Jim Simons: The Mathematician Who Built Wall Street's Greatest Money Machine

From 1988 to 2018, the Medallion Fund earned about 66 percent a year before fees and 39 percent after — a run so relentless that 100 dollars compounded, before fees, into nearly 400 million dollars. It was directed by a man who never took a finance class, never read a balance sheet for a living, and spent his thirties proving theorems in differential geometry. His name was Jim Simons, and he is the closest thing quantitative finance has to a founding legend.

Simons is the answer to a specific question: what happens when a mathematician of the first rank points his whole toolkit at the market and refuses to do it the way Wall Street does? The honest version of the story is more useful than the myth. He did build the greatest money machine in the history of markets. But the machine is closed to you, and the part you can actually keep is not the fund — it is the skill that built it.

Simons was born on April 25, 1938, near Boston. He finished a bachelor's in mathematics at MIT in 1958 and a PhD at Berkeley in 1961, at the age of twenty-three, under the geometer Bertram Kostant. His field was differential geometry — the pure study of curved spaces, about as far from stock prices as mathematics gets. He died on May 10, 2024, at eighty-six, one of the wealthiest people on earth. Almost none of the money came from the work that first made him famous.

That matters, because the temptation with Simons is to treat him as a lucky gambler with a good calculator. He was not. Before he ever traded a currency he was a mathematician other mathematicians took seriously, and the habits of that world — patience, proof, a refusal to believe a result until the evidence forces it — are exactly what he later turned on the market.

While he chaired the mathematics department at Stony Brook University from 1968 to 1978, Simons worked with the great geometer Shiing-Shen Chern. Their 1974 paper introduced what are now called the Chern-Simons invariants — objects that, years later, turned out to be central to theoretical physics, from topological quantum field theory to string theory. It is a rare thing for a piece of pure geometry to reappear as the language of physics, and this was one of them.

In 1976 the American Mathematical Society gave Simons the Oswald Veblen Prize in Geometry, the highest honor in his field. This is the fact to hold onto before the money starts: the man who built Medallion was already, by any measure, a mathematician of the first rank. The fortune was a second act.

There was an interlude that shaped everything after. From 1964 to 1968, Simons broke codes for the Institute for Defense Analyses, a research arm that did cryptography work for the National Security Agency. Codebreaking is, at its core, the art of pulling a faint signal out of what looks like pure noise — and of building small teams of brilliant people to do it. Simons learned both skills there.

He was let go after publicly opposing the Vietnam War. He returned to academia, to Stony Brook, and did the geometry that won the Veblen Prize. But by the late 1970s the itch to try markets had grown too strong to ignore, and in 1978 he left the university for good.

In 1978 Simons founded a currency-trading firm called Monemetrics, renamed Renaissance Technologies in 1982. The early years were not a triumph. Trading on hunches and fundamentals, he made and lost money and came close to quitting. The breakthrough was philosophical: stop trying to predict the news, and start treating the market as a signal-processing problem. Collect enormous amounts of price data, hunt for faint statistical regularities, and trade them thousands upon thousands of times, letting a tiny edge compound.

In 1988 Simons and the mathematician James Ax launched the fund that would make the legend: the Medallion Fund. It did not try to be right about any single trade. It tried to be very slightly right, very often, on a scale no human trader could match.

From 1988 to 2018, Medallion averaged roughly 66 percent a year before fees and 39 percent a year after — and those are arithmetic averages. Measured as a compound rate, the gross figure is about 63.3 percent. Compounded over the thirty-one years of that run, 100 dollars invested gross would have become:

That 398.7-million-dollar figure is gross, before the fund's fees — a true net terminal value is far smaller, and it is best read as a measure of the raw engine, not a payout anyone received. Set the machine beside the vehicles you can actually buy:

The net return investors would have seen is milder but still without precedent. Compounding one dollar at the 39 percent net average for thirty years gives an illustration — not a sourced payout — of the order of:

For contrast, one dollar left in a plain index fund at about 10 percent a year over the same thirty years grows to roughly seventeen dollars:

At a 39 percent net rate your money doubles about every one and a half years, by the rule of 72. And Medallion did this without a single losing year, net of fees — through crashes, panics, and the quiet years in between. The distance between seventeen dollars and nineteen thousand is the whole legend in one comparison.

An ordinary hedge fund charges '2 and 20' — a 2 percent management fee and 20 percent of the profits. Medallion charges 5 percent of assets and 44 percent of the gains (the performance cut was raised to 44 percent by around 2002). No fund on earth charges more. The reason is the opposite of greed, and it is one of the most instructive facts in the whole story.

Medallion's edge lives in strategies that only work at limited size. Every extra dollar of capital crowds the trades and drags returns down, so the fund has a natural capacity — secondary estimates put it around 10 billion dollars. Punishing fees are how you keep the fund small: they price outsiders out. By the early 2000s Renaissance had returned essentially all outside money, and Medallion has been employee-only ever since. The fee is not rent. It is a fence around a machine too good to share.

Simons refused to hire from Wall Street. He staffed Renaissance with mathematicians, physicists, astronomers, statisticians, and — crucially — experts in speech recognition and signal processing. In 1993 he recruited Robert Mercer and Peter Brown out of IBM's speech-recognition group; both would later run the firm as co-CEOs.

The logic was exact. Predicting the next word a person will say, given the words so far, is the same kind of problem as predicting the next move in a price, given the prices so far — both are the search for faint, fleeting statistical structure in a stream of noisy data. Renaissance was doing machine learning years before the phrase became fashionable, staffed by people who had never been taught the 'right' way to think about markets and so never learned the wrong one.

The most serious skeptical argument is about leverage. One analysis estimates that Medallion runs its positions at something like 12.5 times leverage, which would mean the underlying, unleveraged edge is only about 3 to 7 percent a year — genuinely excellent, but far less magical than 66 percent. This is one skeptic's reconstruction, not a disclosed figure, and it is worth stating plainly as that. Even taken at face value, though, it describes a real and durable edge, amplified by borrowing that a fund with Medallion's record can safely carry.

There is a second distinction that matters more for anyone tempted to chase the returns. Medallion is closed. The Renaissance funds that outsiders can actually buy — its public vehicles, sometimes called RIEF and RIDA — are run differently and have performed far worse, losing money in some stretches. 'Renaissance's returns' as sold to the public are not Medallion's returns. Whatever the true edge is, it has never been for sale.

By the time he died, Simons was worth about 31.4 billion dollars, among the fifty or so richest people alive. What he did with it is the part that best fits the mathematician he started as. In 1994 he and his wife Marilyn founded the Simons Foundation, which has given more than 4 billion dollars, largely to basic science and mathematics. In 2004 he launched Math for America to support public-school math teachers. In 2016 the foundation created the Flatiron Institute for computational science. And in 2023 he gave Stony Brook 500 million dollars — the largest unrestricted gift to a US university on record.

He took the market's money and pushed it back upstream, into the pure science he came from. For a man who made his fortune extracting signal from financial noise, it was a fitting closing trade.

The wrong lesson from Simons is 'there is a secret formula, and if I am clever enough I will find it and get rich.' The right lessons are quieter and more useful.

First, the edge was never stock tips or a magic equation; it was the disciplined extraction of tiny, fleeting statistical signals from mountains of data — statistics and signal processing, done relentlessly at scale. Second, the fund that produced 66 percent is closed to you, and the Renaissance products you can buy are not the same machine. Third, and most important: the money followed the skill, not the fund. The people Simons hired — mathematicians, statisticians, physicists, machine-learning researchers — are exactly the people a modern quant desk or data-science team pays for today. You cannot buy Medallion. You can become the kind of person Renaissance would have hired.

Can I invest in the Medallion Fund? No. It has been closed to outside investors since the early 2000s and is now employee-only. Renaissance runs public funds you can buy, but they are managed differently and have performed far worse.

How did he beat 'efficient markets'? He didn't predict; he measured. Medallion found thousands of small, short-lived statistical edges and traded them fast, at scale, letting a tiny per-trade advantage compound. That is closer to codebreaking than to stock-picking.

Was it luck, leverage, or skill? Leverage amplified the result — one skeptic's estimate puts the unleveraged edge at 3 to 7 percent a year. But thirty-one years with no losing year, straight through 2008, is very hard to call luck.

I'm good at math — what's the real lesson? The fortune tracked the skill, not the fund. Probability, statistics, and machine learning are what Renaissance actually paid for, and they are a career you can start on today.