The global guide to
prediction markets
From Kalshi's CFTC-regulated event contracts to Polymarket's on-chain liquidity and Manifold's play-money community — we track every major operator, how they're regulated, and how they actually work, across every major market in the world.
What is a prediction market?
A prediction market is an online platform where people bet real money on the outcome of a future event with a binary, yes-or-no answer that resolves by a specific date — an election, a central bank rate decision, an Oscar winner, whether a stock index closes above a given level, whether a team wins a tournament. The mechanism is a financial instrument called an event contract, which has a nominal value (typically $1) and pays out in full to whoever held the correct side when the event resolves.
A worked example: imagine an event contract on whether the S&P 500 will close above 7,000 by year-end. If you buy 1,000 YES contracts at $0.25 each and the index does close above that level, you collect $1 per contract — turning $250 into $1,000. If you're wrong, the contracts pay zero and you lose your $250. The price you pay is, in effect, the market's current estimate of how likely the event is.
How pricing actually works
YES + NO = $1
YES and NO shares always sum to roughly $1, because anyone holding one of each is guaranteed exactly $1 at resolution. A YES share at $0.62 means the market puts the odds at about 62%.
Orderbook vs AMM
Kalshi and Polymarket use orderbooks where limit orders match when prices cross. Older on-chain venues use automated market makers — smart-contract liquidity pools that always quote a price, even in thin markets.
Settling the market
Winning shares pay $1, losing shares pay zero. Resolution can be handled by a centralized operator, a decentralized oracle like UMA, or a community vote, depending on the venue.
Types of prediction markets
Not all prediction markets work the same way. They differ along three main axes — how the question is structured, how liquidity is provided, and whether real money is at stake.
Binary vs scalar vs categorical
Binary markets have two outcomes (YES/NO) and are by far the most common. Categorical markets have multiple mutually exclusive outcomes (e.g. which party wins an election). Scalar markets let traders bet on a numerical range — will GDP growth land between 2.0% and 2.5%? — and pay out proportionally.
Orderbook vs AMM
Orderbook venues like Kalshi and Polymarket match buyers and sellers via limit orders — tight spreads in liquid markets, nothing quoted in dead ones. Automated market makers (AMMs) use smart-contract liquidity pools that always quote a price, making them better for long-tail questions but worse for large trades.
Real-money vs play-money
Real-money venues (Kalshi, Polymarket, ForecastEx) use actual USD or stablecoins and face the most regulation. Play-money platforms (Manifold's mana, Iowa Electronic Markets' $500 cap, Hollywood Stock Exchange) operate outside gambling law but attract real forecasters who care about their track record.
Centralized vs on-chain
Centralized exchanges hold USD at a regulated custodian, settle in fiat, and resolve markets by operator decision. On-chain venues settle in stablecoins, resolve via decentralized oracles like UMA, and are accessible globally via a wallet — at the cost of adding crypto custody risk.
A short history of prediction markets
Modern prediction markets trace back to the Iowa Electronic Markets, launched at the University of Iowa in 1988 as a research project that has consistently outperformed major election polls. Online play-money markets like Hollywood Stock Exchange followed in the 1990s. Intrade, an Irish real-money exchange, became famous in the 2000s for forecasting US elections before shutting down in 2013. Kalshi launched in 2021 as the first fully regulated event-contract exchange in the US, and Polymarket — built on Polygon and settled in USDC — emerged as the dominant on-chain venue worldwide, processing billions of dollars in volume around the 2024 US election. In October 2025, Intercontinental Exchange (the company that owns the New York Stock Exchange) announced an investment of up to $2 billion in Polymarket, signaling that prediction markets have moved firmly into the financial mainstream.
Major operators worldwide
The four platforms that define the modern global prediction market landscape. Each takes a fundamentally different approach to settlement, regulation, and liquidity.
Polymarket
USDC-settled on Polygon. Largest prediction market by volume.
- Deepest liquidity
- Broad market coverage
- On-chain settlement
Kalshi
First fully CFTC-regulated US event-contract exchange.
- Fully regulated in US
- Bank-grade compliance
- Economics & weather contracts
Manifold Markets
Free play-money social prediction platform with user-created markets.
- Free to use
- User-created markets
- Forecaster community
Myriad Markets
Prediction markets embedded into publisher experiences.
- Publisher partnerships
- Gamified UX
- Multi-chain
What people actually trade
Elections & policy
Presidential winners, parliamentary majorities, cabinet picks, court rulings, foreign election outcomes — politics has historically been the largest category by volume.
Macro data
Will the next inflation print exceed a threshold? Will the central bank cut rates at its next meeting? These contracts let traders express macro views without futures or options.
Pop culture & sports
Award winners, championship results, weekly box-office numbers, and increasingly entire sports leagues. Sports event contracts have grown rapidly as regulation has clarified.
Real-world thresholds
Will hurricane season produce more than 14 named storms? Will a clinical trial hit its primary endpoint? These markets are slower but often the most accurate.
How accurate are prediction markets?
The central claim for prediction markets is that they aggregate dispersed information into a single, continuously updating price — and that this price tends to beat traditional forecasters. The evidence is genuinely strong, though not unlimited.
Better than polls, most of the time
The Iowa Electronic Markets, running since 1988, have consistently produced more accurate election forecasts than major opinion polls in the final week before US presidential elections. Academic studies have found similar results for Intrade and for Polymarket in 2024, though markets have also been wrong — sometimes badly — when liquidity was thin or when one side was dominated by a single large trader.
Roughly in line with odds
On mature sports markets, prediction market prices are close to what the best sportsbooks offer, because the same sharp bettors arbitrage between them. The edge is smaller here than in politics or macro.
Faster than expert panels
For Fed rate decisions, inflation prints, and policy events, prediction markets reprice in seconds when news breaks, while economist surveys update monthly. That speed makes them useful as a real-time sentiment gauge even when their central estimate isn't much better than consensus.
Where they break down
Prediction markets are only as good as their liquidity. Markets with low volume, obscure questions, or a single dominant trader can be manipulated or simply wrong. They also can't forecast events no one is thinking to trade on — the classic "black swan" problem.
Risks and regulation
What they're good at
- Aggregate dispersed information into a single, real-time price
- Often more accurate than polls or expert panels
- Let you express forecasts that traditional markets don't price
- Low minimums on most major venues ($1 on Kalshi and Polymarket)
Where they fall short
- Liquidity dries up fast outside the most popular markets
- Long-dated contracts tie up capital that could earn yield elsewhere
- Resolution disputes can occur in ambiguous events
- Tax treatment is unsettled in many jurisdictions
Things to watch
- You can lose 100% of your stake on any single contract
- Crypto-settled venues add custody, wallet, and chain risk
- Regulatory status varies by country and changes quickly
- Markets can be manipulated when liquidity is shallow
- No insider-trading laws apply to prediction markets in most jurisdictions
Status varies by country
- US: CFTC regulates event contracts; Kalshi is fully licensed, Polymarket is relaunching via a licensed derivatives exchange (QCEX)
- UK, EU: no licensed domestic operators; users access offshore venues in a grey zone
- Singapore, Hong Kong: generally prohibited under local gambling law
- Legal status can change with little warning — check your jurisdiction before depositing
The future of prediction markets
After more than two decades in a regulatory gray zone, prediction markets have moved firmly into the financial mainstream. Intercontinental Exchange's October 2025 announcement that it would invest up to $2 billion in Polymarket — while Kalshi, ForecastEx and Robinhood built fully-regulated infrastructure alongside it — marked a genuine inflection point. But the more interesting shifts are still ahead.
LLMs as traders
Large language models are now good enough to research questions, read news, and price probabilities better than most human participants on many types of markets. Manifold already hosts AI-run accounts that consistently rank near the top of its leaderboards, and researchers have shown that frontier models can match or beat superforecaster teams on short-horizon questions.
Markets as training grounds
Prediction markets are an almost ideal environment for AI agents to learn from — every bet has a clear, verifiable outcome and a continuous real-money feedback loop. Expect to see agent frameworks built explicitly to trade on Polymarket and Kalshi, and for market operators to add APIs that make programmatic participation easier.
From fringe to sentiment gauge
Hedge funds, macro desks, and news organizations increasingly reference prediction market prices in real time. As liquidity deepens, expect event contracts to be treated more like VIX-style sentiment indicators — an input to trading, not just a venue for it.
Beyond binary politics
Watch for scalar markets on economic data, weather-linked contracts for businesses that need hedges, and question types no one is trading today — corporate earnings ranges, drug approval timelines, and even AI benchmark results. The underlying mechanism generalizes far beyond elections.
A few words of caution
Event contracts are short-term, all-or-nothing bets on uncertain outcomes — closer in risk profile to sports betting than to long-term investing, and similarly capable of becoming addictive. Treat them as entertainment, not wealth-building. Don't bet money you can't afford to lose, set a budget and stick to it, and never use prediction markets as a substitute for diversified long-term investing. Nothing on this site is financial advice. Always check the legal status in your specific jurisdiction before trading.
Choose your market
Localized guides to the regulators, legal status, and accessible operators in each major market — with platform availability, tax framing and risks written for that jurisdiction in the local language.
United States
CFTC-regulated exchanges, broker-embedded contracts, state-level enforcement, and US tax treatment.
🇬🇧 RegionUnited Kingdom
FCA and Gambling Commission framing, plus which platforms accept UK users.
🇩🇪 RegionDeutschland
Prognosemärkte zwischen BaFin und GGL — auf Deutsch erklärt.
🇸🇪 RegionSverige
Prediktionsmarknader för svenska användare — på svenska.