2003, Moneyball
Michael Lewis
It was a book about how the American baseball team the Oakland Athletics tried to find a better way to evaluate players and strategy. Oakland had less money than other teams to spend on player salaries, so it had no choice but to rethink the game. The club's management discovered new baseball knowledge in new data, old data, and in analyses done by outsiders, and because of that they were able to run the club far better than other teams. They found value in players other teams had released or ignored, and realized that much of the conventional wisdom of baseball was nonsense.
The project about thinking
Michael Lewis
Over the last decade or so, many people took the Oakland Athletics as a role model and tried to find inefficiencies in the market through better data and better analysis. Up to now I have read countless so-called Moneyball articles in fields ranging from education, filmmaking, elderly medical insurance, golf, farming, publishing, presidential campaigns, government, and banking executives. In 2012, the offensive line coach of the New York Jets complained, "Suddenly everybody is trying to Moneyball offensive linemen?" Comedian John Oliver, watching the North Carolina legislature push a law that maliciously used data to make voting harder for Black people, congratulated the lawmakers for carrying out "Moneyball racism."
And yet the passion to replace old expertise with new-style data analysis often remained shallow. When a risky decision based on data failed to produce immediate success, and sometimes even when it did, it could still draw attacks that would never have come if the decision had been made in the old way.
In 2016, people announced that they would move away from the data-based approach and return to the old method of relying on the judgment of baseball experts. Owner John Berry said it seemed they had perhaps depended too much on numbers. Writer Nate Silver used the statistics he had learned while writing about baseball to predict election results in The New York Times, producing astonishingly successful predictions for years. It may have been the first time a newspaper had surpassed other media in election forecasting. But then Silver left The New York Times, failed to predict the rise of Donald Trump, and unbelievably it was The New York Times itself that began questioning his data-based style of election forecasting. "Politics is fundamentally a human endeavor, and therefore boots-on-the-ground reporting cannot be defeated by prediction and logical judgment, which are impossible here."
That was what a New York Times columnist wrote in late spring 2016. Let us set aside, for now, the fact that almost none of the reporters on the ground predicted Trump's rise either, or that, as Silver himself later admitted, Trump was so unusual that exceptionally large amounts of subjective judgment were inevitably built into any forecast.
Among the criticisms directed at people who try to find knowledge through data, or who exploit inefficiencies in their own industries for their own gain, there are certainly some valid points. But whatever it was that the Oakland Athletics exploited in the human soul for their own advantage, the thirst for experts who offer certainty has a way of lingering and roaming around even when certainty is impossible. It is like a monster in a movie that should have died, but somehow survives and makes one final move.
Even after the loud reaction to my book had faded, one reaction in particular survived for an unusually long time. It came from the assessments of two University of Chicago professors.
One was the economist Richard Thaler, and the other was the law professor Cass Sunstein. Their piece in the August 31, 2003 issue of The New Republic was generous and brutal at the same time. Both professors agreed that there was an interesting truth here: because the professional player market was so broken, a poor club like the Oakland Athletics could beat rich clubs simply by taking advantage of market inefficiency. But they said the author of Moneyball seemed not to understand the deeper reason the baseball-player market was run so inefficiently. That inefficiency came from the way the human mind works.
The project about thinking
Psychologists Daniel Kahneman and Amos Tversky
The way some baseball experts misjudge baseball players, and more broadly the way expertise can distort judgment in any field, had already been explained years earlier by the Israeli psychologists Daniel Kahneman and Amos Tversky. In that sense, Moneyball was not a new book at all. It was only a book that introduced in detail a concept that had been discussed for decades, but that people, and especially I, had failed to appreciate properly.
Thaler and Sunstein's review was restrained. Even then, I do not think I had ever heard of Kahneman or Tversky. Kahneman was a man who had even won the Nobel Prize in economics.
To be honest, I had not thought deeply about the psychological side of the Moneyball story. The baseball-player market was full of inefficiency. Why was that? Oakland's management had spoken about the 'biases' that appeared in the market. For example, a player's speed was overvalued because it was easy to see, while a hitter's ability to draw walks was undervalued because the act was easily forgotten. It might even be better for a hitter to do nothing at all. Fat or awkward-looking players were easily undervalued, while handsome and well-built players were easily overvalued. I found those biases interesting, but I did not go further and ask where those biases came from or why people displayed them. What I especially wanted to talk about was the way a market works, or fails to work, when it evaluates people.
But there was another story hidden inside it. It was the story I had neither examined nor told, about the ways the human mind works, or fails to work, when making judgments and decisions. Whether in investing, in evaluating people, or in anything else, how do people make decisions when faced with uncertainty? How do they process the evidence that comes out of a baseball game, an earnings report, a trial, a medical diagnosis, or a group meeting? How is it that even so-called experts think in ways that lead them to make bad judgments, and end up being exploited by people who ignore expertise and rely only on data?
And how were two Israeli psychologists able to say so much about these questions, and almost foresee that decades later a book about American baseball would be written as a consequence? What had gripped those two people in the Middle East so strongly that they sat facing each other trying to figure out how the mind works when people judge a baseball player, an investment, or a presidential candidate? How could a psychologist possibly win the Nobel Prize in economics? As I tried answering those questions, another story emerged that had to be told.
I want to tell that story now.
Improving existing HR methods, and also managing my own profile
Houston Rockets general manager Daryl Morey
Even Leslie Alexander, the only owner bold and inclined enough to hire someone like Morey in 2006, may have found Morey's view of the world, a view framed in probabilities, disappointing. "He would want a definite answer, and all I could say was that there wasn't one."
Morey attended almost every player-selection interview involving millions of dollars. He tried many things so that decisions would be grounded in statistics as much as possible, and to avoid heuristic guesses and biased choices as much as possible.
During the ten years in which Morey used statistical models with the Houston Rockets, the players he selected performed better than the players chosen by roughly three-quarters of NBA teams, when measured against draft order. In other words, the method he used was far more effective than the methods other NBA teams adopted. He could even point to the exact moment he first felt other teams were copying him. It was during the 2012 draft, when the order in which NBA teams selected players was almost identical to the order the Rockets had produced.
34
Whenever any statistical model hit its limits, human judgment had to be brought back into decision-making, whether it helped or not.
Morey began the hardest effort of his life to mix human subjective judgment into his model. It was not simply about building a better model. It was an effort to listen to both the model and the scouts at the same time.
37
When a scout observed a player, they often formed an immediate impression, and then all the other data tended to be organized around that impression. This was the phenomenon called confirmation bias. The human mind is clumsy at noticing what it did not expect in the first place, and excellent at noticing what it already expected. "Confirmation bias spreads quietly. You do not even realize it is happening," Morey said. A scout forms a fixed view of one player, and then naturally arranges the evidence so that it supports that view. "It is a typical phenomenon. This happens here all the time. If they do not like a candidate, they say he has no position that fits him. If they do like him, they call him a multi-player. If they like a player, they compare his physique to a successful player. If they do not like him, they compare him to a failed one." No matter what prejudice is brought into selecting amateur players, and even if that prejudice negatively affects the selector, it is rarely abandoned, because people keep looking in the direction that confirms it.
Beyond that, people who evaluate talent, including Morey, sometimes prefer players who remind them of their own younger selves. If I see someone who reminds me of me, I will naturally find reasons to like him.
39
The greatest trick our minds produce may be making what is inherently uncertain feel certain. When selecting players, we have seen countless times a crystal-clear picture appear in a basketball expert's mind, only for that picture to later turn out to be a mirage.
40
"His athleticism was never going to be pushed aside. But in reality, miserable people like me judged his athleticism as unimpressive. It was probably because he was Asian. I cannot think of any other reason."
Strangely, when evaluating other people, humans at least recognize what they expect, but fail to quickly recognize what they have never seen before.
40
Morey enrolled in an executive course at Harvard Business School and took a class in behavioral economics. 'They told us to write the last two digits of our cellphone number on paper. Then they told us to estimate, as accurately as possible, how many African countries there were among UN member states.' The result showed that people with larger final digits gave much higher estimates.
Realizing a bias did not mean you could overcome it. That fact made Morey uncomfortable.
42
The expected value of a draft pick far exceeded the value they themselves assigned to the player they were willing to give up in return. Their judgment about Kyle Lowry seemed distorted simply because they already owned Kyle Lowry.
Morey came to understand what behavioral economics calls the endowment effect. To fight it, he forced the scouts, and even his own statistical model, to evaluate the draft-pick value of players Houston already possessed.
The following season, before the trade deadline, Morey wrote all the biases that could distort judgment on a board in front of the staff: the endowment effect, confirmation bias, and others. There is also the bias commonly called present bias, the tendency to undervalue the future relative to the present when making decisions. There is also hindsight bias, the tendency to insist after an outcome occurs that you knew it all along. Statistical models did not show these human swings, but by 2012 the model seemed to have reached the limit of what information could do in evaluating player value. "Every year we talk about what to take out and what to put into the model, and every year it becomes a little more disappointing."
50
What changed decision-making not only in professional sports but in other fields as well was an understanding of how the human mind works in uncertain situations. It took quite a while for that way of thinking to seep into society, but now it has naturally mixed into the air we breathe.
People became newly aware of various kinds of systematic mistakes that both individuals and entire markets can make when they do not examine judgment itself. There were reasons basketball experts failed to recognize Jeremy Lin as an NBA player, dismissed Marc Gasol's true value from a single photograph, and failed to see that an Indian player could become the next Shaquille O'Neal. Speaking about people's failure to recognize what is happening inside their own minds, Morey said, "It is like a fish that does not know it is breathing water until someone tells it." And there was, in fact, someone who told them that.
57
From childhood, Dany had shown an abstract interest in people. Why do people think the way they do? Why do people behave the way they do? He had limited chances to meet people directly. He went to school, but avoided private contact with teachers and classmates. So he had no friends. Simply knowing people could put your life in danger. Still, from a certain distance, he observed many interesting kinds of behavior. He thought teachers and bar owners alike would inevitably know he was Jewish.
59
By December the Germans had withdrawn from France, and Dany and his mother, newly free, wandered through Paris looking to see whether anything remained of their old home and property. Dany carried around a notebook titled What I Write of What I Think ("I must have found it unbearable"). While in Paris, he read Pascal in his sister's textbook and, inspired by it, wrote in his notebook. At the time, Germany was attempting one last counterattack to retake France, and Dany and his mother lived in fear that the attempt might succeed. While writing, Dany tried to explain why human beings need religion. He began by quoting Pascal: "Faith is making God plausible in the heart." Then he added his own words: "Exactly. Churches and organs are devices for artificially producing that feeling." He no longer regarded God as someone to pray to. Later, looking back on his life, he remembered the bravado of his childhood, proud of it and embarrassed by it at the same time. He thought his unchildlike writing was deeply tied to his sense of being "a Jew with only a head and no useful body, and someone who would never be able to fit in with other children."
65
At fourteen, Dany already seemed like an intellectual trapped inside the body of a boy. "Dany was always absorbed in some problem. One day he showed me a long piece he had written about himself. It was very strange. Writing was something you did reluctantly when a teacher at school gave you a topic, a nuisance. The very idea of writing at length on a topic simply because it interested you, even though it had nothing to do with class, was extremely impressive. Dany compared the character of an English gentleman to Greek nobility from the age of Hercules." Shamir said that while other children looked at the people around them to find their direction, Dany seemed to look for it in books and inside his own head.
"Dany seemed to be looking for ideals. Role models, maybe."
67
He refused Israel's offer to give him a new identity. He chose instead to define his own place of belonging.
It was difficult to state exactly what that identity was. Dany himself was hard to define precisely. He did not seem like someone who wanted to settle in any particular place. The things he attached himself to seemed loose and temporary. Ruth Ginsburg, who was dating Ariel Ginsburg at the time and later soon married him, said this: "Dany decided very early on that he would not take on responsibility. I felt there was always a need inside him to rationalize his rootlessness. He was the kind of person who did not need roots, who saw life as a series of accidents, something that had unfolded this way by chance but could have unfolded differently. It is choosing the best in a world where God is denied."
In a land full of people starving for belonging, place, and nation, Dany's lack of that urgency stood out all the more.
Dany took a vocational aptitude test at fifteen, and it said psychology would suit him. He was not surprised. He had always felt that in any field he would become a professor, and the questions he held about human beings interested him more than any other. He said, "My interest in psychology was my way of doing philosophy. I wanted to understand the world by understanding the question of why people, especially I, look at the world the way they do. By then, whether God existed had left my list of interests, but why people believed God existed was still very fascinating to me. I was not very interested in right and wrong, but I was extremely interested in anger. Looking back now, what else is that if not a psychologist?"
In Israel, almost everyone is conscripted into the military as soon as they finish high school. But Dany's intellectual talent was recognized, and he was allowed to go straight to university and earn a psychology degree. 69
School 1. Behaviorism
If you cannot directly observe what happens inside the mind, how can you even pretend to study it? The only things worthy of scientific attention, and scientifically studyable, were the ways living creatures behave. The representative school that studied thought in this way was called behaviorism. Skinner began studying the field during World War II after the U.S. Air Force instructed him to train pigeons to guide bombs. Skinner taught pigeons to peck accurately at a target when it appeared on a screen, and rewarded success with food. The pigeons grew passive when anti-aircraft shells exploded around them, so they were never deployed in combat. But once Skinner succeeded in this experiment, his idea that all animal behavior is triggered not by thought and feeling but by external rewards and punishments began to exercise astonishing influence. 71
Behaviorist psychologists assumed that what they found in experiments on rats and pigeons applied equally to humans. At the same time, they thought experimenting directly on humans was unrealistic for various reasons. Skinner wrote in a paper titled How to Teach Animals: "Sometimes one must introduce a program that applies a relevant reinforcement, and sometimes suppresses it. In doing so, one is likely to produce a certain emotion in the human subject. Unfortunately, behavioral science has not yet been as successful in controlling emotion as it has been in controlling behavior." The appeal of behaviorism lay in how clear its scientific method seemed. Stimuli could be observed, and responses could be recorded. It looked, at a glance, 'objective.' Behaviorism did not rely on verbal reports in trying to grasp thought or feeling. Only what could be observed and measured mattered. There is even a joke, one Skinner himself liked, about this seemingly spotless behaviorism. After a couple makes love, one partner says to the other, "That was good for you. How was I?"
The leading behaviorist psychologists of the 1950s were all WASPs, that is, white Protestants. But none of the young people at the time noticed that fact. Looking back on psychology from that era, it almost feels as if there should have been entirely separate traditions of 'WASP psychology' and 'Jewish psychology.' The WASPs roamed brightly through white laboratories in their coats, clipboard in hand, thinking up new ways to torture rats, while steadily avoiding the troublesome problem of human experience. By contrast, the Jews, even while despising Freud's methods, longed for 'objectivity' and were willing to take on that troublesome problem in order to find scientifically verifiable truth.
School 2. Gestalt
This psychology, which began in Berlin in the early twentieth century, was led by German Jews who wanted to explore the mysterious human mind scientifically. Gestalt psychologists discovered fascinating phenomena and demonstrated them with astonishing talent.
For example, light appears brighter when it shines in darkness, gray looks green when surrounded by purple and yellow when surrounded by blue, and when someone says "Don't step on the banana cel!" people are certain they heard "Don't step on the banana peel!"
Gestalt psychologists showed that because the human mind often intervenes in strange ways when responding to outside stimuli, there is no clear one-to-one relation between the stimulus itself and the sensation it produces inside a person.
In particular, Dany was deeply impressed by the way Gestalt psychologists let readers directly experience the strange phenomena taking place inside their own minds. For example, when you look at a clear night sky, some stars immediately stand apart from the rest and appear as one cluster: Cassiopeia, the Big Dipper. 74
Gestalt psychologists thought about the very questions behaviorists had decided to ignore. How does the brain create meaning? How does it turn the fragments collected by the senses into a coherent picture of reality? Why does that picture seem less like the world injected itself into the head and more like the head imposed itself on the world? How do people turn fragments of memory into a coherent life story? Why do people interpret what they see differently depending on the surrounding context?
Dany , like other Jewish psychologists, also tried to pursue objectivity. He found the questioning style of the Gestalt school deeply attractive. He entered psychology driven by questions that arose from phenomena he had lived through himself, such as why, when regimes obsessed with exterminating Jews rose to power in Europe, some Jews recognized their true nature and fled while others remained and were slaughtered.
Dany later said that he thought of science as a kind of conversation. If so, psychology was like a noisy dinner party where people talk about the past and change the subject at any moment. Gestalt psychologists, behaviorists, and psychoanalysts filled a building labeled 'psychology,' yet paid no attention to each other's stories. Psychology was unlike physics, and even unlike economics. There was no single persuasive theory that could systematically organize psychology, and no agreed-upon rules for debate. Major psychologists could look at another psychologist's work and say, in effect, "What you are saying and doing is basically complete nonsense," and in fact they did say such things, but those comments had almost no effect on the other side's behavior.
Israel, meanwhile, remained less a nation than a fortress, and the military was a kind of institutionalized chaos. Soldiers were poorly trained, and units were in terrible shape. In the early 1950s, there was no formal war between Arabs and Jews, but meaningless violence repeated itself regularly, exposing the helplessness of the Israeli army. Soldiers ran the moment trouble seemed likely, and officers did not want to step forward. 80
Since the First World War, the work of drafting and sorting young people into the military had been handed over to psychologists. Some ambitious psychologists had successfully lobbied the U.S. military to win that job. 80
Dany had little difficulty identifying people who would become good officers. "We would declare without hesitation, 'This one will never make it,' 'That one is just so-so,' or 'That one will be a star.'" But then, comparing his predictions with actual results, he found a problem. When the candidates actually entered officer school and underwent training, their performance did not match his forecasts at all. The predictions were garbage. But it was the military, and the work had to be done, so he had no choice but to keep doing it. And being Dany, he noticed that in spite of all this, he still felt confident in what he was doing. To him, the situation resembled the famous Muller-Lyer illusion.
Dany thought the hard part was accurately measuring those traits through an ordinary interview. The subtle difficulty that appears when evaluating other people had already been described in 1915 by the American psychologist Edward Thorndike.
A soldier first judged outstanding overall would then be judged tougher than he really was in a specific area. "The halo of a general virtue affects judgments of specific ability, and conversely the halo of a specific ability affects judgments of general virtue." Thorndike's conclusion was this: "Even an outstanding field supervisor, employer, teacher, or department head cannot perceive a person as a mixture of separate traits and assign scores to each trait independently." That is where the term still used today, the halo effect, came from.
Dany knew about the halo effect. He also understood that it was appearing in Israeli military interviewers at the time. 83
To minimize the halo effect, he taught the military interviewers, mostly women, how to write lists of questions to ask each recruit. The questions had to be highly specific, and designed to uncover how the recruit actually behaved, not how the recruit thought of himself. The questions had to draw out facts while disguising the facts they were trying to extract. Before moving from one section to the next, the interviewer would assign a score from 1 to 5, ranging from 'absolutely does not behave this way' to 'always behaves this way.' 84
Dany's personality test did predict one thing: whether a recruit was likely to do well at almost anything. More surprisingly, the result had very little correlation with intelligence or education. In other words, it provided information that could not be captured by the simple measures of intelligence and education. This scale, also called the 'Kahneman score,' helped make use of the entire population in more useful ways for the military, and especially in selecting military leaders, by reducing the importance of raw measurable intelligence and raising the importance of the qualities listed in Dany's checklist. 86
He later went to Washington, D.C. as a research fellow. One day he received a call from a senior official at the U.S. Department of Defense asking whether he could come and talk. When he arrived, he was bombarded with questions in a room full of American generals. He later described the scene like this: "The wording varied, but it was always the same question. 'You have the same guns, the same tanks, and the same airplanes we do, so how is it that you win battle after battle and we do not? We know it is not a weapons problem. It must be a psychological one. How do you select the soldiers who go into battle?' For five hours all they wanted to know was one thing: the selection procedure."
Dany later told his university students: "When someone says something, do not ask yourself whether it is true. Ask instead in what case it would be true." That was his intellectual instinct, the natural first step into the loops of the mind.
Whatever someone had just said, take it in and try to understand it rather than dismantling it. The question the Israeli army had put to him, which sort of personality best fits which military role, was in fact absurd. So Dany asked a more productive question. How can we stop interviewers from ruining their evaluations of recruits through intuition? He had been asked to divine the personality of Israeli youths. Instead, he learned a new fact about people who try to divine the personality of others: if you abandon intuition, you can make better judgments.
He had been given a concrete problem and discovered a broad truth.
The University of British Columbia psychologist Dale Griffin explained it this way: "What makes Dany different from the other 999,999 psychologists is that when he discovers a phenomenon, he can explain it in a way that allows it to apply in other situations too. It looks like luck, but he does it over and over."
Four years later, in 1961, when he returned to Hebrew University as a young assistant professor, he was stimulated by fresh ideas from personality research being conducted by psychologist Walter Mischel. In the early 1960s, Mischel developed a remarkably simple test for children that revealed many things. It was the marshmallow experiment. The experimenter placed three-, four-, and five-year-old children alone in a room, set something they liked, such as pretzels or marshmallows, beside them, and told them that if they waited a few minutes without eating it, they would receive another one. Children's patience turned out to correlate with IQ, family environment, and other factors. When Mischel later tracked their lives, the children who had resisted temptation better at age five tended to have higher SAT scores and higher self-esteem, and lower chances of addiction and lower body fat.
Personality
He would become consumed by some idea or ambition, rush at it fiercely, and then give it up in deep disappointment. He said, "I always think there are plenty of ideas. If one of them does not work, you should not cling to it desperately. You should move on to the next one."
In an ordinary society, it would be hard to find great practicality in Daniel Kahneman. But Israel was not an ordinary society.
Dany, seized by a new passion, developed many experiments similar to the marshmallow experiment. He even coined a phrase for his own line of research: the psychology of single questions. In one of them, he gathered Israeli children and told them they were going camping, then asked whether they would prefer to sleep in a one-person, two-person, or eight-person tent. He thought their answers might reveal something about their sense of belonging. But he learned nothing. Just when it seemed he had discovered something, the result failed to repeat in follow-up experiments. He gave up on the experiment. "I wanted to be a scientist, and I thought I could not be one unless I could get my findings to repeat. But in the end, I could not repeat them." He doubted himself once again, gave up on personality research, and concluded that he simply had no gift for it.
