Transcription
Who will Donald Trump insult next? Will GTA VI really be released on time? Will the Ukraine war end this year? Or will China attack Taiwan? Welcome to the world of prediction markets. Here you can bet on almost anything: Those who bet correctly can potentially achieve high profits. But not everyone plays with the same cards. "The vast majority lose money on prediction markets." In this video, we'll look at why so many people are suddenly betting on the future, how these platforms really work – and whether we can make money with them. I'll put 100 in now, because: No risk, no fun. I'm glad I didn't bet more. It's best if we look at an example. For that, we'll go to the platform Polymarket. It is blocked in many European countries and only accessible through detours. First, the many bets on political topics stand out here. The first bet I see is whether the British Prime Minister will be deposed or not. Here I now see the different data that is being traded. The next bet is about USA–Iran, so about the peace agreement. In the past, such prediction markets were an experiment for economics nerds at universities. Today, they are marketed much more broadly and aggressively – especially to a young, online-savvy audience. Kalshi and Polymarket are currently the two best-known providers. And they are growing rapidly. Every month, bets worth more than 25 billion dollars are traded on these platforms alone. That's ten times as much as a year ago. This brings the platforms a lot of money: They earn primarily through transaction fees. This means: Whenever users buy or sell on their markets, Polymarket or Kalshi profit from it. However, the founders of both platforms also see their platforms as "truth machines." "Prediction markets provide a societal added value because they make accurate predictions. These could be questions about economic development, political development, demographic, even technological development. These are all important questions for our society, for which it is very valuable if we have good future scenarios that we can adjust to and prepare for." We want to try our luck with a bet on US politics. A speech by Trump is very likely to be the best bet. The next speech Trump will give will probably be during his meeting with Xi Jinping in China. Therefore, we are now betting on what Donald Trump will say during his stay in China. So, here I can now bet on individual words: Tariffs, Ships, Crypto, Six Seven, Hormuz, Taiwan. I don't understand what Peng means. IQ, Soybean, North Korea, Rare Earths. "Well, I'm actually not the right person to give advice, because I usually lose when I bet myself. I would definitely not bet money on 'Crypto/Bitcoin'. I would rather bet on 'Tariffs', 'Pandemic', or 'Covid'. Words for which you have past experience. I have of course also done a quick search and I would choose the word... Ah, I think I'll no longer take 'Fentanyl', but... Yes, I'll take 'Fentanyl'. Donald Trump has long accused China of smuggling Fentanyl into the USA, and I've also read that Fentanyl could be a topic at the meeting. And therefore, I believe he will say the word." Currently, a "yes" share costs 30 cents. That is significantly cheaper than a "no" share, because most users do not believe that he will say the word. I'll bet 100 dollars on it now, because: No risk, no fun. So, if Donald Trump says the word "Fentanyl," I will win 270 dollars. But if Donald Trump does not say Fentanyl, I will lose the entire 100 dollars. Now we have to wait and then watch live how Donald Trump will say the word: Fentanyl. Our bet is only about a single word. Often it's about much more: for example, about wars, elections, or court decisions. So, about real events with real consequences. And when such events become tradable, a problem arises: Some people not only try to predict what will happen, but want to influence the event itself. A particularly absurd example takes place in April 2026 in France: A user on Polymarket bets on a sharp temperature increase at Paris airport. The chance of this is less than 1 percent at that time. Nevertheless, the user bets 120 dollars that the temperature will rise above 18 degrees Celsius. This weather station at the airport is decisive. Shortly after the bet is placed, the temperature suddenly rises very briefly from 16 to 22 degrees – and then drops again. The user wins 21,000 dollars in the meantime. The suspicion: Someone might have deliberately heated the sensor, for example with a hairdryer or a lighter. This has not been proven yet. But the French authorities are investigating. The second major problem is insider trading: When a market participant knows more than others. This is exactly what happens in January 2026, when Venezuelan President Nicolás Maduro is overthrown by the USA. Shortly before his arrest, a Polymarket user bets a total of around 34,000 dollars that Maduro will be deposed. He places the last bet hours before the attack on Venezuela begins. This is shown by the transaction history. In total, the account makes more than 400,000 dollars in profit. It later turns out: The account is said to belong to a US soldier who was involved in Maduro's abduction with his special unit. He naturally knows about the plans in advance. On traditional financial markets, such insider trading is strictly regulated: Anyone who uses non-public information to make money can be prosecuted. The situation is more complicated on prediction markets. Insider trading is not a flaw in the system here, but part of the principle. "Actually, on prediction markets, you want to aggregate as much information as possible. And therefore, insider trading is actually good for predictions. Insider trading is of course a problem when other actors are harmed. So, if, for example, people on Polymarket have more information than normal participants, then the normal participants lose money simply because others have information that the public cannot have." Not only that: On prediction markets, insider trading can even be life-threatening under certain circumstances. For example: If information from military operations reaches prediction markets, enemy intelligence services can observe course anomalies and prepare for them. Now a few days have passed and Donald Trump's state visit to Xi Jinping is over. Let's see if he said the word "Fentanyl." Okay, but it's faster if I just look at the transcript of the speech and search for "Fentanyl." He didn't say it anywhere here. We can now go to Polymarket. I can already see, my money is gone. The portfolio was once more than 100 US dollars. Now I have 2.37 US dollars here. It shouldn't look like this if I had won. Let's go to it and see directly: We have lost. Apparently, Donald Trump did not say Fentanyl anywhere. At least not in the official recordings. I have lost my money – and most users on these prediction markets are in the same situation. "It's only a very small part and probably already quite professionalized actors who make profits on these large prediction markets. The vast majority lose money on prediction markets." An analysis by the Wall Street Journal shows how unevenly profits are distributed: Only 0.1 percent of Polymarket accounts collect two-thirds of the total profits. Professionals with better data and insiders skim the money from amateurs. Polymarket and Kalshi like to market themselves as truth machines. As platforms that bundle the knowledge of the masses. According to this study, however, the masses are unimportant for correct predictions on Polymarket. A small, professional minority of users is decisive. And these are also the ones who make the big money, while everyone else believes they have a fair chance. Prediction markets, AI, and all other questions of the future – that's also what our Quantensprung newsletter is about. Sign-up link here: