SIG Founder: Why Do I Believe in Prediction Markets?

Source: Generating Alpha Podcast

Compiled by: Jiahuan, Chaincatcher

Who is Jeff Yass?

This week on “Generating Alpha,” we have a unique guest - Jeff Yass, the founder of Susquehanna International Group (SIG), one of the most successful trading firms in the world.

Jeff is a legendary figure in the financial world, applying the principles of poker, probability, and decision theory to the markets. Over the past forty years, he has quietly built a global trading giant, discreetly active behind the scenes on Wall Street, trading everything from options to cryptocurrencies, all based on mathematical precision and rational thinking. He is also one of the most influential yet most enigmatic and low-profile figures in modern finance, and this episode marks his very first podcast interview.

In this episode of the short series, we only talked about one thing: prediction markets—why Jeff believes they are the future of humanity's quest for truth, how they can improve business and government decision-making, and the immense power they reveal in incentives, information, and human behavior.

Recording this episode was a great pleasure for me, and I hope you enjoy it just as much after listening.

Host: Jeff, thank you so much for being here, and thank you for taking the time.

Jeff Yass: It's my honor, Amir, let's get started.

The Core Value and Significance of Prediction Markets

Host: To lay a foundation for this conversation, I would like to start by asking: What is your overall view on prediction markets right now? How important are they to you and SIG?

Jeff Yass: Prediction markets have been our true passion for many years. They bring immense value to the world. You can't make good decisions without accurate probabilities of events occurring. And prediction markets are currently the most accurate way we know to estimate these probabilities. Therefore, we believe this is a remarkable tool that will bring great benefits to society.

Host: From a macro perspective, how do you think the prediction market will evolve in the next decade, especially regarding regulation and gambling legislation?

Jeff Yass: To be honest, we are not entirely sure in the traditional betting field. But the world is gradually realizing that exchange models like Betfair in Europe, which allow people to buy and sell from each other, are a fairer system that can significantly reduce costs. Currently, the vig in traditional betting is around 5%, but if people can trade with each other on an exchange, we believe it can be reduced to 1-2%. This is a huge victory for those who want to participate in sports betting.

But our true motivation for promoting prediction markets is to uncover the truth. Our favorite example is the Iraq War. When Bush first invaded, he said it would cost only 20 billion dollars, while his economic advisor Lawrence Lindsay estimated it could reach 50 billion. As a result of telling the truth, he was punished. The actual figure later calculated was between 2 trillion and 6 trillion dollars.

If there had been a prediction market at that time asking, “What is the total cost of this war over/under?” I dare say it wouldn't be 2-6 trillion, but it would definitely be far higher than 50 billion, possibly up to 500 billion. Then the common people would see this number and say, “Wait, we don't want this war! Politicians always say wars will be quick and cheap, but that's never the case!” So we need a credible source, and prediction markets are an objective and trustworthy source—because if the bettors are wrong in their judgment, they will lose money.

If such terrifying numbers were seen at that time, the anti-war voices would be much louder. Predictive markets can really be powerful enough to slow down the lies that politicians continuously tell us, and that is the main reason I want to see it thrive.

Host: This is almost the “truth of the people”, rather than the polluted, watered-down version fed to the public.

Jeff Yass: Absolutely correct! And it's not just for the average person, but even for experts as well. You and I may not know how much a war truly costs, but there's always a small group of people who do know, and they will place bets to push the price to a reasonable level. The average person has no way of knowing the cost of war, but when you see experts battling in the market, voting with real money, you can trust that number. Just by looking at the prices in the prediction market, you can be more professional than those politicians who either make up numbers or deliberately lie.

Manipulation and Risk

Host: I guess that prediction markets will be used in the future to price more financial tools and support other decisions. But how can we prevent prediction markets from being manipulated?

Jeff Yass: It's similar to preventing any market from being manipulated—if you want to manipulate prices, as long as there are enough participants in the market, you have to take a huge loss yourself. For example, if you want to drive the cost of the Iraq war down to “below 50 billion,” fine, we can bet several hundred million against you, saying you're wrong. Your manipulation plan would be outrageously expensive, possibly hundreds of times more than running a few misleading ads (ads only cost a few million, this would be several hundred million). So this itself protects the integrity of the market.

Host: Move forward a bit, you were an early professional gambler, playing poker and betting on horse racing. What similarities do you see between gambling and prediction markets? What systemic risks and opportunities does this bring?

Jeff Yass: I really don’t see any systemic risk. What I see is more truth and a more rational and objective probability entering the market. The real systemic risk is politicians using lies to deceive us, and the prediction market is precisely the antidote. Of course, there may be a small amount of manipulation, but it is negligible compared to the amount of manipulation we are facing now. A competitive market will smooth out any issues.

Business Applications and Hedging

Host: So overall, how do you think companies like yours will integrate prediction markets into everyday decision-making in the future?

Jeff Yass: For example, in 15 days, New York City will have an election (the podcast was released on October 23). If you only watch TV and read the news, it’s hard to judge the real probability—some say “it's too close,” while others say “New York could never elect someone like Maami.” But if you look at the prediction market, he has over a 90% chance of winning. If you need to decide whether to move to New York or relocate your company there, you must know this probability; just reading the newspaper or listening to the news won't clarify things. Having this clear number will greatly assist your decision-making.

For example, if you are a real estate developer and you believe that your property will drop by one million dollars after Maami comes to power, you can directly hedge against that. More importantly, you can instantly get the most reliable probabilities without having to read a million articles or call polling companies; all the work has already been done for you, and you get the best numbers directly to guide all your decisions.

For SIG, we have also been observing the probabilities of the presidential election, how the stock market fluctuates based on who is leading, and we use the probabilities from the prediction market to determine whether a particular stock is overreacting or underreacting to political events.

Host: I can imagine that as the trading volume in prediction markets continues to grow, large institutions will also begin to participate, using prediction markets instead of traditional financial tools for hedging. You recently collaborated with Kalshi and became one of their main market makers. How do you think the participation process of companies like yours will evolve as the market develops?

Jeff Yass: The prediction market is still a customized product; institutions have not really entered yet, and the trading volume mainly comes from relatively small players. No giant institution has come in to place large bets on events like “Will the Federal Reserve raise interest rates?” But we believe that as regulation becomes clearer and the market becomes more popular, institutions will flock in, leading to Wall Street-level huge bets. Currently, Goldman Sachs and Morgan Stanley are still a bit cautious, but they will change their stance sooner or later.

What I truly look forward to is how prediction markets can impact the insurance industry. In many places, insurance is simply unavailable; the government has driven prices too low, causing insurance companies to leave—like in Florida. But if you use prediction markets for insurance, you could have a contract that says: “Will the wind speed in your area exceed 80 miles per hour in the next 48 hours?” Assuming a 10% probability, if you're worried about your house being destroyed, you could bet $10,000 to win $90,000, effectively covering your losses. Moreover, you only buy when there's a threat of a typhoon, eliminating all claims, advertising, and operational costs, making it much cheaper and fully customizable to your needs.

Host: And it's much more quantifiable! Traditional insurance companies always want to assess how much you need and how much they will compensate you, etc., while the prediction market is clear at a glance. As these markets eventually become fully regulated exchanges, do you think liquidity will primarily come from Wall Street institutions or retail investors?

Jeff Yass: Both will. And they will create enormous opportunities. For example, if you are an enthusiastic weather fan living in Florida and you have a deep understanding of hurricane probabilities, you can put your expertise into the market and say, “I think the disaster probability for this area is this much.” In the past, such specialized knowledge could never be monetized, but now you can make money from your expertise and also help ordinary people lower their insurance prices.

Impact, Barriers, and Learning

Host: Do you think that predicting the market in the future will affect the outcome of the events themselves?

Jeff Yass: No. For example, the French guy who was crazy buying Trump on Polymarket before was just nonsense. We bet against him directly, he drove the price up, and we pushed it back down, it really didn't affect the outcome. This kind of concern is not zero probability, but it has been seriously exaggerated.

Host: So what do you think is the biggest obstacle to the widespread adoption of prediction markets right now? How can it be eliminated?

Jeff Yass: The biggest obstacle is that, like the questions you are asking, you can see where things might go wrong. These thoughts immediately come to mind psychologically — that things might go wrong. Some things might go wrong, but things are already going wrong now. So, as we get used to it, that obstacle will disappear. It may take time, but people have fears and they tend to exaggerate negative impacts. However, as the product is accepted and people understand its value and how much it can save them, those fears will dissipate. It may take years, but I am very optimistic about our ability to reach our goals.

Host: Some people are concerned that certain decisions should not be quantified. What decisions or predictions do you think we should intentionally avoid quantifying?

Jeff Yass: Good question. Theoretically, you could even set up a market: “Should I marry this girl?” Maybe your friends and family would be more objective than you… but that does seem a bit excessive. So my answer is basically no.

Host: What is no one talking about now, what can prediction markets achieve in the future?

Jeff Yass: The most important point: it can prevent wars. In every war, politicians exaggerate, saying it will end quickly, cost very little, and have few casualties—all lies. During the American Civil War, the Lincoln administration temporarily halted conscription in 1862, thinking the war would be over in a few weeks, resulting in the loss of 650,000 lives. If the public knew the true costs and disastrous consequences in advance, they would desperately seek alternatives to war.

Another example is self-driving cars. Many people oppose them now because they can imagine robots going out of control and killing people. But in the next 12 months, about 40,000 people will die on American roads; if we fully implement self-driving technology, I guess it could be reduced to 10,000, saving 30,000 lives. If the prediction market shows that traffic deaths will significantly decrease by 2030, policymakers will rush to accelerate the deployment of self-driving cars—because we can clearly see how many lives can be saved. Right now, everyone is still hesitant: “I don't know if it's good or not.” Once there is an objective number, we will move much faster.

Jeff Yass wants to say a word to the world

Host: Before we conclude, if you could only convey one message about prediction markets to the entire world, what would you say?

Jeff Yass: My mother used to tell me, “If you are really that smart, why haven't you made a fortune yet?”

Prediction markets are objective. If you think the market probabilities are wrong, then go bet and correct them to where you think they are right. If you are truly smarter than the market, you will make a lot of money while also helping society get the prices right. If you can't make money, then maybe you should be quiet—perhaps the market understands better than you.

This will drive all university professors crazy because they want to be experts, but they are not. A group of speculators who risk real money every day will be much stronger than any professor. Driving professors crazy is, in my opinion, a good thing.

For example: When my daughter was 12 years old, Obama was competing against Hillary in the Democratic primary. At that time, the most famous political scientist in the U.S. appeared on TV and said, “Hillary is leading by 30-40 points, it's settled.” I had my daughter check TradeSports (the only prediction market at that time), and she said, “Obama has a 22% chance of winning.” The market had already recognized Obama's uniqueness and personal charm, while Hillary, despite her high profile, had a lead that was actually meaningless. My 12-year-old daughter was more accurate than the top political experts in the world. This is the power of prediction markets.

Learning and Life Advice

Host: Last question: If you were a high school student right now, based on your years of success and recruitment experience, what do you advise today's young people to study?

Jeff Yass: Of course, you need to learn computers, and you must understand programming and AI. But if you really want to become someone who makes decisions under uncertainty—which is the essence of being human—you must master probability and statistics.

There are too many decisions in the world made under uncertainty, and if you do not understand the mathematical foundations of probability and statistics, it is easy to make disastrous decisions. For example, hurricane season has arrived, and there are many hurricanes—does this matter? Has there always been this many in previous years? How much variability is there? Is this evidence of global warming, or just random fluctuation? Distinguishing between signal and noise requires knowledge.

In 1958, the Soviet Union launched Sputnik, and the United States feared being left behind in the race to the moon, so the entire country promoted calculus for everyone. As a result, now to get into a good university or medical school, one has to learn calculus—it's absurd because 99% of people will never use it in their lives. However, probability and statistics are treated as secondary subjects, and no one is required to learn them. Thus, our country has a lot of people who know calculus, but almost no one understands probability and statistics, which is completely backward.

You must take the initiative to learn probability and statistics, and you must understand Bayesian analysis. Students at Harvard Medical School can get disease probability questions wrong by a factor of 100—these people are super smart, but the school didn't teach them. Even if you ask a doctor, “What do you think the probability is that I have this disease?” he will only say, “It might be possible, it might not be.” You might think: Doctor, can you tighten up the market a bit?

Host: I am currently learning calculus… It seems I need to supplement my statistics myself.

Jeff Yass: Calculus is beautiful, it is my favorite subject, it is art, it is the key to science. But for the vast majority of people, its practical use is limited.

Host: The last traditional question (I have already asked 39 people, and I am now 16 years old): If you were to give today's 16-year-olds one piece of life advice (it could be about career, relationships, or any aspect), what would it be?

Jeff Yass: If it’s relationship advice — I trust the market. Don’t date someone that all your friends think is crazy. Many people get trapped in that, and at that point, you need to ask your friends for the truth. You can do it anonymously and have them create a small prediction market: “Am I making a huge mistake by dating this person?” How many people have been ruined by the wrong person their whole life, just because no one dares to tell the truth. You need to design a mechanism to bring the truth to the surface.

We have a problem: the bigger the decision, the less we think about it. We might ponder for half a day over buying a single stock (which actually has a negligible impact), but decisions that affect our entire lives, like whom to marry or who to date, are often made in a muddled way. We have completely reversed the allocation of our time.

Host: I have little life experience, but I completely agree! I also recommend everyone to listen to my episode with Annie Duke about decision-making, it's a perfect match with this one. Jeff, thank you so much today!

Jeff Yass: Good luck to you, I am also very happy, goodbye!