What Is a Prediction Market?

A market where you trade on the probability of real events. Not a bet. Not a poll. Something more interesting than both.

A prediction market is a platform where people buy and sell contracts tied to real-world outcomes. The price reflects the crowd's probability estimate for that outcome. When a contract trades at 68 cents, it means traders collectively put the odds at 68% - not because someone decided that, but because thousands of people with real money at stake pushed it there.

You've probably seen a headline like this: "Polymarket gives Trump 67% odds of winning." Or maybe a friend mentioned they made money "trading on the election."

You had questions. What does 67% actually mean? Who decided that number? Is this gambling?

Here's what's actually happening - and why it's worth understanding even if you never trade a single dollar.

The One-Sentence Definition

A prediction market is a platform where people buy and sell contracts tied to real-world outcomes. The price reflects the crowd's probability estimate for that outcome.

That's the whole thing. Everything else is mechanics.

The key word is price. When a contract for "Will the Fed cut rates in March?" trades at 68 cents, it doesn't mean some analyst typed in 68%. It means thousands of traders - each putting real money behind their belief - have pushed the price to that level through buying and selling. The crowd set that number. And the crowd is accountable to it.

The contract price and the implied probability are the same number. A 68-cent YES contract means the market believes there's a 68% chance. If you genuinely believe the true probability is higher, buying YES gives you a positive expected value over time. If you're wrong about your edge, you lose money.

That feedback loop - profit for accuracy, loss for error - is what separates prediction markets from a casual opinion.

What the Price Actually Means

The price is a probability estimate. Nothing more, nothing less.

A contract at 63 cents implies a 63% chance the event happens. It does not mean the event is likely to happen. It does not mean it will happen. It means that, right now, the collective judgment of everyone trading this market puts the odds at 63%.

Compare this to a poll. A poll asks people what they think will happen. A prediction market asks people to put money behind what they think will happen. The financial stake filters out lazy opinions and wishful thinking. People reason more carefully when they're accountable to the outcome.

That's the underlying idea: prices encode more honest beliefs than words.

When a price moves, the crowd's collective belief has changed. Something shifted - a new data point, a new development, a new risk. The price is updated in real time. A traditional poll is a snapshot taken once a month. A prediction market is a live feed.

Types of Prediction Markets

Binary markets are the most common type you'll encounter - yes or no, one side wins $1, the other wins nothing.

Multi-outcome markets cover events with more than two possible results. "Who wins the 2028 Democratic primary?" might list eight candidates, each with their own tradeable contract. The prices across all candidates should roughly sum to $1, since exactly one of them will win.

Scalar markets pay out on a sliding scale. Instead of yes or no, you might trade on where exactly inflation lands - above 3.5%, above 4%, above 4.5% - with each threshold carrying its own market.

Play-money markets, like those on Manifold Markets, use virtual currency. No financial risk, but also a weaker accuracy signal. Useful for learning the mechanics before committing real money.

A Short History

Prediction markets aren't new. The idea traces back to the economist Friedrich Hayek, who argued in the 1940s that prices encode dispersed knowledge that no single person or institution could ever fully collect.

The first structured modern prediction market was the Iowa Electronic Markets (IEM), launched in 1988 by the University of Iowa. It allowed small-stakes trading on US elections and became a major academic dataset for studying whether markets could outperform polls. They often could.

In the 2000s, major corporations ran internal prediction markets for the same reason. Eli Lilly used them to forecast which drug candidates had the best chance of advancing through clinical trials. Google ran them to forecast product launch dates and office openings, using employee trades as a forecast.

The modern public era began with platforms like PredictIt, and then Polymarket and Kalshi - expanding the scope from politics into macroeconomic data, sports, weather, crypto, and cultural events. According to PredictionCircle's topical research, the industry grew from roughly $15.8 billion in trading volume in 2024 to $63.5 billion in 2025, driven by the 2024 US election cycle.

The technology changed. The underlying idea - aggregate distributed knowledge into a price - has been the same since 1988.

What Prediction Markets Get Right - and Where They Fall Short

The research is generally positive. A 2020 meta-analysis published in the ICEB proceedings found that prediction markets tend to outperform alternative forecasting methods - polls, expert panels, Delphi surveys - when studied in controlled settings. They're not always better, but the meta-analysis found prediction markets substantially outperformed Delphi panels and expert surveys across studied settings.

That said, they have real limits.

They're probabilities, not predictions. A market at 70% will be wrong roughly 30% of the time. That's not a flaw - that's the whole point. A well-calibrated market should be wrong exactly as often as its prices imply. When people say "the market was wrong about X," they often mean the lower-probability outcome occurred. That's expected to happen.

Thin markets can mislead. In low-volume markets - niche events, obscure questions - a small number of traders can dominate price formation. The signal weakens when fewer people are participating. On Polymarket's largest markets (major elections, Fed decisions), liquidity is deep and prices are more reliable. On newer or niche markets, treat prices with more skepticism.

Manipulation is possible in small markets. A motivated actor with relatively little capital can move prices in a thinly traded market. Informed traders typically correct the price - but the window of distortion can last hours or longer in illiquid markets.

Access isn't universal. Polymarket is unavailable to US residents due to regulatory restrictions. Kalshi requires identity verification. Crypto requirements on some platforms create friction. The market may be more accurate than a poll; it's also more complicated to participate in.

Prediction markets show you what informed people with money at stake currently believe. That's more signal than a poll. It's less than a crystal ball.

The Platforms You'll Encounter

Kalshi operates as a federally regulated event-contracts exchange under CFTC oversight. It lists markets on macroeconomic indicators (inflation, interest rates, employment), weather, and politics. Fiat currency (USD), straightforward onboarding, published contract rules. The most regulated option in the US market.

Polymarket is a crypto-native platform with global reach, using USDC (a dollar-pegged cryptocurrency) on the Polygon network. Unavailable to US residents. Fastest to list new markets, broadest range of topics, highest trading volume globally. Requires a crypto wallet to participate.

PredictIt focuses primarily on US political events. Fiat currency, higher fees than competitors, more limited market selection. In 2022, the CFTC withdrew PredictIt's no-action letter, nearly shutting it down. It continues to operate under a legal cloud.

The Bigger Picture

There's a reason serious institutions - from the University of Iowa to Google to Wall Street trading desks - have used prediction markets for decades before they became a consumer product.

When people put money behind their beliefs, they reason more carefully. When thousands of people do it simultaneously, the aggregate price encodes more information than any single analyst, any poll, any pundit can hold alone.

That number on the screen - 63%, 41%, 78% - isn't a guess. It's the current consensus of everyone with enough conviction to back it up financially.

That's worth knowing. Whether you ever trade or not.