How Jim Simons Built a Machine That Beat the Market

26:14That is the end of the break. Now we're going to talk about game theory and why it's good tax like a casino if you were Jim Simons in Renaissance. The year is 1988. Jim Simons Quant Investing Fund that started out as Monometrics has changed its name to Medallion because of all the mathematical prizes his employees have won, including himself. and they brought in this expert in game theory, Elwin Berlekamp. And Berlekamp ran in the same circles as a famous investor and gambler named Ed Thorpe, who you've met. I interviewed him, yeah. He's a super interesting guy, known now as an investor. He made a ton of money as an investor.

26:53But he, earlier in his career, went to Vegas, learned how to count cards, wrote a book called Beat the Dealer, did this amazing thing in, I think, the 60s with Claude Shannon and other super interesting guy where they built this machine to like understand how the roulette ball was going to break and like basically beat roulette like extremely interesting finance math guy yeah so elwin berlecamp knows ed thorpe and is thinking about gambling strategies right and he looks at the medallion fund and he says you're making all these big bets on market moves right And you win sometimes and you lose sometimes.

27:32But the key is, if you're going to rely on probabilities, you need a lot of bets. Imagine a casino that rolled the roulette wheel once a day. Yeah. Right? They could win money from the customers. They could lose a bunch of money. But you wouldn't know. It's random at that point, right? Right. But you do 10 ,000 roulette wheel spins, 100 ,000, a million roulette wheel spins. and the casino has an edge and they will win money after a million turns. It's the law of large numbers. If the odds are in your favor, you want to be making essentially as many bets as possible. Yeah. And he says, let's do at the Medallion Fund what casinos do.

28:13Get a tiny advantage, figure out how to win slightly more than 50 % of the time and then just make a lot of bets. So instead of, you know, making one bet and seeing how it's going to turn out. Going all in on main potatoes and seeing how the harvest comes out next year? No. You are going to make bets on the market that last maybe a couple of days, one day, intraday, this sort of high-speed trading that we would eventually see in the market. And here's an example of, like, what they started to see when they did this. They needed just tiny little advantages they could exploit in a small way. So the computer spotted something that was futures traders out in the market.

28:55if they had a good week, you know, made money on their contracts over the week, they would tend to sell at the end of the day on Friday and then rebuy the contracts on Monday morning. So if they're selling on Friday, the price goes down a little bit. If they're buying on Monday, the price goes up a little bit. And they didn't really know why. Maybe, you know, the futures traders wanted to chill on the weekend or, you know, maybe they were worried about world events, whatever it was, the Medallion Fund was like, well, we're just going to buy a bunch of future contracts on Friday. We're going to sell them on Monday morning.

29:31We're going to make a little bit of money. And that is consistent, at least until someone else discovers it. Yeah. I mean, that one is interesting, right? Because you can tell it as a story and it makes some sense. And there's like human beings taking actions for reasons. Presumably, the best things, most of the things they do, there is no story. It's just the computer is like, do this, and they do it, right? That, to me, is the ideal version. Because if you can tell a story about it, then somebody else is going to figure out that story. Why do you think I picked it out? I am a storyteller. I want to explain this to you.

30:05That's why I picked that example. No, presumably, thousands, thousands of bets were things you could not tell a story about because they were correlations of an interest rate in Japan affecting the commodity prices in Mexico. and you can't see the connection. Matt Levine, the finance writer you and I both like a lot, he had this thing once about how Renaissance, Simon's firm, actually at some point didn't want finance people, right? Because finance people are always looking for the story. Why? What is the story of this? But if you can tell a story, there is no edge, right? Because somebody else can tell the story.

30:41So you just want to trust the machine. In Greg Zuckerman's book, one of the traders for the firm said It would probably be better if stock prices didn't have names on them. Like if the company didn't have a name? If we were saying, like, oh, what should we do about Oracle stock? Yeah. So if you're trading Oracle, you probably have some deep emotional feelings. You like Larry Ellison. You don't like Larry Ellison. But if it were stock number 87, you would be able to just be like, well, this is the movement in the numbers. And I don't care what the company does or why. Basically, the machine says buy stock 87.

31:14Okay. Buy stock 87. Yeah, it would be fewer employees stopping the machine, right? So this casino tweak that Brilla Camp brought, it actually worked. And by 1990, they're having very good trading days. They're like going up 1 % a day sometimes, and they were celebrating. They had to start a new rule, which is you cannot break out the champagne unless you go up 3 % a day. In one day. In one day. A 3 % return in one day. For the whole year of 1990, Medallion Fund, it's finally churning, right? It returned 55%, and that's after fees. And they were charging a ton of fees. I would pay them a lot of fees to return by 55 % after fees.

32:00I guess I'd pay them any amount of fees. Yeah, exactly right. So, obviously, Jim Simons, Renaissance Technologies, the Medallion Fund, they are not the only people doing this. This is the 1990s, right? There's David Shaw starts D.E. Shaw, big quant trading firm. Kev Bezos worked there early in his career. Yes, that's correct. Ken Griffins had just started Citadel. Kepler, Morgan Stanley, everyone's playing around with these techniques. But Jim Simons and the Medallion Fund, they managed to stay ahead of them all. In a field where everybody's looking for an edge in the data, they had an edge on the edge.

32:36So how did they do it? How did they do it? So there's a couple of things. This is my opinion, right? So one thing they did was to stay small and focused. They closed the Medallion Fund to outside investors and said, we are going to trade solely for the employees of Renaissance itself, for the people who work here. Because if you have tens of billions of dollars,$100 billion, it's hard to take advantage of these tiny little moves. Well, right, because once you start buying, you drive the price up or you start selling, you drive the price down. You get big enough, you're actually moving the market and losing your edge, right?

33:13Classic problem for big funds, actually. Medallion also focused their efforts on where their data was the best, which was commodities, currencies, and bonds, mostly. They would eventually do stocks, but we'll get to that in a moment. And Jim Simons had this other edge that I feel like is super rare on Wall Street. His employees loved him, and they were loyal. And almost no one ever left. I mean, they liked the intellectual atmosphere being run by mathematicians. The money, specifically. They liked the money. And remember, they were sort of sequestered out on Long Island in a small town, right?

33:49They weren't having beers with the guys from Goldman Sachs. They weren't getting job offers on the street because no one knew their names. It's almost like the machine, the machine learning is a black box. But then the firm itself, Renaissance, is like a black box around the black box. Black boxes all the way down. Until the pile of gold at the center. But, you know, his employees were happy and content and obviously becoming very rich until Jim Simons made one hire that maybe wasn't the best idea. In a minute.

34:37Women deserve tools that support their health at every stage of life. Oura Ring is more than a wearable. It's a powerful health companion, helping women take control of their well-being with personalized insights. From sleep to stress, activity to cycle tracking, Oura Ring gives you over 50 metrics that help you understand your body better so you can be your own health advocate. With its accurate heart rate tracking and stylish, comfortable design, Oura Ring is made to be worn 24-7 and actually looks good. It's even FSA and HSA eligible. Oura doesn't just show you the data. It helps you turn insight into action so you can show up as your best self every day.

35:22This is personalized health powered by data designed for women. Be the expert in you. Visit auraring.com forward slash decoding to get 10 % off the new Aura Ring 4.

35:59Jim Simons:you'll ever need. It's an all-in-one, fully integrated platform that handles everything. CRM, accounting, inventory, e-commerce, HR, and more. No more app overload. No more juggling logins. Just one seamless system that makes work easier. And the best part? Odoo replaces multiple expensive platforms for a fraction of the cost. It's built to grow with your business, whether you are just starting out or already scaling up. Plus, it's easy to use, customizable, and designed to streamline every process so you can focus on what really matters, running your business. Thousands of businesses have made the switch, so why not you?

36:36Jim Simons:Try Odoo for free at odoo.com. That's O-D-O-O dot com. Every business has an ambition. PayPal Open is the platform designed to help you grow into yours with access to business loans so you can expand and hundreds of millions of PayPal customers worldwide. Your customers can pay all the ways they want today with PayPal, Venmo, Pay Later, and all major cards so you can focus on the future. When you need a partner trusted by millions, there's one platform for all business. PayPal Open. Grow today at paypalopen.com. Loan subject to approval in available locations.

37:24Finally, Jim Simons is going to make a mistake. As I recall from before the break, Robert Smith, tell me what happens next. Such a small mistake. He still makes a ton of money. Spoiler, right? There is one place in New York State with as many math geniuses as Renaissance Capital, IBM. Specifically, the Watson Research Center Upstate. And it was there that Jim Simons found a pair of geniuses working on speech recognition algorithms. Peter Brown and Robert Mercer. And they arrived at the Medallion Fund in 1993, and they solved this last big challenge the fund had, trading stocks. Now, of course, you may notice that linguistic speech recognition software is very similar to what the astronomers were doing, to what the codebreakers are doing.

38:14So it's classic machine learning, classic machine learning problem. them and they come in and they look at the problem of trading stocks so of all the things on wall street apparently like stocks are the most kludgy and human you know you've got to have someone make the trades for you right okay at that time at that time and the computer would have these fancy optimal trades that you had to do but then sometimes in the real world it would just be like no, you can't short that. Or, you know, we can't get the leverage necessary, the margin on the stock to make that trade happen. And the computer would just sort of stop, you know, and couldn't do the trades they needed to do.

38:57Remember, there's a lot of things that have to happen for all of these to work. So it's like the second order effects in some like, imagine a frictionless plane universe. Yes. The model could win in the stock market, but the frictions of the actual stock market were stumping it. Yeah. So Brown and Mercer come in and they programmed their computers so that it was all working together. I guess it was in pieces before, but they're like, we have one model for everything. And what that meant is that the model itself could change its algorithm as it went. And so if certain trades were not working, it could find the second best, the third best, and then adapt all the strategies depending on what they were actually able to do in the real world.

39:43So it's sort of like real time or almost real time updating and feedback. And if certain trades were working, the computer could allocate more money to those trades without a human being like, yes, no, spend my$10 million. Don't spend my$10 million. dollars. And the two of them were legendary because they worked as this sort of pair. In the Greg Zuckerman book, The Man Who Solved the Market, he calls them like Penn and Teller. So Peter Brown was the, I forget who's who in that. One of them talks and one of them doesn't. They're both magicians. So which Peter Brown was the one who talked or the one who didn't?

40:17Yeah, the one who talked. So Peter Brown did all the talking, right? He never stopped moving. He He slept at the office. He yelled at people. He inspired people. He was just this big character. And Robert Mercer, Bob Mercer, was the silent one. He was quiet. In the telling, Peter always says, well, Bob comes up with all the ideas silently, and I make them work. Silent Bob. Penn and Teller would eventually double the profits of the stock trading arm, and they would run the firm eventually when Jim Simon stepped back. They moved the Medallion Fund into foreign markets. It was this constant move of collecting more data, more sources to have their giant computer system look at, and then trading in more and more and more markets, right?

41:01And, you know, I would love at this point to have, you know, some sort of crisis in the computer, some sort of moment where they stop making money. You would love it for narrative purposes, to be clear. Please. Yeah. But, you know, it didn't really happen. In the worst possible markets, the Medallion Fund was doing well. I mean, there would be some dicey days, like literally days like during the tech crash in 2000 where they're like, oh, we're losing a bunch of money. What are we going to do? And, you know, a few weeks later, oh, yeah, we are fine. Right. It kept going up, as we said, you know, sort of more than 50 percent a year and with low volatility.

41:34Like there's no risk to this. It's just insane when you think about it. You shouldn't be able. Like, I don't mean morally. I mean, just kind of the laws of investing physics suggest that you shouldn't have that kind of return year after year after year. It's just somebody should catch you, but it doesn't happen. And it's interesting, right? I think anyone else who did this would say, I am a genius. I have insight. I know what I'm doing. But you'll notice in this story, like Jim Simons himself has sort of stepped back a little bit. He's running the machine. He's hiring the people. He's letting them run the actual machine, right?

42:13And even the people who are making these decisions at this point, like Robert Mercer, he has a line that I love from the Zuckerman book. I'm going to read this here. We're right 50.7 % of the time, but we are 100 % right 50.7 % of the time. You can make billions that way. Isn't that basically the anchorman line? There's that line from Anchorman. There's an anchorman joke. I think it's 60 % of the time it works every time. Exactly right. So it's funny, right? Because that is the philosophy that gets you this money. That's the casino. Like, that's how casinos work. You just need the little bit of edge, and then you need to ruthlessly exploit that, and you have to keep the edge.

42:59So at this point, there's really no drama from the algorithm. It is working. It is succeeding. it is evolving, really. The only drama really comes from the personalities. As Jim Simons gets older, he starts to step back from the business. And I will say some minorly dodgy things start to happen. At one point, they get caught booking short-term gains in the market, which means immediate trading profits, as long-term gains. Essentially, paying less tax than they should. And the way they do this is they're trading like with the stocks in a basket stored at the bank. They don't really take the profits out till after a year.

43:40Well, whatever it was, the IRS is like, no, no, you can't do that. And so they had to pay a reported$7 billion to the IRS. A lot. Sounds like a lot. I don't know how much they have, but$7 billion is a big number. They have more than that. They have more than$7 billion, yeah. And then Bob Mercer, Silent Bob, starts to get a little bit weird at the office, right? He starts to get into politics. Always a bad idea. So apparently he's very conservative. He's a right-wing guy. And he starts to talk about conspiracies at work. Now, this is puzzling to everyone, right? Because we have a firm of brilliant mathematicians, Bob Mercer, brilliant, logical guy, right?

44:20But he'd get into these arguments at the firm about how, oh, you know, global warming is overblown or that like radiation exposure is actually good for you. This is all stuff that's like circulating the Internet and fringe, you know, science stuff. He actually funded a scientist who collected human urine to extend human longevity. Okay. I guess it didn't work as far as I know. I don't know. But it started to rub people the wrong way in this like really smoothly functioning, you know, black box out on Long Island because he was now super rich. Robert Mercer, he started using the money to actually influence politics to support right wing causes.

45:04He and his daughter especially created a political action committee. They started to work with Steve Bannon and Kellyanne Conway before Donald Trump was running for election in 2011. and apparently he was the one that urged Donald Trump to hire them, right? That sort of sent the Donald Trump campaign in sort of a rightward direction. So he becomes a big donor to Trump. This comes out and all of a sudden Bob Mercer and Rebecca Mercer, his daughter, become this sort of symbol of the shadowy right-wing money-making forces propelling this president that certain people don't like into power. people start to picket outside the firm and appearances by Robert Mercer.

45:47There's a huge piece in The New Yorker about him. And eventually Jim Simons, who at this point, you know, is leaving the working of the firms to Bob and Peter. He has to step in and say, Bob Mercer, you need to resign. And Bob Mercer steps back. And in case you were wondering, they still made money. They still made money through the whole thing. whole thing. At this point, you can have all the drama in the world because the computer has no drama anymore. It is just getting better and better. How about this for a fuego take? Maybe the human drama is actually helpful for the firm. Yes. Because it distracts people from wanting to question the results of the computer.

46:30The more they fight, the less. The better their returns. The less they want to fiddle with the computer. That's exactly right. So what did you say, Robert, the return has been for whatever 30 years. Did you say 66 % a year? 66 % were the trading returns. Okay. And then there were fees, which got bigger and bigger every year because they're building this giant computer system, right? And so after fees, 39%. Still unbelievable. So, okay. So that's since 1988 to 2021. 21. Okay. So let's just play around with that for a sec because I think it's hard to really grasp, right? So I was a teenager in 1988.

47:11I bought CDs, right? This was the CD era. What did CDs cost in 1988? $14.99 at Tower Records. Okay. Sam Goody. I wasn't cool enough to shop at Tower Records. I shopped at Sam Goody. 1988 albums. Oh, Green. R.E.M. Green, which I loved. I'm already doing the math for you. Instead of buying R.A.M.'s green for$14.99 in 1988, I had put$14.99 into the Renaissance Medallion Fund and left it there and let them charge me their fees. How much would that$14.99 from 1988 have turned into by, what did you say, 2021? 33 years at 39 percent annually,$785 ,000. You could buy a house instead of green, although green was a fabulous album.

48:08It was a great album. But I'd rather pay for my children to go to college than have listened to green. That is amazing. And then, of course, the wild thing is if you'll leave the$700 ,000 in there for one more year at 39%, it's like a million, basically. Yeah, exactly. Exactly. I also bought Rattle and Hum that year. I'm just looking at the list. So that's another$700 ,000. grand. So needless to say, Jim Simons made a ton of money out of his money-making machine and made money for a ton of other people. He died in 2024 and spent his last few years giving away a lot of that money and traveling on his yacht.

48:46I mean, he earned it, right? He was a big donor to Stony Brook University, where he had run the department. He supported research into autism. And I love this part. He had a fund to help math teachers stay in public education by basically giving the money to still teach in high schools instead of, you know, going to Wall Street like he did. Fine. At the end of these investing episodes, I usually say that the investor that we're featuring found an inefficiency in the market, right? That they somehow did a bunch of research, found data that no one else had, and then they exploited that in the market to make a lot of money.

49:29So when I talked about Jesse Livermore, the speculator back in 1929, he was a chalkboard boy who knew the numbers of the stock market inside and out. He had that intraday trading data in his brain because he was writing it down on the chalkboard all day long. Exactly right. And when we talked about Warren Buffett, he would actually physically go to the CEO of a firm and get information that no one else had at the time. Like on Saturday, you just go on the weekend and knock on the door and say, tell you about your company. The thing about those two examples is that eventually any sort of edge you have in the market gets competed away.

50:07Other people see the edge. Other people get the data. Other people see what you're doing and just do the exact same thing. But Renaissance Technologies and the Medallion Fund, at least so far, are in the middle of this. And I don't know if it's that they continue to ingest so much more data that they're finding new inefficiencies all the time, or if their computing power is just getting more and more powerful so that they're able to spot things that no one else has. But when it's like the Medallion Fund, you know, it's still going strong. It's still making money for its employees, its investors.

50:42One of the things I do love about this story is that even the people involved, the mathematicians, are just like stunned by the amount of money that they make. You know, in Wall Street, there's a kind of feeling like, oh, we earned it. But these are logical, rational people who at various points in this story stop to ask themselves, what are we doing? Like, are we doing good? Are we doing harm? If we're making all this money, who is losing the money? Who's on the other side of these trades? You know, if we have 50.7%, who's at 49.3? and they decide at some point, they're like, yeah, you know, we're adding liquidity to the market.

51:27We're, you know, closing inefficiencies. But really what we're doing is there are people out there making emotional decisions and we're not making emotional decisions. And when they get a little too optimistic or a little too pessimistic, they sell early, they buy too soon, whatever it is, we spot it, we take a little percentage of that trade, and we win. That is their theory about what's happening out there. It's them versus the emotions. That sounds like a suspiciously human-centered narrative story, right? There's another version of the story, which is maybe it was that at the beginning, but surely at this point, a lot of the time.

52:09It's the Renaissance computers against somebody else's computers. And it's not that the other computers are more emotional. It's just that the Renaissance computers, the Renaissance algorithm, the Renaissance AI is just better. Their computers are just better than the other computers for now. And the key to that is you would have to read this story in machine language. Why are the Medallion computers better than everyone else's computers? Only the computers know. Our producer is Gabriel Hunter Chang. Our engineer is Sarah Bruguer. And our showrunner is Ryan Dilley. I'm Jacob Goldstein. And I'm Robert Smith.

52:46We'll be back next week with another episode of Business History. A show about the history, wait for it,

52:53Jim Simons:of business.

52:57Jim Simons:You know that feeling when a story just grabs you and won't let go? That's the kind of drama that's waiting for you on Disney Plus Hulu. Mysterious post-apocalyptic thrillers like the acclaimed Hulu original, Paradise. Action adventure dramas like Daredevil Born Again. And iconic medical dramas like Grey's Anatomy. Or maybe you want your drama with a side of comic relief with shows like High Potential. Find the drama you want on Disney Plus and Hulu with a bundle subscription. Terms apply. This is Matt Rogers from Lost Culture East. us with Matt Rogers and Bowen Yang. This is Bowen Yang from Lost Culture Research with Matt Rogers and Bowen Yang.

53:36Hey, so what if you could boost the Wi-Fi to one of your devices when you need it most? Because Xfinity Wi-Fi can. And what if your Wi-Fi could fix itself before there's even really a problem? Xfinity is so reliable, it does that too. What if your Wi-Fi had parental instincts? Xfinity Wi-Fi is part nanny, part ninja, protecting your kids while they're online. And finally, what if your Wi-Fi was like the smartest Wi-Fi? Yeah, it's Wi-Fi that is so smart. It makes everything work better together. Bottom line, Xfinity is smart and reliable. You deserve the peace of mind of having Wi-Fi that's got your back.

54:10Xfinity, imagine that. When you buy business software from lots of vendors, the costs add up and it gets complicated and confusing. Odoo solves this. It's a single company that sells a suite of enterprise apps that handles everything from accounting to inventory to sales. Odoo is all connected on a single platform in a simple and affordable way. You can save money without missing out on the features you need. Check out Odoo at odoo.com. That's odoo.com.