This week on generating alpha in an episode unlike most, I'm joined by Jeff Yes. Founder of Susquehanna International Group, one of the most successful trading firms in the world. Jeff is a legendary figure in finance, known for applying the principles of poker, probability, and decision theory to markets. Over the past four decades, he's built a global powerhouse, quietly operating behind the scenes of Wall Street, trading everything from options to crypto, all grounded in mathematical precision and rational thought. He's also one of the most influential and private figures in modern finance, making this conversation one of his first interviews ever. In this short episode, we talk about prediction markets. Why Jeff believes they're the future of how we understand truth, how they can improve decision-making and business in government, and what they reveal about the power of incentives, information, and human behavior. I really enjoyed recording this episode and I hope you guys enjoy listening. Thank you, Jeff, for coming on. I really appreciate you making the time.
本周,我们带来了一期与众不同的《生成阿尔法》节目,我邀请到了杰夫·耶斯。他是全球最成功的交易公司之一——Susquehanna International Group(锡古国际集团)的创始人。杰夫在金融界是个传奇人物,以将扑克、概率和决策理论的原则应用于市场而闻名。在过去的四十年中,他悄然在华尔街的幕后建立了一个全球强大的企业,涉及的交易从期权到加密货币,所有操作都以数学精确性和理性思维为基础。他也是现代金融界最具影响力且极为低调的人物之一,所以这次对话是他首次接受采访之一。在这简短的节目中,我们讨论了预测市场。杰夫为何相信预测市场是我们理解真相的未来,它们如何能在政府和商业中改进决策,以及它们揭示了激励、信息和人类行为的力量。我非常享受这次录音过程,希望你们也喜欢收听。感谢杰夫的参与,非常感谢您抽出时间。
Jeff It's my pleasure, Amir. Let's go. So to set a foundation for this conversation, I'd love to kind of start simple. What's your current perspective on prediction markets as a whole, and how are they significant to Susquehanna and yourself? Well, prediction markets have been a great passion of ours for years. They add tremendous value to the world. You basically can't make a good decision without knowing the probabilities of events happening. Prediction markets are the best way we know how to get the most accurate guess of what those predictions are. So we think it's a fantastic tool that will add tremendous benefits to society.
In from a broad perspective, how do you see the kind of evolution of prediction markets playing over the next decade, especially in terms of regulation and gambling legislation? Well, in the gambling world, we're really not sure. I think the world is coming to the conclusion that a system like the one they have in Europe, like Betfair, where people combine sell amongst themselves is a much fair system. We'll reduce costs tremendously for customers. Currently, the big is somewhere around 5%. If you can trade amongst yourself on exchange, we think they'll go down substantially probably to 1% to 2%. So that will be a big win for people who want to engage in sports. But our real motivation for prediction markets is to get the truth out there.
Our favorite example is during the Iraq war, when George Bush first went into Iraq, he said it would cost $20 billion. Lawrence Lindsay, his economic advisor, said, I think of my cost as much as $50. He was sort of punished for saying that. The true number has come in somewhere between $2 and $6 trillion. So had the people had a prediction market, what's the over on the line and how much this would cost? I don't think it would be anywhere near $2 to $6 trillion, but it would have been substantially higher than $50 billion. Let's say it would have been $500 billion. Then the people might have said, look, we don't want this war. Politicians always tell us that the wars are going to be cheap and quick and fast, and they never are. So we need a trusted source and prediction markets would be an objective trusted source, because anyone betting on them, it's going to lose money if they get their analysis wrong.
So had we seen this gigantic number, I think there would have been much more pushback against this war. And prediction markets could be that powerful, where they can really slow down the lies that politicians are constantly telling us. And that's really sort of my number one reason why I want to see them thrive. It's almost the people's idea of the truth, rather than the kind of tainted idea of the truth. That's given to the general population. But also with experts. I mean, you may not know what wars course, the vast majority, people don't know, but there's a small group of people who do, and they would be betting it, and they would be bidding it up to a price that makes sense.
So the public who may not, you know, how the hell you're going to be informed about what a war is going to cost if you're a regular person. But if you see experts battling it out and betting on it, then you can trust that number, and you could be more of an expert by looking at a prediction market than a politician can, who's just either making up a number or purposely lying. And I also assume in the future that prediction markets can be used, and we'll be used to price more like financial instruments and support other decisions.
But how can we protect from prediction market manipulation? Well, in the same way you're protecting as any other manipulation, if you're manipulating the price, you're going to lose money. If there's enough players out there and you want to get a price up to something for some nefarious reason, you're going to have to lose a lot of money to do it. So if you want it a bet that it's going to be under $50 billion spending, well, we will bet you hundreds of millions of dollars that you're wrong.
So your plan is going to be very, very expensive, and it'll probably be more expensive than just a misleading advertising campaign, which is just my cost in the millions, this would cost in the hundreds of millions. So that will protect the integrity of the markets. And I want to take a step back for a moment. Early in your career, your professional gambler, specifically in poker and horse betting, what do you see as the parallels between gambling and prediction markets and what systemic risks and opportunities do you think are introduced as a result?
I don't really see any systemic risks. I see more truth, more rational objective probabilities getting out into the marketplace. And I see this systemic risk as politicians telling us stuff that they're trying to trick us. And this is the antidote to that. So I see, obviously, there could be some tiny amounts of manipulation, but that's going to be trivial compared to the amount of manipulation that we have now. Competitive markets will wipe out any problems that we may see.
And from kind of like a broad overview, how do you think your firm infirms like yours will incorporate prediction markets into their daily decision making? Well, for example, there's an election in New York City in 2015. If you listen to TV, cable news, you get very hard to figure out what the probability is. Some people say, no, it's going to be too close. Come on, New York's not going to elect someone like Ma'am Dammy.
But when you look at the prediction market, you see he's a little over 90% to win. If you're making a decision and you want to move to New York, you want to move to business to New York or whatever, you need to know that probability. It's very hard to know just by reading the newspapers or listening to the news and to have that actual number helps you dramatically in that decision. Plus, let's say you're a real estate developer and you think that your value of real estate is going to go down by a million dollars if Ma'am Dammy wins.
You can hedge it so you can buy insurance. But more importantly, you can get, you know, you can find out what the best guess is and you can do it in an instant. You just look at the price, you have the probability, you don't have to do all kinds of work, you don't have to read a million articles and call posters and posters and do all the work. All the work is done for you and you get the best possible number you can. And that will guide you through all your decisions.
For Susquehanna, you know, we're constantly looking at what are the odds, let's say the presidential election and stocks are going up and down based on who's going to win and who's going to lose. And we use that number to determine if we think a stock has overreacted or underreacted to the political odds. And I imagine as prediction markets become bigger and bigger and there's more volume that larger firms will start participating and actually hedging on the prediction markets rather than using outside financial instruments to hedge.
So my question kind of around that is you recently joined forces with Kalshi to provide liquidity as one of its primary market makers. How do you believe the involvement of firms like yours will evolve with the markets? Yeah, that's a great question. We, right now it's still a bespoke product. Institutions aren't really using it. There's a lot of action, but it's mainly relatively small bettors. No giant institution has really showed up and wanted a hedge.
You know, will the fed raise rates or not yet on these things? But we think as they get regulatory clarity and as they grow in popularity, institutions will show up and there will be Wall Street size bets placed on these things. But that has not yet happened. I mean, if you're, you know, an investment bank, if you go to the Saxon Morgan Stanley, you're a little cautious about betting on these things, which you haven't done it yet, but eventually they'll go away.
What I really hope that prediction markets could influence is the insurance business. You know, insurance in some places is impossible to get the government caps the rate that you can sell it at. So a lot of insurance companies have left Florida, for example, and you can't ensure your home because the price is too low. But if we had insurance bets on prediction markets, you can live in an area we can put up a price and say, will the wins get above 80 miles per hour in the next two days in your area? And let's say there's a 10% chance of it happening. If you think that if that happens, you're, you know, you may suffer serious damage to your home, you might want about $10,000 on it to win $90,000. If it happens and now cover your cover most of your insurance costs, and you only have to buy it when there's a, you know, when there's a problem coming, they will take out all the adjustable claims, all the expense, all the advertising of insurance, making much, much cheaper, and much, more sort of bespoke to what you need to have happened.
So, you know, there's enormous expense in insurance, and this would reduce, reduce a lot of it, making much easier for people to, to ensure what they really have. It won't be as perfect as an insurance claim where my roof low off, give me my money, it'll be more like, well, the win was really bad. I know my house is messed up, how messed up I don't, I don’t exactly know, but because it's so much cheaper, you could, you could much more easily hedge your risk than you can with typical insurance products. And it's so much more quantifiable. The insurers will obviously try to see how much you need, and how much they'll give you, and stuff like that, and with prediction markets, it's so much more quantifiable. And as these prediction markets kind of evolve in maturing to some day fully regulated exchange that I imagine, do you think the majority of liquidity will come from large Wall Street firms, or do you think they're going to come from retail flows?
I think it's going to come from both. And I think that it's going to create tremendous opportunities. Let's say you're just a weather person that loves following weather and hurricanes and probabilities, and you live in Florida, and you can put out your own markets and say this area, I think is these are the odds of disruption, and in these areas, these are the odds, and you can have relatively small businesses who have expertise in this stuff, who now have no way to make any money from their expertise, and they can be putting out markets, and they can be making a lot of money and reducing the price for regular people. Yeah, I think it's incredible.
And do you think at all in the future prediction markets can influence outcomes? No, that's like one of the myths that that someone's going to bet, there was that story on polymarket that the French guy was betting Trump, and it was just nonsense. We bet against them. If he bids it up, we'll bet it down. It's not going to influence anything. So all that's a fear that comes to mind, and it's not a zero probability that can happen, but it's really vastly overstated. And what do you think is the most significant obstacle to broader participation in prediction markets, and how can one go about removing that obstacle?
The biggest obstacles, like as you ask these questions, you can see what could go wrong, what can go wrong, what can go wrong? Those things psychologically come right to your mind that these things could go wrong, and yes, something could go wrong, but something is already going wrong. So that obstacle, as we get used to it, will go away. It's going to take time, but people have fears, and they overstate the downside. But as the product takes place, and people learn how valuable it isn't how much money it can save them, those fears will dissipate. This may take years, but I'm very optimistic that we're going to get there.
Before we go back to the episode, I want to take a short break to talk about my sponsor, Row. The generating alpha podcast is presented by Row, the all-in-one banking platform for startups. Thousands of startups like their complexity, product tons, and more use Row. You get everything you need to manage your startups cash, fast banking setup, cards with a 2% cashback, a yield that turns company cash into extra runway. All super important in the early days of launching. But the thing founders really love about Row is their team. Their obsessed with helping founders disrupt the status quo, and will go to the end of the earth's health and the do so.
And exclusively for generating alpha podcast listeners and viewers, you'll get a $1,500 statement credit plus a ton of exclusive perks when you manage your company cash with Row. Terms in condition to apply. To learn more, visit rho.co slash generating alpha. Row is a fintech, not a bank. Checking and card services provided by Webster Bank, member of FDIC. See your award terms for details. Thank you, and back to the episode.
And I know you come from a very probabilistic background, but with the rise of decision markets, there's under-prediction markets. Is there any type of decision, or even prediction that we should deliberately avoid quantifying? That's a good question. You could put up on a prediction market. Should I marry this girl or not? No. And maybe some, you know, your friends and your relatives might be more objective than you are. But I'd say that's going a bit too far. So my answer sort of would be no.
And what is possible with prediction markets that no one is talking about right now? Or what do you think is possible prediction markets no one's talking about? I think the number one thing it will stop wars because every war is exaggerated. How quickly it'll be ended. It'll be ended. And how little it'll be cost. It will cost. And how many lives will be lost is always lied to us by our politicians.
And Abraham Lincoln, you know, in the civil war in 1862, the war department, stop taking in the north, stop taking recruits. They said this war will be over in a couple of weeks. You know, he was off by 650,000 deaths and stuff. So he honestly believed that it was going to be a short, quick war. And obviously it wasn't. And it still reverberates now, you know, the horror of the civil war. If the people knew how expensive it's going to be and how disastrous it's going to be, they will try and come up with other solutions besides going to war.
Another example I can give is driverless cars. There's a lot of opposition to driverless cars because people can imagine a robot going crazy and killing somebody. But, you know, this year, in the next 12 months, about 40,000 Americans will die on their roads. If we had driverless cars, I'm guessing that number would probably be about 10,000. You know, down 75%, we'd say 30,000 lives. If we put that up in prediction markets and said how many lives, you know, in 2030, how many people will die in car accidents? And the numbers vastly lower than it is now, because people expect driverless cars to happen.
It would make policy makers hustle and hurry up and getting driverless cars there. But we're going to have this gain of tens of thousands of people who aren't going to die right now. You know, you so I say, I don't know, maybe driverless cars will be good. Maybe they won't. If we had an objective number on it, I think we would see how great it is. And we move much, much faster. I think it's an incredible kind of use case of it, especially for quantifying things for policy makers to make decisions based off of.
And before I move on to kind of a question or two about advice, what is the one message that Jeff Yass wants to tell the world about prediction markets? If you give one message to the world, if you were selling the world in prediction markets, would it be? It is my mother used to say to me, if you're so smart, how come you're not rich, the prediction markets are objective. If you think the odds are incorrect, then go bet it and go put it, go put it back into where in line where it should be. If you really are smarter than the markets, you'll make a lot of money. You'll do society a favor, which you'll get the price.
You get the price right. And if you can't make money, you may want to consider being quiet. Like maybe the market knows more than you do. Now this is going to infuriate every college professor you're ever going to have because they want to be the experts, but they're not. A bunch of speculators battling it out every day in the market place will be vastly greater. It will insult the college professors, which as far as I'm concerned is a good thing.
I agree. I'm going to give you an example when my daughter was 12 years old. Obama was running against Hillary Clinton in the primary. One of the most famous political scientists in America was on TV saying, no, Obama Hillary Clinton is going to win. She's up by 30 or 40 points. My daughter said, go check trade sports. It was just the only place at that time to look. And she said, Obama has a 22% chance of winning.
So the market place knew that Obama was special. That he was charismatic. He didn't have any name recognition. And Hillary did. So the fact that he's down by 35 with months to go doesn't really mean anything. So I use that as an example that my 12 year old daughter had a better guess of who's going to win that primary than the world's foremost expert in polyps high. And that's the power of prediction markets. That's that's an incredible incredible incredible example.
And I'm going to ask two questions about advice. The first one being as a high schooler today, I'm a high schooler and I'm given all the success that you've had and given all the hiring that you've done. What should students today study? I would really strongly suggest. I mean, obviously, you know, computer science, you got to be computer literate and you got to you got to understand what AI is coming from. But if you really want to be a decision maker under uncertainty, which is what humanity is, you have to learn probability statistics so much of what happens, you know, in the world is you are making a decision.
And if you're not really informed on the mathematics behind probability statistics, you can make a terrible decision. So when you see that there's a hurricane season, there's a lot of hurricanes like, well, is this a big deal or there are always a lot of hurricanes? And what's the volatility around hurricanes is it vary by a lot? Is this such an outline? Does this prove this global warming or is this just a blip? So it's sort of the signal versus the noise. And to be able to distinguish which is which takes some some some knowledge and some learning, but you really can interpret events in the world unless you have a firm background in probability statistics.
And I'll give you another little anecdote that like the Russians in 1958, you know, had Sputnik and we were afraid that we're going to beat us to the moon. And they did beat us to the moon, but not a man on the moon. So the United States put in a science program where everybody has to learn calculus. I've heard about that. Okay, so we all got to learn calculus, which we can't let the Russians beat us. So now everyone has to, you know, get into a good college, you're going to have to learn calculus. Again, to med school, you have to learn calculus, which is absurd, but you're never going to use it.
But no one learned how to use probability statistics, because it was not it was considered secondary to calculus. So we have a country that sort of knows, you know, fair amount of calculus, but very little probability statistics. And it's just not the way it's just not what's necessary to be a good decision maker to be a good to be a good citizen, but it's almost impossible to change these things. So you have to take the effort yourself to make sure that you are literate in probability statistics.
And you certainly understand Bayesian analysis, because there's you know, all these studies done that they as Harvard kids in Harvard medical school who are going to be researchers, some basic questions after they got the data about about diseases and they were off by a factor of a hundred. I need to very, very smart people, but they didn't know Bayesian analysis and they were and they were ridiculous, but they was not taught to them in medical school. And if you've ever had the frustration of talking to a doctor and saying, Doc, what's my chance of having this? He goes, Oh, I don't know. Oh, you may or may not have it. It's like, I'd like you to I'd like you to tighten that market up a little bit, Doc, but they're not trained that way.
And that's a tragedy. You have to make sure that you go out of your way to get that to get that training. I think that's a very valid point. And I'm currently learning calculus. I think I might get to do a little statistic education on my own. Calculus is wonderful. It's my favorite subject and it's great. It's beautiful. It's art. It's that's key to science and everything like that. But it's it's of limited value to most people.
And I want to ask one more question that I do at the end of every interview. I think I asked it to 39 people so far. I'm 16 right now. If you would to give one piece of advice to a 16 year old today, it can be life advice, career advice, even romantic advice. What would it be? I take it as romantic advice. I mean, I believe in markets. It's like, don't go out with somebody that your friends think is a nutcase. But you can get caught up. And if you say to my friends, be honest, I won't punish you, try and do it anonymously. Give me your marketplace. Am I making a gigantic mistake?
So many lives are ruined because you get involved with the wrong person. And no one wants to speak up. So you got to come up with a mechanism. Hey, friends, you're my friends. I trust you. And I'll do this anonymously. You know, should I should I know? Is this person too nutty for me to be going for me to going out with you could prevent a lot of horrible relationships from happening there? That would be my number one advice because one of the things that we do in reverse, the bigger the decision, the less time we think about it.
You know, if you're buying a selling a stock and it's basically irrelevant what you're doing with the markets of fair, you'll spend a lot of time on it. If you're deciding who to marry or who to have a relationship or whatever, as you basically just plop into it without much thought. And one has a gigantic impact on your life. And one has a very small impact on your life. Yet we spend much more time worrying about the minor things and not enough time worrying about the big things.
I mean, from a living life experience, I think I agree. And I recommend anyone who's listening to this to listen to my episode with any you about decision making. I think it's an excellent compliment to this episode. But Jeff, it was an absolute pleasure having you. Thank you for coming on. I really appreciate it. Good luck. I really appreciate it. It was fun. Okay. Bye.