational human behaviors and a hidden game or two that will help us uncover the underlying rationale behind these seemingly irrational behaviors.

That’s the plan. Shall we get to it?



IN THIS CHAPTER, WE WILL LAY OUT JUST HOW POWERFUL LEARNING processes are at getting us to behave optimally. We will see that the power of learning is not limited to cases where we have any awareness of what is being optimized. Nor is the power of learning limited to our behaviors; often our beliefs and preferences are sculpted as well.

Why talk about any of this? Because it lays the groundwork for using game theory even when people aren’t rational and even when we are trying to explain their puzzling preferences and beliefs.


In video footage from the 1950s that can still be found floating about on YouTube,1 B. F. Skinner stands before some green and white lab equipment in a dark tie and a white button-down, with the sleeves pulled down all the way to his wrists. Speaking into the microphone in a pleasing mid-Atlantic accent that was made for television, Skinner narrates as he trains a pigeon to spin about in a counterclockwise direction. Skinner’s strategy is simple. Each time the pigeon turns to her left, Skinner opens a trough and gives the pigeon a treat of a few grains, but if the pigeon stands in place or turns to the right, Skinner leaves the trough closed.

The camera zooms in on the pigeon, who, at first, just bobs back and forth. Eventually, randomly, she shuffles half-heartedly toward the left. Click. The trough opens, and the pigeon quickly nabs the snack. The trough closes, and the pigeon looks about. Again, she bobs a bit, and shuffles about somewhat aimlessly.

“I’m waiting for it to turn counterclockwise now,” Skinner intones, “then I reinforce that movement.” Soon, the pigeon shuffles toward the left and—click—Skinner opens the trough.

The pigeon has caught on now. As soon as the trough shuts, she shuffles to the left expectantly.

“You see, the effect is instantaneous,” Skinner says with some pride. The pigeon pauses, and Skinner adds, “I’m waiting for a more pronounced movement than that. It’s got to be more than that.” The pigeon shuffles a bit more. Click. The pigeon hurries back to grab her prize.

The trough closes. The pigeon immediately turns to the left and, with barely a pause, makes a full circle.

“There we go, all the way around,” Skinner announces flatly. The whole training session has taken less than a minute.

Skinner’s video is a stark illustration of reinforcement learning, one of the key processes that shapes animal and human behavior. The core essence of reinforcement learning is straightforward: when a behavior (like shuffling counterclockwise) leads to a reward (like a morsel of food) then it is reinforced and is more likely to be repeated.

Reinforcement learning is ubiquitous. Anyone with a pet has employed reinforcement learning. Dogs learn to sit and stay because we give them treats when they do what we say. Cats learn to stop scratching the couch when we spray them with a water bottle. If you don’t have a pet yourself, an adorable way to see reinforcement learning in action is to search YouTube for videos like “How to train a pig” or “Teaching pigs to sit.” Then again, you’ve already seen reinforcement learning in action since you—and everyone around you—learns via reinforcement. Kids learn to use the potty in pursuit of gummy bears, and they learn addition and subtraction in pursuit of gold stars. Kids and adults both learn when they’ve told a good joke based on whether they’re rewarded with a hearty laugh or nervous silence. And whether an outfit should be worn again or taken to the thrift shop based on whether it is complimented.

Reinforcement learning is powerful. Kids don’t just learn simple arithmetic through reinforcement, they also learn long division, algebra, geometry, and precalculus. YouTube’s pigs have been trained to complete obstacle courses, to score goals, or to ring a bell when they need to be let out to go potty, all through reinforcement. Skinner, famously, taught his pigeons to play ping-pong; in the mid-1990s, a team of Japanese researchers one-upped him by training their pigeons to accurately distinguish Picassos from Monets.

Although reinforcement learning can help animals end up on YouTube, this is obviously not what it’s there for. It’s there so that they learn functional behaviors that are critical for survival in a changing environment. Through reinforcement, animals learn where to find food, shelter, and mates, how to avoid predators, and which foods are poisonous or nutritious.2


The Yasawa Islands are a beautiful, remote volcanic archipelago in the Pacific. Although technically part of Fiji, for the better part of the twentieth century, the islands were ruled autonomously, and their king did not allow tourism on the islands. When the anthropologist Joe Henrich and his PhD student James Broesch visited Yasawa in the mid-2000s, here’s what they found:3

Economically, Yasawans rely primarily on horticulture, fishing and littoral gathering. Fishing is the most important source of protein, and spear-fishing is the most productive form of fishing for those with sufficient skill. People also fish with lines and nets. Yams and cassava provide the caloric staples, although yams are preferred, traditional and necessary for ceremony life. Men compete informally to grow the largest yams. Political units are composed of interrelated clans called Yavusa, which are governed by a council of elders and a hereditary chief. Social life is organized by a complex web of kinship relations and obligations. At the time of the study there were no cars, TVs, markets or public utilities in these villages.

In Yasawa, Henrich and Broesch set out to understand how the Yasawans learned the skills necessary for survival: to fish, grow yams and cassavas, and use medicinal herbs. They surveyed the local population, asking questions like: Who would you go to for advice if you had a question about fish or fishing? Who would you seek advice from about planting or growing yams? Who would you ask about using which plants to use for medicine? They also asked questions like: Who are the best line fishers? Who are the best yam growers? Who knows the most about medicinal plants? They jotted down the names people mentioned and collected data about those individuals. How old were they? What was their sex? Were they in the same village? The same household? Their findings are straightforward. By far, Yasawans tended to ask advice from the individuals who were the best at fishing, growing yams, and using medicinal plants.


On Sale
Apr 5, 2022
Page Count
368 pages
Basic Books

Erez Yoeli

About the Author

Moshe Hoffman is a research scientist at the Max Planck Institute for Evolutionary Biology, a research fellow at MIT’s Sloan School of Management, and a lecturer at Harvard's department of economics. His research focuses on using game theory, models of learning and evolution, and experimental methods to decipher the motives that shape our social behavior, preferences, and ideologies. He lives in Lubeck, Germany.
Erez Yoeli is a research scientist at MIT's Sloan School of Management, the director of MIT’s Applied Cooperation Team (ACT), and a lecturer at Harvard’s department of economics. His research focuses on altruism: understanding how it works and how to promote it. Yoeli collaborates with governments, nonprofits, and companies to apply the lessons of this research towards addressing real-world challenges. He lives in Cambridge, Massachusetts.

Learn more about this author

Moshe Hoffman

About the Author

Moshe Hoffman is a research scientist at the Max Planck Institute for Evolutionary Biology, a research fellow at MIT’s Sloan School of Management, and a lecturer at Harvard’s department of economics. His research focuses on using game theory, models of learning and evolution, and experimental methods to decipher the motives that shape our social behavior, preferences, and ideologies. He lives in Lubeck, Germany.

Learn more about this author