Finance experts share what they learned from poker, chess, puzzles and more.
Markets are deeply tied up with puzzles, wagers and games. Ask a hedge fund manager and a blackjack player how much money to put on a bet, and there’s a good chance they’ll both lean on a mathematical formula called the Kelly criterion. Ask a stock analyst if Nvidia Corp. is a buy, and she might remind you of economist John Maynard Keynes’ description of the market as a beauty contest where the judges have to guess whom the other judges will like best. Games can be tools for understanding what action to take amid complexity and uncertainty.
As the stories that follow show, many investment pros are passionate game players and puzzle solvers. What they get from their pastimes varies widely. Poker teaches risk management and emotional regulation. Crosswords are exercises in pattern recognition and understanding how other people think. Sports, too, can be a mental training ground.
Finance is not actually a game. It’s pensions, clients’ savings and transactions with counterparties who may forgive you for outsmarting them but not for tricking them. Gamesmanship can cause traders to lose sight of the real world behind their screens. (Recall how FTX fraudster Sam Bankman-Fried stayed glued to video games.) But for as long as investing means taking chances, a good gamer is likely to have an edge. —Pat Regnier
What Pro Poker Taught a Trader
By Isabelle Lee and Lu Wang
IN MORE THAN A DECADE as a professional poker player, Vanessa Selbst learned a valuable skill: how to make peace with her own ignorance. When playing a hand, you know your cards, the bets on the table and not much else. What your opponents hold, how much they’ll risk, whether they think you’re an idiot or not—all that has to be figured out. That makes the game as much a test of psychological composure as it is a problem in math and probabilities.
Selbst, 41, is now an options trader at Jane Street Group, and she says acting with limited information is key to both fields. One mistake poker players tend to make is shying away from a bet until they’re sure they’ll win. While risk aversion is sensible, she says successful players bet whenever the odds are above 50% and size their wagers accordingly. “Many people are uncomfortable with that: ‘What if they call me, and I lose?’” says Selbst, once the highest-earning female professional poker player in the world. “That’s going to happen. If that never happens, you are not betting enough.”
Audacity of this sort doesn’t come easily. In poker, it requires millions of hands to learn. “A lot of a poker-player skill set is understanding trends and getting a feel for an intuition,” Selbst says. “It’s the same in markets. You see certain traits, and you understand what they are, what they mean, how often they’re likely to be something that you want to take the other side of or not.”
For instance, when someone wants to buy an options contract for 50¢ over fair value, she asks: Under what conditions does the contract go to, say, $10, and what are the chances of that happening? That analysis is often done with the help of algorithms before the order is filled. Selbst advises traders to practice introspection so they can excel. “A lot of people only question the hands that they lose. But questioning the hands you win—if you win them poorly—can be just as important,” she says.
Over the course of her card-playing career, Selbst captured three World Series of Poker bracelets and racked up roughly $12 million in tournament winnings. In 2014 she became the first woman to top the Global Poker Index, a ranking of the game’s best players. She attributes part of her initial passion to the engagement on internet forums of the early 2000s, when players dissected and debated strategies in a uniquely collaborative spirit. As the game turned increasingly competitive, she found her interest waning.
When she retired from pro poker in 2017, the Brooklyn, New York, native landed her first finance job, at hedge fund Bridgewater Associates LP, through a poker buddy. She joined Jane Street a few years later, and the trading firm’s emphasis on teamwork and process over outcomes reminded her of the online poker forums. “Coming to a place like Jane Street was really invigorating, because it was like a return to what that era was for me,” Selbst says.
After Hours, We Stalk Monsters
By Alfred Cang
ON A MUGGY SINGAPORE evening, 12 of us shuffled into the apartment and took our seats in a circle, like students on the first day of school. By daylight we were chief executive officers, heads of trading desks, risk managers, an intern and a journalist. That night we checked our titles at the door. All that mattered was finding the werewolves before they killed the rest of us.
Here the party game is known as Langren Sha, which is Mandarin for “Werewolf.” Everyone is assigned a secret role, and the people playing the killers must trick the others to stay alive. In our version the cast also includes hapless villagers, plus four roles with unique powers—seer, witch, hunter, idiot. In each round the lights go out, and the hidden beasts claim a victim and notify the referee. Then everyone has to talk things over and pick someone to hold responsible. The game ends when the group successfully “kills”—or votes out—the werewolves, or when they mistakenly execute too many of their fellow innocents.
Some traders say the game, which has many variants and is also called Mafia, has some obvious parallels with their work. In a market with asymmetrical information, you sometimes have to conceal your role while identifying who’s on the same side. The game is also notoriously difficult to model mathematically given the high number of variables. That inherently attracts people who are drawn to puzzles and problem-solving.
The ritual for my group began online during Covid-19, when living rooms turned into offices. But scripted murder games have been a craze in China for the past decade and have recently inspired globally popular TV series such as The Traitors.
For Frankii Lin, a Shanghai-born consultant who organizes games in Singapore, Werewolf’s appeal lies in its constant freshness. “Each round lasts about an hour, and everyone takes on a different role each time—no two games are the same,” he says. He sees it as a training ground for logical thinking, speaking skills and psychological resilience. Lin, 30, has seen people get job offers through Werewolf, as well as a few relationships that blossomed into marriage.
No one talks shop in the apartment where I play. But the game room eventually gives way to the real world. A week later, I keep bumping into my Werewolf teammates at a conference. One trader introduces me to a cluster of peers with a grin: “Alfred is lethal.”
The Crossword Quant
By Justina Lee
HEDGE FUND MANAGER Pete Muller is known for his eclectic range of hobbies. He surfs, is a singer-songwriter, runs a monthly music-themed crossword contest and publishes crosswords in the Washington Post.
“I still oversee the company,” Muller says of his firm, PDT Partners, which invests based on quantitative strategies. “Whenever people ask me things like ‘How much time do you spend on X, Y or Z?’—time is just the wrong dimension. I can spend an insightful five minutes and come up with an amazing solution to a business problem, a song or a puzzle, or I can spend 15 hours banging my head against the wall.”
Muller is a notoriously secretive businessman. He’s coy when asked about his firm; he’ll only say its assets under management exceed $10 billion. But he’s surprisingly open about his hobbies—and that time he beat the North American Scrabble champion. “I got great tiles, and he didn’t get very good tiles at all,” he says in an interview on Zoom from his home in Santa Barbara, California.
As a crossword constructor, Muller says his aim is to wage a “battle with the solver” that he ultimately loses. His monthly music puzzle is what’s known as a meta-crossword, which has the additional element of a riddle that can be solved only once the grid is filled in. One of his recent puzzles spelled “MICHELLE” on black squares that looked like branches on the grid, a nod to Grammy Award-winning pop singer Michelle Branch.
When he’s solving puzzles, Muller will engineer a sufficiently formidable enemy if he can’t find one. For example, he tries to solve the daily New York Times crosswords from Monday through Wednesday using only half the clues. He recently started putting his own spin on another Times game, Letter Boxed, where players make words by connecting letters. In Muller’s version you must use all the letters in only two words. He calls it “Think Twice.”
The Card Game for Traders
JANE STREET GROUP, a trading house with almost $7 billion in third-quarter revenue, invented its own card game called Figgie. Players dissect strategies in academic papers and on Reddit. It’s hard for an outsider to get their hands on a set of game cards, but a mobile app made its debut last year where anyone can play.
The game requires four or five people who each begin with chips representing $350. At the start of a round, everyone antes up an equal share to form a $200 pot. They deal 40 cards and then rapidly swap them for four minutes to maximize the value of their hands. The game uses a specialized deck of cards marked with standard suits but not numbers. Two suits have 10 cards each, one has 8, and one has 12. One way to score is to identify and acquire cards in the “goal suit,” which has the same color as the 12-card suit. For example, if the 12-card suit is hearts, then the goal suit is diamonds.
Players also make money by selling cards—goal suit or not—to their opponents. At the end of the round, each goal suit card is worth $10 from the pot, and the person with the most takes the remainder. Other suits are worth nothing.
Like in a live market, there’s no playing order, just shouting out what you’re looking to buy or sell and for how much. The goal is to get the highest score over multiple rounds. Players get a fresh deck for every round, and each deck has a different combination of cards.
Jane Street senior trader David Vincent, a baritone-singing math competition whiz who joined the quant firm straight out of Massachusetts Institute of Technology, created Figgie in 2013. It was named for a stock that traded at the time.
“Figgie is most importantly a game of action, and making quick decisions that are ‘good enough’ is far better than thinking for a long time to get the perfect answer,” says Ross Rheingans-Yoo, who helped run Figgie-themed events for college recruitment during his more than five years as a Jane Street trader and educator. He’s now an investor and freelance biotech consultant who also teaches trading boot camps. New Figgie players can grasp the basic strategies within hours, Rheingans-Yoo says. Poker, on the other hand, takes months of practice to get good at intense mental math and bluffing skills.
Jane Street has included Figgie in the curriculum for summer internships and new-hire training, and it’s also popular after-hours entertainment. Victory in the annual office tournament for interns and employees is coveted for bragging rights, though winners might also get a trophy, hoodie or other novelty prize.
“Figgie is incredibly simple,” says Anthony Ozerov, a doctoral candidate in statistics at the University of California at Berkeley who co-authored a paper on Figgie strategies. “The trading feature makes it a much more direct imitation of real trading than most other games.”
Jane Street has built its fortunes on a willingness to move fast when it spots an opportunity and take risks, qualities Figgie helps traders hone. But those same attributes can also draw controversy. In India, regulators have accused Jane Street entities of using their trading and technology to gain an unfair advantage in the world’s largest options market, booking hefty profits at the expense of ordinary investors. The firm is fighting the claims in court and says it was operating in its legitimate role as a market maker.
This year, Jane Street introduced a new card game on college campuses called MegaGem. The notoriously secretive company declined to share details about the game, but it’s also designed to simulate elements of markets and trading. A pending trademark application in the US references “coins” among the game equipment.
Liar’s Poker Enters the AI Era
By Richard Dewey
THE FIXED-INCOME arbitrage group at Salomon Brothers in the 1980s became famous for betting with each other on the eight-digit serial numbers on currency. The game was called liar’s poker. Each player would hold a dollar bill, and winning required correctly guessing how often a particular digit appeared across all of the players’ hands.
As a game that rewards logic and wits, liar’s poker offered a new way to establish dominance on the trading floor, becoming a cultural signpost of the power shift to quants from jocks on Wall Street. The game spread across Salomon and eventually to other investment banks and trading houses. Initially, traders pulled dollar bills from their wallets to play. But as liar’s poker grew in popularity, clerks made runs to the bank to keep the game going—a mistake, as the bills were often sequential, making it easier to guess other players’ numbers. Before long, computers that priced bonds by day were repurposed to produce random numbers known as Salomon Liar’s Poker Strips, or “Slips,” after the close.
Players would guess numbers based on each others’ bids. A bid of “four sixes,” for instance, meant that the bidder believed there were at least four sixes among all the bills. The next player had two options: Challenge the bid or top it with a higher wager, such as “five sixes” or “four sevens.” The game ended when all players challenged a bid.
Michael Lewis’ first book, Liar’s Poker, thrust the game and Salomon Brothers into popular culture. Lewis says his reporting on the game was based more on observation than experience. “I only played once, at the Dorchester hotel in London,” he says. “My boss fleeced me for something like $250, and I never played again. To this day, usually when I’m at some conference, someone will pull out dollar bills and ask me to play,” he adds.
The game was so central to Salomon’s culture that after Victor Haghani, a star trader on the arbitrage desk, resigned from the firm, he received a call from Chief Executive Officer John Gutfreund. Its purpose: to congratulate him on winning a five-figure payout in his final liar’s poker hand at Salomon. Haghani became a co-founding partner, along with legendary Salomon trader and liar’s poker player John Meriwether, of Long-Term Capital Management, the hedge fund that blew up so badly in 1998 that the Federal Reserve intervened to prevent a potential contagion by getting banks to recapitalize it. Later Haghani founded wealth adviser Elm Wealth and wrote about the lessons of that episode—especially the importance of not risking too much of one’s personal wealth. I worked for Haghani for five years at Elm Wealth and played liar’s poker with him and his friends a few times.
After artificial intelligence systems conquered Go and Texas No-Limit Hold ’em, I began to wonder how they would fare at the game. I enlisted my former professor Ciamac Moallemi and a few friends to build, with advice from Google DeepMind’s Marc Lanctot, a liar’s poker AI program we dubbed Solly. Earlier this year, Solly took on some of the original Salomon crew and others, playing hundreds of hands. The algorithm and experiments are detailed in a paper on arxiv.org, “Outbidding and Outbluffing Elite Humans: Mastering Liar’s Poker via Self-Play and Reinforcement Learning,” and we plan to make the AI agent available for online play.
Google DeepMind founder Demis Hassabis recently observed how games are a safe way to learn and improve decision-making. The Salomon old-timers agree. “We didn’t know about biases like herding and anchoring in the 1980s,” Haghani says. “But the best liar’s poker players grasped those ideas and how their actions influenced their opponents. Those ideas are central to understanding how markets work.”
Dewey is a researcher, investor and founder. He directed the documentary Radical Wolfe, about the journalist and novelist Tom Wolfe.
The Mental Game of Tennis
By Betty Hou
BEFORE MATT CHENG was writing checks to startups in Silicon Valley, he was a teenager winning international competitions on the tennis court. The 48-year-old Taiwanese investor credits his years as a junior player for teaching him the skills he needed to be successful in finance. Cheng’s Taipei-based venture capital firm, Cherubic Ventures, was an early backer of unicorns—startups valued over $1 billion—such as logistics company Flexport, telehealth firm Hims & Hers Health and meditation app maker Calm.
The first lesson tennis taught Cheng was learning to play solo. It’s a physical game, but as with other sports, you’re also solving mental puzzles. Players must analyze situations, develop solutions and make decisions in split seconds on the court. Timeouts are allowed only for injury, and players get just a couple minutes to hydrate and towel off during set breaks and changeovers.
Cherubic Ventures, with $462 million in assets under management, operates under a rare single-general-partner structure, with Cheng acting as the sole decision-maker for the company’s portfolio investments. “There are no teammates to share the pressure, no coach to call timeouts—it’s just you on the court, trying to win the match,” he says.
The second lesson is what Cheng calls the “best out of five” strategy. In a Grand Slam men’s singles tennis tournament, a player wins the match by beating his opponent in three out of five sets. To win a set, he needs to outplay his counterpart by at least two games. In other words, you don’t need to win every rally. “I can’t be shortsighted and let a bad play get into my head when I’m on the court,” Cheng says, reflecting on grueling matches that sometimes lasted nearly five hours. “I always need to focus every second on the next swing and not the point that I had lost.”
Cheng was once the No. 1 junior tennis player in Taiwan for singles and doubles. Juniors typically play best-of-three sets. During one critical match in the 1991 Junior Davis Cup, Cheng fell behind early, where he lost the first set, then found himself trailing in the second. His forehand wasn’t firing as it should, but he remained patient. He analyzed his opponent’s weaknesses and gradually turned the match around with strong serves, precise drop shots and aggressive net play.
That was the strategy Cheng adopted when he provided angel funding to Hims & Hers Inc. in 2017. Co-founder Andrew Dudum had a prior venture that wound down, and many of his previous backers stayed on the sideline when he sought funding for Hims.
But it takes three sets to win in a best of five,” says Cheng, so he backed Dudum again. Hims later became one of Cheng’s most profitable investments and had a market value of roughly $11 billion as of late October—more than 500 times the valuation at which Cheng first came in.
These days, Cheng shuttles between Japan, the US and Taiwan to meet founders, visit companies and scout for new deals. He doesn’t have much time to play anymore, but his company runs a tennis academy with more than 1,000 students including children and adults.
What Jeopardy! Taught Me About Risk
By Kerry Benn
MY REIGN ON Jeopardy! lasted two days, or about 60 minutes of filming. Since then, I’ve thought a lot about how players manage risk.
As an editor at Bloomberg and a lifelong fan of the TV show, I’ve always had a head for facts and figures. When I was a kid, I’d read the encyclopedia when I was bored. I got my chance on Jeopardy! in 2017 after taking an online test and doing an in-person audition.
The game features clues that are each assigned a dollar value. The first player to buzz in with the correct response in the form of a question earns the amount listed. Wrong guesses subtract from your score. There are also three Daily Doubles—so named because the player who selects the clue can double their score—and a Final Jeopardy! round. All involve wagering all or part of your total. At the end, the winner gets to keep their entire haul (on average about $12,000); second place takes $3,000; third place is worth $2,000.
Wagering requires tricky mental math. You have to decide what you can risk, balancing your knowledge of the category with the need to end up on top to take home a more hefty sum (and play again the next day). And players are doing those calculations under the lights in front of an audience, all while trying to remember the capital of Liechtenstein (What is Vaduz?) or the alliteratively named former NFL wide receiver who won Dancing With the Stars (Who is Donald Driver?).
Benn with former Jeopardy! host Alex Trebek Source: Kerry Benn
The single-game record is $131,127, set by James Holzhauer in 2019, and it strikes me as no coincidence that he’s a professional sports gambler. Most contestants don’t have such a high appetite for risk—they’re regular people who flew to Los Angeles on their own dime. They’re inclined to protect winnings as if they’re real dollars, even when they’d be better off considering them points or play money.
Holzhauer, however, routinely went all-in on Daily Doubles, eliciting gasps from a crowd that has long been used to more conservative bets. His success with this strategy over a 32-game winning streak changed Jeopardy! for players in the years since. If one player in a match is willing to take chances, it behooves others to do the same.
Holzhauer and other players act a lot like some Wall Street pros. If they’re too cautious while others are making money, they’ll be out of a job. Since they’re betting other people’s cash, they can always try something else if a bubble bursts.
As for me, I won a little more than 10% of that record total, easily the most money I’ve ever made for less than an hour’s worth of work. I took my husband on a lovely two-week trip to Central Europe—after paying about 50% of my winnings in taxes, of course.
Solving the Subway Problem
By Hema Parmar
CHRIS SOLARZ, a former pension investing consultant-turned-cryptocurrency investor, once ran for 44 hours straight across three mountain summits in the Canadian Rockies. He’s completed more than 300 marathons and broken nine Guinness World Records—including fastest swim around Staten Island (it took 14 hours). But some of his feats were a matter of calculation as much as endurance, such as most pubs visited in 24 hours by a team (250) or quickest navigation of all 468 New York City subway stops (22 hours and 52 minutes).
“They are forms of the traveling salesman problem,” says Solarz, 47. The classic version asks how to guide a salesman through a list of cities along the shortest route. It’s used for thinking about puzzles in many tasks, including computer programming and engineering.
As a child, Solarz was “mildly obsessed” with The Guinness Book of Records. Years later, when he was 30, the subway task would be the first manifestation of his childhood dream. Solarz and a friend who also worked in finance, Matthew Ferrisi, spent four to five months obsessing over every detail. The subway is open 24 hours: On which of the 1,440 minutes in the day is it most optimal to start? At which station and going in which direction, with which route? They downloaded train schedules and built a model on computer-programming app MATLAB to run through trillions of solutions and solve for the shortest route. They also did two test runs and noted which subway cars to ride for minimizing running distance between stations.
The duo conceived of the math themselves and finished two minutes later than their predicted time, with a single bathroom break. But they beat the previous record holders by more than two hours. The record they set in 2009 held for almost five years until a group of six people shaved off an additional 26 minutes. The time has been lowered on a few more occasions since then, and the challenge now includes 472 stations.
For more than a decade, Solarz was a global macro specialist at Cliffwater LLC, an investment adviser and fund manager that specializes in alternative assets. There he advised some of the US’s biggest pensions on investing in hedge funds. In 2022 he made the switch to cryptocurrency, and he’s now chief investment officer of digital assets at Amitis Capital, a fund of funds, where he selects which crypto funds to back.
Solarz sees a similarity between the sheer effort of the subway challenge and his finance job. Over the past 20 years, he’s reviewed about 5,000 fund managers. “I’ve always had this internal motivation to do exciting, challenging things that seem intimidating at first,” he says. “If it were purely about a recall game or pure intellect, I wouldn’t be able to be at the top. But if I can add in this extra little superpower I have, which is just being the most diligent over the longest amount of time, then maybe I have a chance to compete.”
A Grandmaster of Economics
By Enda Curran
INTERNATIONAL NEGOTIATIONS require patient, strategic thinking with clear insight into the mind of the party across the table. One game excels at perfecting those skills: chess.
Harvard University economist Kenneth Rogoff would know. His career has included roles at institutions that steer the global economy, including the International Monetary Fund and the Federal Reserve. He’s also a chess grandmaster who’s represented the US in the world championships.
Rogoff first learned the game at age 6 from his father growing up in Rochester, New York, then became more immersed when he received a chess set for his 13th birthday. He struggled over the decision to give up chess for a career in research economics. “I realized that both professions demand total focus,” says Rogoff, 72. “I think what tipped the balance, perhaps, was that I thought I could do something more important with my life in economics.”
Rogoff is known for his research on debt crises, central bank independence and exchange rates. His work is a go-to for both policymakers and Wall Street strategists. The lessons he learned from the board game have never been far from his mind. “When you’re playing chess, you almost view it as fighting for your life, even though, of course, you’re not,” he says. “But you’re in this intense, competitive situation, and when you learn to control your nerves in that situation, it can translate into many other things.”
Betting games teach players to act fast whenever they see an edge. Chess, by contrast, encourages cool reflection at key moments. As chief economist of the IMF, Rogoff needed that discipline to make policy decisions with incomplete information. The game teaches you how to think methodically and avoid impulsive actions.
There’s a balance to strike. Rogoff recalls negotiations with Argentina for an IMF bailout he opposed in 2001. “When the clock is ticking and you have to make a move, sometimes in chess you need to rely on intuition and analogies,” he says.
Yet the game also trained Rogoff to pause and review before making a final decision. “That’s a good discipline, and chess really forces you to do that, because if you don’t, you just end up making mistakes,” he says.