The Machine That Ate Wall Street

There's a story Warren Buffett tells about a guy who made more money than him.

This doesn't happen often. Actually, it barely happens at all. But Jim Simons did it.

Simons ran a hedge fund called Medallion that returned 37% per year, after fees, for three decades. Buffett's Berkshire Hathaway returned about 20% per year over the same period. Both are extraordinary. But 37% versus 20% compounds into a different universe of wealth.

Here's what's interesting: Buffett and Simons couldn't be more different.

Buffett reads annual reports and thinks about businesses for years. Simons fed price data into algorithms that held stocks for minutes.

Buffett wants to understand what he owns. Simons built a black box so complex that even he didn't fully understand how it worked.

Buffett looks for companies he'd be happy owning for decades. Simons looked for patterns that might repeat in the next few hours.

Yet both got rich. Absurdly rich.

This tells you something important about markets: There isn't one way to win. There are many ways. And the way that works for you might be totally different from the way that works for someone else.

The Math Genius Who Broke Wall Street

Simons wasn't supposed to be a trader. He was a mathematician. A code-breaker. A university professor.

But in 1988, he started Medallion with a simple idea: What if we could use math to predict short-term price movements?

Not revolutionary companies or great businesses—just prices. Numbers on a screen, moving up and down in patterns that might, if you squinted hard enough through enough data, become predictable.

It sounds crazy. Except it worked. Better than anyone imagined possible.

Medallion's approach was the opposite of everything Wall Street taught. Instead of research reports, they used terabytes of data. Instead of company visits, they hired astrophysicists. Instead of understanding businesses, they understood correlations between soybean futures and Japanese yen movements at 2:47 PM on Tuesdays.

Their motto was "there's no data like more data." They collected everything: stock prices, weather patterns, news stories, internet posts, insurance claims. If it could be quantified, they wanted it.

Then they fed it all into algorithms that could spot patterns invisible to human eyes.

The Uncomfortable Truth

Here's what makes people uncomfortable about Medallion's success: It suggests that most of what we think we know about investing might be wrong.

We tell stories about great investors studying companies, understanding competitive advantages, thinking long-term. These stories feel good because they make investing seem like a skill anyone can develop with enough effort.

Medallion's story is different. It suggests that the biggest profits might come from finding mathematical patterns in data that require PhD-level expertise in fields most people can't even pronounce.

That's not inspiring. It's intimidating.

Warren Buffett acknowledged this when he said of quants like Simons: "They're very, very smart ... we don't think we know how to do it."

Even Buffett—Warren Buffett—admits there are investing approaches he doesn't understand and can't replicate.

The Paradox of Prediction

But here's where Simons' story gets really interesting.

Medallion was extraordinary at predicting what would happen in the next few hours or days. But when they tried to apply the same methods to longer time periods—months or years—they became ordinary. Their long-term fund delivered returns barely better than index funds.

This reveals something profound: The shorter your time horizon, the more predictable markets become. The longer your time horizon, the less predictable they are.

This seems backwards. We intuitively think long-term should be easier to predict than short-term. But markets don't follow intuition.

In the short run, markets are driven by mechanical patterns, algorithmic trading, and predictable human behavioral biases. Medallion learned to exploit these with mathematical precision.

In the long run, markets are driven by innovation, economic growth, and changes in human behavior that no algorithm can anticipate. Here, the advantage goes to people who can understand businesses and think through how the world might change.

What This Means for You

Medallion's success teaches us several things:

First, don't try to be Jim Simons. His approach required hiring some of the smartest mathematicians in the world, spending hundreds of millions on data and computing power, and developing algorithms over decades. As the document notes: "Don't try this at home."

Second, there's room for different approaches. Simons made money in minutes. Buffett makes money over decades. Both work, but they work for different people with different skills and resources.

Third, know your edge. Medallion's edge was mathematics and data. Buffett's edge is understanding businesses and thinking long-term. Individual investors' edge might be patience, or knowing particular industries, or simply not panicking when others do.

Fourth, respect what you don't understand. Simons revolutionized trading with methods most people can't comprehend. That doesn't make those methods wrong. It makes them specialized.

The Real Lesson

The most important lesson from Jim Simons isn't about algorithms or data or mathematical models.

It's about intellectual humility.

For decades, value investors believed their approach was not just the best way to invest, but the only sensible way. Buy good companies at fair prices and hold them forever. Everything else was speculation.

Simons proved them wrong. Not by beating them once or twice, but by beating them consistently for thirty years while doing everything they said you shouldn't do.

This doesn't mean value investing is bad. Buffett's track record speaks for itself. But it does mean that markets are complex enough to reward multiple approaches.

The investors who do best over time aren't those who find the one perfect strategy. They're the ones who find a strategy that matches their skills, temperament, and resources—then stick with it through the inevitable periods when it looks foolish compared to whatever is working best at the moment.

Jim Simons found his way. Warren Buffett found his way. Now you have to find yours.

Just remember: In a world where machines can process billions of data points per second, your advantage probably isn't going to come from trying to predict what happens in the next few minutes.

It's going to come from thinking about what happens in the next few decades.

The machines can have the minutes. We'll take the years.

Buy the book on Amazon: https://amzn.to/43NRkDS