Prediction Markets: Forecasting Tool or Gambling?

Prediction markets have become one of the most talked-about — and fastest growing — corners of the financial internet over the past year.

During the 2024 US election alone, more than $3.3 billion was wagered on Trump versus Harris. By early 2026, prediction markets were processing over $22 billion a month — an 11x increase in volume in under a year.

Supporters insist these are not gambling platforms. They describe them as something closer to financial markets: a place where people buy and sell probabilities about the future, generating forecasts that are often more accurate than polls or expert panels.

The distinction they draw is simple. Gambling is entertainment. This, they say, is a forecasting tool.

So let's unpack it.

In a traditional sportsbook, the bookie sets the odds. If you want to bet on a football team, the price is already posted, so you can accept it or walk away. The bookmaker takes the other side of every bet and makes money from the spread built into those odds.

Prediction markets work differently. Instead of betting against a house, traders take positions against each other. If someone believes a particular outcome is likely, they buy a contract that pays out if it happens. Someone else has to sell that contract, effectively betting against it. So prices move as people trade back and forth — just like a stock.

Advocates argue this structure makes prediction markets fundamentally different from a casino. Prices emerge from supply and demand rather than being set by a bookmaker, and in theory the market price becomes a real-time probability estimate.

A contract trading at 60 cents means the market believes there's roughly a 60% chance of that event happening. And because traders are putting real money behind their beliefs, the argument goes, they process information more carefully than a pollster or pundit with nothing at stake.

That financial incentive is supposed to produce better forecasts. And in some cases, it has. Polymarket gave Trump a meaningfully higher probability of winning the 2024 election than most national polls, which had the race essentially at 50-50 throughout.

But the accuracy argument has limits. Across more than 2,500 markets and $2.5 billion in volume, Polymarket correctly predicted outcomes only 67% of the time. Not dramatically better than chance on many questions.

And the problem ran deeper than the number. When new political information emerged, prices often didn't react. Or they spiked one day and reversed the next. Visibility and hype were moving prices more than actual news. Which is a fairly significant problem for a tool whose entire value proposition is that money makes people think more carefully.

But even setting that aside — even if the forecasts were consistently good — usefulness doesn't automatically make something not gambling. A lottery that funds schools is still a lottery.

Then there is the claim that prediction markets reward skill, not luck. Someone with better information or stronger models can profit by identifying mispriced probabilities, and in that sense, supporters say, prediction markets resemble financial trading rather than gambling.

The problem is that the same argument applies to professional sports bettors, poker players, and horse racing analysts. Skill exists there too. It doesn't automatically change what the activity is.

And finally, there’s the argument that prediction markets let people protect themselves against real-world outcomes.

A political consultant whose entire client base could disappear if the wrong candidate wins could buy contracts on that candidate winning. If they do win and he loses clients, at least he makes money on the contracts.

The problem is that almost nobody is actually doing that. The vast majority of people on these platforms are just speculating on outcomes because they think they know what's going to happen.

The CEO of Kalshi has publicly described his long-term vision as wanting to "financialise everything and create a tradable asset out of any difference of opinion."

Not just elections or economic policy. Literally anything two people might disagree about. That's a very different pitch from "we help businesses manage risk."

And the revenue breakdown backs that up. If prediction markets were genuinely being used as risk management tools, you'd expect most of the money to come from people hedging real economic or political exposure. But 89% of Kalshi's revenue comes from sports betting.

Prediction markets may use financial language, peer-to-peer mechanics, and academic research. But at their core they still involve people risking money on uncertain future events, which is, by most definitions, gambling.