Edward Thorp: The First Quant — From Beat the Dealer to Beat the Market

In the spring of 1961, in a basement laboratory in Cambridge, Massachusetts, two grown men in cufflinks and white shirts were trying to get a small computer the size of a cigarette pack to fit inside a shoe. The shoe was an ordinary leather brogue. The computer had twelve transistors, a microswitch hidden in the toe, and a tiny earpiece that delivered musical tones to its operator. The two men were Edward Oakley Thorp, twenty-eight, mathematics professor at MIT, and Claude Shannon, the founder of information theory. The shoe was for cheating at roulette.

The plan, worked out in a quarter-million dollars’ worth of stolen Bell Labs evenings, was straightforward. One operator stood at a roulette wheel and tapped the microswitch in his shoe each time the ball passed a fixed reference point. The computer measured the velocity of the ball and the wheel, calculated which octant of the wheel was most likely to receive the ball, and transmitted that information to a second operator, sitting elsewhere, via the earpiece. The second operator placed bets on the favoured numbers. The expected edge was roughly forty-four percent. They tested the prototype in Las Vegas. It worked. The earpiece wire, made of household speaker wire smaller than a hair, kept breaking in the heat. They went home, fixed it, and stopped before the casinos noticed.

This was the same Edward Thorp who, three years earlier, had become the first person in human history to mathematically prove that the house edge in blackjack could be overcome by a small player armed only with the cards he could see. The same Thorp who, eight years later, would launch the first market-neutral hedge fund in history. The same Thorp who, in 1991, would look at the reported track record of a New York investment manager named Bernard Madoff and conclude, seventeen years before anyone else figured it out, that the entire operation was a Ponzi scheme.

If Benjamin Graham was the constitution of value investing, Edward Thorp is its first hacker. He started by asking a simple question that nobody else dared ask: can the math actually be beaten? Then he proved it could, in casinos. Then he asked the same question of Wall Street. Then he proved it again. He compounded his hedge fund at roughly nineteen percent a year, after fees, for nineteen years, with no losing quarters. He met Warren Buffett in 1968 at a bridge game and told his wife the same evening that the fellow from Omaha was going to be the richest person in the world. He was right about everything except his estimate of how much trouble it was going to bring him.

The Question Nobody Else Asked

Graham worked with the assumption that markets were eventually rational and that the patient analyst could profit from the gap between price and value. Thorp worked with a different assumption. He assumed that any system in which money changed hands probabilistically could, in principle, be reverse-engineered. The casino, the warrant market, the option market, the convertible bond market, even the supposedly random equity market: each of them was a mathematical structure, and structures had weak points, and weak points could be exploited if you were willing to do the work.

This is a small distinction in summary and an enormous one in practice. Graham looked at a balance sheet and asked what the business was worth. Thorp looked at a deck of cards and asked what the dealer’s expected loss would be if the player adjusted his bet size to the running count. The two men were doing fundamentally different things. Graham was building a profession of valuation. Thorp was building a profession of edge detection. Modern quantitative finance descends from Thorp in almost every important respect: statistical arbitrage, convertible arbitrage, delta hedging, the use of computers to find pricing anomalies that humans cannot see, the use of the Kelly criterion to size bets so as to compound capital without ruining yourself. None of this existed as a discipline before Thorp. After Thorp, every successful quant fund in the world has been a refinement of his original move: find a measurable edge, hedge what you cannot predict, size your bets according to the math, repeat.

Chicago, the IBM 704, and the Cards That Were Already Gone

Thorp was born in Chicago on the fourteenth of August, 1932, to working-class parents. His father, a security guard, had read widely and pushed his son to do the same. By seven, Edward could compute the number of seconds in a year in his head. By his teens he was building rockets in the backyard, distilling his own gunpowder, and racking up first-place finishes in California state science fairs. UCLA. Master’s in physics. PhD in mathematics in 1958, at twenty-six. Then a junior appointment at MIT, the temple of mid-century American science.

The cards entered his life by accident. In 1958, on holiday in Las Vegas, his wife Vivian had given him a few dollars to play with at the blackjack tables. He had read a paper by four military mathematicians, the so-called Baldwin paper, which gave a basic strategy for blackjack that reduced the house edge to about half a percent. Thorp played the strategy. He lost a little money, slowly, exactly as the math said he should. While he sat there, something occurred to him that should have occurred to anyone. The deck was not reshuffled after every hand. As cards came out of the shoe, the composition of the remaining deck changed. The probabilities at any given moment were not the probabilities of a fresh deck. They were the probabilities of whatever was left.

If you could track what was left, you could compute how much your edge had moved, hand by hand. If your edge was negative, you bet the minimum and waited. If your edge was positive, you bet more. If your edge was very positive, you bet a lot more. Over thousands of hands, the small adjustments would compound into a real, calculable advantage over the house. The mathematics of this was simple. The execution was not. To execute it you needed to know exactly how much each card removed from the deck shifted the player’s expectation, and to know that you needed to compute, for every possible deck composition, the expected return of optimal play. The combinatorial space was too large for a human to do by hand.

Thorp went to the IBM 704 at MIT, taught himself FORTRAN, and did the calculations. The computer ran for hours. Out of it came a system that, in a single deck game where the dealer hit on soft seventeen, gave the player an edge of roughly half a percent over the house, swinging higher when the deck became rich in tens and aces. He called it the Ten Count. The casino edge had been overcome for the first time in recorded history.

He wrote up the result and submitted it to the American Mathematical Society conference in 1961. The mathematicians read it and shrugged. Thorp expected an audience of his peers in a quiet seminar room. He arrived to find reporters, gamblers, and a queue of strangers wanting to know more.

Manny Kimmel, the $10,000 Bankroll, and the Disguises

Thorp needed capital to test his system in the field. A salary of a few thousand dollars a year as a junior MIT professor was not going to survive a casino weekend, even a winning one. Manny Kimmel, a wealthy New York gambler with a colourful past in numbers running and labour racketeering, agreed to back him with ten thousand dollars. The two of them went to Reno first, then Lake Tahoe, then Las Vegas. Thorp won eleven thousand dollars in a single weekend. The system worked exactly as the IBM 704 had predicted. So did Kimmel’s instinct that this was going to attract attention.

It did. Thorp’s Beat the Dealer, published in 1962, sold seven hundred thousand copies and remained on the New York Times bestseller list for years. It was the first book in history to give the public a mathematically proven winning strategy against a casino game. Kimmel, lightly disguised in the text as Mr. X, was furious. Casinos were furious. The Las Vegas authorities tried, briefly, to change the rules of blackjack in 1964. Regular customers boycotted the new rules. The casinos backed down, and instead introduced the multi-deck shoe, which made counting harder but did not make it impossible.

What happened next reads like a Cold War spy novel and is almost entirely true. Thorp was barred from casino after casino. He grew a beard and got coloured contact lenses. He wore wraparound glasses. He played in a striped shirt one night and a black turtleneck the next. He kept a notebook of his disguises so as not to repeat them at venues that had recently barred him. Casinos sent young women dealers to his tables to flirt with him in the hope of distracting his count. They watered his drinks. On one occasion at the Dunes in Las Vegas, while playing baccarat with a team in 1963, he became disoriented in a way he could not explain by alcohol because he had not been drinking, and concluded later that something had been added to his glass. On another occasion in Las Vegas, returning to a parking garage in his rental car, the brakes failed at low speed in a way that an examination later suggested was not random.

None of this stopped him. He kept playing. He moved to baccarat and developed a counting system there. He played on roulette wheels with Shannon’s wearable computer in his shoe. He compiled enough data to publish three editions of the book. By the late sixties he was a celebrity in a field whose practitioners did not generally seek celebrity. And then he stopped playing entirely. The casinos had won the practical arms race even if they had lost the mathematical one. There were better games elsewhere.

The Real Casino: How Beat the Market Predated Black-Scholes

Thorp’s transition from blackjack to Wall Street was not a career change. It was the same activity applied to a larger casino. In the mid 1960s he became interested in stock warrants, which were essentially long-dated call options issued by companies as sweeteners on bond financings. Warrants were notoriously hard to value. Most analysts treated them as a kind of speculative side bet. Thorp, working with an economist named Sheen Kassouf at UC Irvine, set out to build a proper model.

What they produced, published in 1967 as Beat the Market, was a delta-hedging framework that anticipated almost all the key insights of the Black-Scholes options pricing formula, which Fischer Black, Myron Scholes, and Robert Merton would publish six years later in 1973. Thorp’s version did not include the risk-free rate explicitly, because the bond market in the 1960s did not lend itself to a clean risk-neutral derivation in the way it would after the introduction of listed options exchanges. But the core move was there. If you could write down a continuous mathematical relationship between an option’s price and the underlying stock’s price, you could construct a hedged portfolio that was exposed neither to the stock going up nor to the stock going down. You were exposed only to the option being mispriced relative to the model. If the option was too expensive, you sold it short and bought the stock. If it was too cheap, you bought it and sold the stock short. Either way, you collected the convergence as the option moved towards fair value.

Thorp implemented this in his own brokerage account starting in 1965. He made roughly twenty-five percent a year for several years, in a market that was going essentially nowhere. The dean of UCI’s graduate division noticed and asked Thorp to manage some of his money. So did the chancellor’s secretary. So did several other faculty members. By 1968, Thorp was running what amounted to a small unofficial investment partnership. The dean, in a separate transaction, was a limited partner in another small investment vehicle out in Omaha, run by a man named Warren Buffett. Buffett was about to wind that vehicle down and return the capital because he could no longer find anything cheap. The dean asked Buffett where to put his money next. Buffett asked to meet Thorp first.

The Bridge Game with Buffett

The introduction took place at the dean’s house in Newport Beach, California, in the summer of 1968. Buffett and his wife Susie were in Southern California for the wedding of friends. They came over for dinner and a few rubbers of bridge. Buffett, a serious bridge player, played carefully but warmly. Thorp, also a serious bridge player, watched the way Buffett bid and discarded and recognised, immediately, what Charlie Munger would later describe in Buffett: a mind that compounded information faster than the room around him.

The two men talked for hours about probability, about non-transitive dice, about a particular bridge convention. They did not actually talk all that much about stocks. Buffett did not pump Thorp for trade ideas. Thorp did not press Buffett on Berkshire’s positions. They were two probabilists comparing their tools. At one point Thorp mentioned the Kelly criterion. Buffett knew it already. He did not use it explicitly but he had been sizing his concentrated bets in ways that approximated it for over a decade.

That night, driving home with Vivian, Thorp told his wife that Warren Buffett was going to be the richest person in the world someday. Buffett at the time was managing a partnership of perhaps sixty-five million dollars. Berkshire Hathaway, the textile mill he had taken over and was using as a holding company, was trading near book value at perhaps fifteen dollars a share. Thorp’s prediction was not based on any traditional valuation work. It was based on what he had seen at the bridge table: an unusually clear mind, a temperamentally honest accounting of risk, and a sufficiently young age that the compounding could run for forty or fifty years. Vivian, who had heard a lot of confident pronouncements about people from her husband over the years, took note.

Buffett told the dean that Thorp was sound. The dean and several others gave Thorp their money. By the following year, in 1969, Thorp and a New Yorker named Jay Regan had launched what would become the first market-neutral hedge fund in history, Convertible Hedge Associates, later renamed Princeton/Newport Partners.

“The more I learned about gambling and investing, the more I realised that they were two ends of the same continuum. The casino is a small market with simple rules and short time horizons. The stock market is a large market with complicated rules and long time horizons. The math is the same. Only the noise is different.”
— Edward O. Thorp, A Man for All Markets (2017)

Princeton-Newport: Twenty Years Without a Losing Quarter

Convertible Hedge Associates, renamed Princeton/Newport Partners in 1974 when Regan moved the trading desk to Princeton, ran for nineteen years. From its launch in November 1969 to its forced wind-down in late 1988, it produced an annualised return of approximately nineteen percent after fees, depending on which precise calculation you use, with no losing quarters across roughly seventy-six consecutive quarters. The S&P, over the same period, returned roughly ten percent annualised. PNP did this at lower volatility than the index. The Sharpe ratios it generated were unprecedented at the time and remain extraordinary today.

The strategy was a refinement of the Beat the Market framework, scaled up. Convertible bonds and warrants, hedged against the underlying stock. Mispriced options, hedged against the underlying stock. Index arbitrage, hedged against the cash market. By the mid eighties the firm had moved into early statistical arbitrage, an entire portfolio of small mean-reverting trades across the equity universe, executed by computer in volumes that would have been physically impossible by hand. The firm employed dozens of people in its Princeton trading room and its Newport Beach research room. Thorp ran the research and the math. Regan ran the trading.

And here, in 1987, the trouble began. Federal prosecutors led by Rudy Giuliani, then U.S. Attorney for the Southern District of New York, began an aggressive investigation into Michael Milken’s junk bond operation at Drexel Burnham Lambert. PNP had done some perfectly legal trades with Drexel several years earlier. Giuliani’s office expanded the investigation to include PNP’s Princeton office. In 1988 FBI agents raided the Princeton trading room. Five PNP employees, including Regan, were charged with stock parking and tax-related offences related to the Drexel trades. Thorp himself, in California, was never charged. The convictions of the Princeton employees were eventually overturned on appeal in 1991, and the cases were quietly dropped. But by then the damage was done. The fund’s investors had been spooked. Legal costs had run into the millions. Thorp shut PNP down at the end of 1988, returned all capital to investors, and walked away.

It is worth pausing on this. PNP had compounded at nineteen percent for nineteen years without a losing quarter. It had been killed not by the markets but by a regulatory sweep that had targeted Milken and incidentally caught the firm in the dragnet. Thorp, who had done nothing wrong personally and was never accused of anything, lost the vehicle of his life’s work. The closest historical parallel is the way Long-Term Capital Management would be killed a decade later by markets rather than by prosecutors. PNP was killed by lawyers.

Madoff: The Seventeen-Year Warning

In the spring of 1991, Thorp was hired as a consultant by a major institutional investor (the firm has since been identified in his memoir as McKinsey and Company) to perform due diligence on a portfolio of hedge fund investments. One of the funds in the portfolio was managed by a New York brokerage firm called Bernard L. Madoff Investment Securities. Madoff’s firm was reporting smooth, consistent monthly returns of roughly one to one and a half percent. The strategy was described as a split-strike conversion: long the S&P 100 index, hedged by selling out-of-the-money calls and buying out-of-the-money puts. The reported returns had no losing months and almost no volatility.

Thorp, who had spent twenty years pricing and trading exactly these kinds of options, asked for a sample of the actual trades. He obtained a list of one hundred and sixty trades that Madoff claimed to have executed. Thorp went to his contacts at the options exchanges and pulled the actual exchange data for the days in question. Roughly half of the trades had not occurred at all. Another quarter showed sizes larger than the entire day’s volume on the relevant exchange. The remaining quarter were inside the realm of possibility but did not match the reported prices. Madoff had also claimed to have purchased one hundred and twenty-three call options on Procter and Gamble on the sixteenth of April 1991. Total exchange volume in P&G calls that day, across all strikes and all expirations, was twenty contracts.

Thorp wrote up his findings. He told McKinsey: do not invest. He told the friends and family members who asked him about Madoff over the next seventeen years: do not invest. He told one fund-of-funds manager, repeatedly, that Madoff was a fraud. The fund-of-funds manager continued to allocate to Madoff and continued to raise money for him until the morning of the eleventh of December, two thousand and eight, when Madoff confessed. Thorp’s son Jeff called him that day and said: it’s what you predicted in 1991. Thorp had been right for seventeen years before anyone else figured it out. Harry Markopolos, the Boston quant who would later win fame for his repeated SEC complaints about Madoff, did not begin his own investigation until 1999, eight years after Thorp had already finished his.

The Madoff episode is the cleanest possible illustration of why Thorp matters. Madoff was reporting impossibly smooth returns for fifteen years before he confessed and got away with it because almost no one in the entire investment industry had the combination of options-pricing expertise, willingness to do basic verification work, and personal indifference to the social pressure of disagreeing with a famous, wealthy, well-connected money manager. Thorp had all three. He had been the hacker. He had spent two decades building precisely the toolkit needed to detect a fake hedge fund. When the data did not add up, he said so, and he kept saying so.

Ridgeline, Buffett, and the Long Tail

In 1994, after a few years of managing his own money, Thorp launched Ridgeline Partners, a pure statistical arbitrage fund. From August 1994 to September two thousand and two, Ridgeline returned roughly eighteen percent a year, with low correlation to equity markets and no losing years. By the early two thousands, the strategy was being arbitraged away by the next generation of quant shops, including Renaissance Technologies, D.E. Shaw, and the rapidly growing high-frequency trading firms. Returns compressed. In two thousand and two, Thorp closed Ridgeline. It had, like PNP, done what it was designed to do and stopped doing it as the market evolved.

By that point, Thorp had also been quietly compounding personal money in Berkshire Hathaway since 1982, when he had read an article about Buffett and recognised, fourteen years after the bridge game, that the prediction had been right. Berkshire’s stock had risen from roughly twelve dollars when Thorp first met Buffett to nine hundred and eighty-two dollars by 1982. Most investors would have looked at that eighty-fold rise and concluded they had missed the trade. Thorp looked at it and bought. Berkshire continued to compound at roughly twenty percent a year for the next forty years. The position made him rich on a different timescale than PNP did.

Today, in his nineties, Thorp continues to manage his own capital from Newport Beach, California. He has reported personal annualised returns averaging roughly twenty percent over twenty-eight and a half years through the late nineties, and on a cumulative basis over his entire investment life, returns measured in the tens of thousands of percent. He still corresponds with the small number of practitioners still alive who knew the original generation of quants. He has outlived almost everyone he started with.

What We Cannot Know

Thorp’s record contains genuine uncertainties that an honest profile has to acknowledge.

The first concerns the precise return of Princeton/Newport Partners. The figure of nineteen percent annualised is consistent with the partnership’s audited records and is widely cited. Some reconstructions, including a careful analysis by William Poundstone, place the figure closer to fifteen percent net of all fees and incentive compensation, with the higher numbers reflecting gross returns or returns to certain limited partners. The seventy-six-quarters-without-a-loss figure is robust. The exact compounded return depends on which fee class and which time slice one uses.

The second concerns how much of Thorp’s edge is replicable. His memoirs are extraordinarily honest about the role of obscure technological advantage. In the 1970s and eighties, Thorp had access to early computer systems, options models, and execution capabilities that almost no other market participant had. By the time those advantages had diffused, around the late 1990s, the spreads on which he made his living had compressed materially. The methodology is real. The unique historical window in which it printed money was finite.

The third concerns the closeness of his calls. Thorp predicted that Buffett would be the richest person in the world; Buffett came close but never quite achieved that ranking against Bezos and Musk and others. Thorp identified Madoff as a Ponzi seventeen years early; he could not, and did not, predict precisely when the scheme would collapse. Thorp consistently warned about excess leverage in episodes including Long-Term Capital Management and the two thousand and eight crisis; he did not, however, time those collapses with enough precision to position aggressively against them. His honest framing in his memoir is that detection is much easier than timing, and that the most a careful analyst can usually do is stay out of the wreckage.

The fourth concerns the human cost. Thorp’s life looks, from the outside, like one of the great success stories of twentieth-century finance: long marriage to Vivian, three healthy children, no scandal, no professional disgrace, regular tennis until his mid-eighties, twenty-one completed marathons, and as of this writing he is alive and lucid in his ninety-third year. By his own account, the years of harassment by casino security in the 1960s were genuinely frightening. The federal raid on PNP in 1988 cost him sleep, friends, and the life’s work he had built around Regan. He is one of the rare protagonists in this series for whom the verdict on a good life appears to have come down clearly on the positive side. We acknowledge this with a small note of caution. Many of the cleanest narratives in finance have turned out, on closer examination, to contain quiet costs that the protagonist did not advertise.

What Thorp Teaches: Four Lessons in Order of Depth

1. The most important lesson is structural: every game has a structure, and structures can be measured. Thorp’s first move was always the same. Look at the rules. Compute the actual probabilities, not the assumed ones. Compare the price to the expected value. If the price is wrong, exploit it. If the price is right, walk away. This sounds like advice for a casino. It is also advice for every market a retail trader will ever look at. The implied volatility on an option is a price. The premium on a futures contract is a price. The fee structure on a prop firm challenge is a price. Every one of those prices is a function of underlying mathematics. If you are unwilling or unable to do the math, you are paying the price the structure asks you to pay. If you are willing to do the math, you are sometimes given the gift of finding that the structure is wrong.

2. The deeper lesson is procedural: hedge what you cannot predict, and bet only on what you can. Thorp almost never took a directional view on the equity market. He thought it was hard, the data were noisy, and the timescales were too long for his methods to compound efficiently against them. What he did instead was build positions in which the directional risk was hedged out and what remained was a bet on a specific, narrow, measurable mispricing. Convertible bonds against the underlying stock. Options against the underlying stock. A statistical arbitrage book against itself. The retail equivalent is to ask, before placing any trade, what exactly am I betting on. If the answer is, the market goes up, ask whether you have any actual edge in predicting that. If you do not, do not size the bet as if you do. Hedge what you cannot predict. Concentrate only on what you can.

3. The deeper lesson still is mathematical: bet small enough to survive the variance. The Kelly criterion is the technical answer to the question of how much to risk on a favourable bet. Bet too little, and you fail to compound. Bet too much, and you get ruined by an unlucky sequence even though your edge was real. Thorp used the Kelly criterion at the blackjack table and at PNP and at Ridgeline and to this day in his personal account. Most retail traders do not know the formula and do not need to memorise it, but they do need to internalise its central insight: position sizing is not a separate problem from edge. It is part of the same problem. An edge that is real but is bet too aggressively becomes a losing strategy. Most blown retail accounts are not the result of bad analysis. They are the result of correct analysis bet too large.

4. The deepest lesson is epistemic: do the verification work yourself. When McKinsey asked Thorp to look at Madoff in 1991, he did not read the marketing material and call it a day. He pulled the actual trade tickets, checked them against the actual exchange data, and discovered that the trades had not happened. The verification took perhaps two days. It saved his clients hundreds of millions of dollars. The principle generalises. When a backtest looks too good, look at the underlying trades. When a strategy looks like it has no losing months, ask why. When a guru posts smooth equity curves, ask for the broker statements. Do not rely on social proof, on credentials, on the size of someone else’s portfolio, or on the polish of their marketing. The math either checks out or it does not. The verification is the discipline. Most fraud, and most flat-out wrong investment thinking, dies on contact with five minutes of honest checking.

Frequently Asked Questions

Was Edward Thorp really the first hedge fund manager to use a market-neutral strategy?

By his own claim and by the consensus of historians of the industry, yes. Convertible Hedge Associates, launched in November 1969, was the first fund whose entire architecture was explicitly designed to remove broad market exposure and isolate specific mispricings. Earlier funds, including A.W. Jones from 1949, had used long-short structures, but Jones’s fund was not strictly market neutral and used hedging primarily to permit greater long exposure. Thorp’s structure was the first true market-neutral derivatives-arbitrage hedge fund.

Did Thorp really predict Buffett would be the richest man in the world?

According to multiple interviews with Thorp and his memoir, yes. After their bridge game in 1968, Thorp told his wife Vivian that Buffett’s combination of intelligence, compounding rate, and scalability made it likely he would eventually become the wealthiest person on earth. Buffett did become one of the three or four richest people in the world for extended periods, peaking near the top before being overtaken by tech founders. The directional prediction was correct.

How did Thorp catch Madoff?

In 1991, McKinsey asked him to perform due diligence on a portfolio that included Madoff’s fund. Thorp pulled a sample of 160 trades Madoff claimed to have executed and checked them against the actual exchange records. About half the trades had no record of having occurred. Another quarter showed sizes larger than the day’s total exchange volume. Thorp told McKinsey it was a fraud and warned anyone who asked over the next 17 years until Madoff confessed in December 2008.

What is the Kelly criterion and how should retail traders use it?

The Kelly criterion, derived by John Kelly Jr. in 1956, gives the optimal fraction of capital to bet on a favourable wager to maximise expected logarithmic growth of capital. In its simplest form for a coin-flip-style bet: Kelly fraction = (probability of winning × payoff ratio − probability of losing) divided by payoff ratio. Most professionals use a fractional Kelly (typically half-Kelly) to reduce variance. The retail takeaway is not the formula itself but its underlying principle: bet sizing must be calibrated to your edge and your odds, not to your conviction or your account balance.

Why did Princeton-Newport Partners shut down?

Federal prosecutors led by Rudy Giuliani, in the course of the Michael Milken investigation, raided PNP’s Princeton office in 1988 and charged five employees with stock parking offences related to trades with Drexel Burnham Lambert. The convictions were eventually overturned on appeal in 1991, but legal costs and investor concern forced Thorp to wind the fund down at the end of 1988. Thorp himself was never charged with anything.

Did Thorp invent the Black-Scholes formula?

Not exactly, but he came remarkably close. His 1967 book Beat the Market, co-authored with Sheen Kassouf, presented a delta-hedging framework for warrant pricing that anticipated most of the structural insights of the Black-Scholes-Merton formula published in 1973. Thorp’s version did not include the explicit risk-free-rate-based derivation that Black, Scholes, and Merton would later formalise. He has consistently been generous in crediting Black-Scholes as the cleaner and more general formulation.

What is statistical arbitrage and how did Thorp pioneer it?

Statistical arbitrage is a quantitative trading strategy that constructs a portfolio of many small, short-term, mean-reverting trades across a large equity universe, designed to be neutral to broad market movements. Thorp’s PNP began running an early version in the mid-1980s, originally developed by Gerry Bamberger. The strategy depends on cheap computation, fast execution, and large numbers of small bets, and it has been the workhorse of many of the largest quant funds since.

Is Thorp’s approach still usable by retail traders today?

The specific edges he exploited at PNP have been arbitraged away by larger and faster firms. The principles, however, remain entirely usable: do the math, hedge what you cannot predict, size positions according to edge and odds, verify before you trust, and walk away when the edge is gone. These are not technologies. They are habits of mind. Retail traders who develop them have a real, durable advantage over those who do not.