I've been thinking a lot about prediction markets lately, and the more I think about it, the more something feels off.

Everyone talks about how powerful prediction markets are—beating polls, outperforming experts, crushing traditional forecasting tools in the 2024 U.S. election. Suddenly, platforms like Polymarket are hailed as “truth discovery machines.” Sounds great, right? Markets aggregate dispersed information, people put real money behind their beliefs, and ultimately, the prices converge to the truth.

But here’s the problem. Last year, a trade made me realize what “getting the early water’s edge” really means.

Someone bet $30k on a prediction market that Venezuelan President Maduro would step down by the end of the month. At the time, the market priced that probability very low—seemed like a bad trade. But what happened? A few hours later, the police arrested Maduro. That account closed out with a profit of over $400k. The market was right, but that’s precisely the issue.

If the market is accurate, it’s because someone has access to information that everyone else doesn’t—so it’s not really about “discovering the truth,” but about monetizing an informational advantage. This isn’t “collective intelligence,” but blatant information asymmetry.

Supporters might say, if someone trades on insider info, the market will react earlier, helping others. Sounds ideal, right? But in reality? If a market’s accuracy depends on leaked military operations or government secrets, then it’s not an information market—it’s a black market. There’s a fundamental difference, but many choose to ignore it.

And it gets even more absurd. The Zelensky suit event is a textbook example. In 2025, a prediction market had a question: Will Ukraine’s president wear a suit before July? Hundreds of millions of dollars in volume. Zelensky appeared publicly in a black jacket and trousers, and media and fashion experts said that was a suit. But guess what? Because a big trader held a huge short position, they enforced a settlement that benefited their own interests. The system ran perfectly—except what was “discovered” wasn’t the truth, but who had more money.

This isn’t a failure of decentralization; it’s a failure of incentives. When you flood the system with large capital, ambiguous language, and unresolved governance, these kinds of outcomes become inevitable.

Honestly, we’ve overcomplicated this. Prediction markets are simply places where people bet on future outcomes. Win, make money; lose, lose money. All those words about “truth,” “information discovery,” “collective intelligence”—they’re just embellishments. Your profit isn’t because you have insight; it’s because you bet correctly.

This “masking” is the root of the problem. When platforms claim to be “truth machines,” every dispute becomes an existential crisis; but if you’re honest that it’s a high-risk financial product, disputes are just disputes, not philosophical crises. Acknowledging this can actually lead to clearer regulation and more rational design.

Prediction markets themselves aren’t the problem—they’re honest ways to express beliefs amid uncertainty. But we shouldn’t pretend they’re more sophisticated than reality. They’re just financial tools linked to future events, nothing more. Once you accept that you’re running a betting product, you won’t be surprised when betting behaviors occur.