What Are Prediction Markets? A Plain-English Guide

Been reading about prediction markets but still not sure what they are or how they work?

Here's the plain-English version: a prediction market is a place where you trade on the outcome of a real-world event instead of betting against a bookmaker. Every contract is tied to a yes-or-no question, like "will the Fed cut rates in September," and its price is the market's live estimate of how likely the answer is. A $0.60 Yes contract means traders think there's a 60% chance it happens. If it does, the contract pays $1. If it doesn't, it expires worthless.

Based on volume, five platforms run these markets at scale today: Polymarket, Kalshi, Limitless, Predict.Fun, and Opinion.

What Is a Prediction Market?

A prediction market is a way to price the odds of a future event without a bookmaker: traders buy and sell contracts tied to a yes-or-no outcome, and the price settles wherever the crowd's money agrees the probability sits.

That's different from a poll, which asks what people think, and different from a sportsbook, which sets its own price and takes the other side of your trade. Across the five major platforms, Polymarket, Kalshi, Limitless, Predict.Fun, and Opinion, traders set the price themselves by buying and selling against each other, the same way a stock's price comes from buyers and sellers rather than the company's own estimate of its worth.

How Do Prediction Markets Work?

Every market centers on one specific, verifiable question. Not "who will win the election" in the abstract, but "will Marco Rubio win the US Presidency in November 2028," for example. Vague questions can't resolve. Specific ones can.

The contract itself is simple: a Yes share and a No share, together worth $1. If the event happens, Yes pays $1 and No pays $0. If it doesn't, the reverse. The price of the Yes share between now and resolution is the market's running estimate of the odds, expressed as a number between $0 and $1.

You don't have to hold a contract to resolution. If new information moves the odds in your favor, you can sell and take the gain immediately. If it moves against you, you can sell and cut the loss rather than ride it to zero. That's what separates a position in a prediction market from a bet placed and forgotten.

A Live Example

Take the market on who wins the 2028 Republican presidential nomination. On Limitless, a Yes contract on Marco Rubio trades at $0.50, meaning the crowd there is pricing him as a coin flip.

Check the same contract elsewhere and the picture changes. Kalshi prices Rubio at $0.29. Polymarket at $0.23. Opinion at $0.21.

Four platforms, one event, a 29-point gap on the frontrunner. Each platform is its own market, and its price reflects whatever pool of traders is active there, not one objective forecast.

J.D. Vance tells a different story. Three platforms land within a point of each other at 41 to 42%. Opinion prices him at 28%, twelve points below the rest. Even the closest thing to consensus in this market has an outlier.

$742 million in total volume and $48 million in live liquidity back this market. That 29-point gap on Rubio reflects real disagreement between four active pools of money, not a glitch in thin trading.

That's why checking one platform tells you what one crowd thinks. Checking all of them tells you where the disagreement actually is. (Prices as of July 5, 2026, per PredictionHero.)

How Big Is the Prediction Market Industry Right Now?

Prediction markets moved more than $91 billion in taker volume in the last 30 days alone, per Dune Analytics. Kalshi ($47.1B) and Polymarket ($24.1B) lead. Opinion ($12.7B), Predict.Fun ($1.9B), and Limitless ($1.7B) make up most of what's left. That's the same five platforms PredictionHero covers. (Figures are community-maintained on Dune Analytics, not an official industry benchmark, and update continuously.)

That volume didn't build up gradually. Combined monthly trading on Kalshi and Polymarket went from under $5 billion in September 2025 to about $24 billion by April 2026, roughly a fivefold jump in seven months, according to Pew Research. Pew's own comparison point puts that in perspective: US legal sportsbooks averaged around $14 billion a month in wagers through 2025. Prediction markets now move more money every month than the entire regulated sports betting industry.

TRM Labs tracked the same acceleration from the wallet side: unique addresses trading on these platforms nearly tripled to 840,000 over the six months ending February 2026, and monthly volume crossed $20 billion that January. The activity isn't crypto speculation wearing a new label, either. Geopolitics, US politics, and macro releases like Fed decisions drive most of the volume; crypto price markets are a minor slice by comparison, despite running on the same blockchain rails.

Investors are pricing in more growth, not less. Kalshi closed a funding round at a $22 billion valuation in March 2026, with Sequoia, Andreessen Horowitz, Coatue Management, and Morgan Stanley on the cap table, and by June was in talks to nearly double that to $40 billion, per CoinDesk. Polymarket was reportedly seeking $15 billion over the same stretch. The valuation gap tracks the volume gap above: Kalshi's federal regulation is pulling in institutional money that Polymarket's crypto-native structure has a harder time attracting.

Where to Trade: the Five Platforms

Polymarket is the volume leader, and it earned that spot honestly: crypto-native, global, and covering everything from elections to entertainment to sports before most competitors existed. The catch for US traders is that the platform you've heard of isn't the one you can use, US residents trade through a separate, CFTC-regulated product, Polymarket US, not the original global exchange. $24.1B in taker volume moved through Polymarket over the last 30 days. Trade on Polymarket

Kalshi has been playing a different game from day one: built for regulation instead of around it, the only one of the five that's operated as a CFTC-regulated exchange since launch. That's made it the natural home for economic, policy, and legislative questions, categories regulators watch more closely than "who wins the game." It's also the outright volume leader: $47.1B over the same 30 days, more than Polymarket and Opinion combined. Trade on Kalshi

Opinion is the platform most explainers skip, and that's a miss: $12.7B in 30-day taker volume puts it ahead of Predict.Fun and Limitless combined. Its US regulatory status isn't publicly documented, so treat that as an open question, not an assumption. Trade on Opinion

Predict.Fun moved $1.9B over the same 30 days, smaller than the other four but not a rounding error. Like Opinion, it doesn't publicly state a US regulatory status. Trade on Predict.Fun

Limitless moved $1.7B over the same period and runs one of the more interesting fee structures of the five: a flat 0.25% on its AMM markets, and a sliding 0.03% to 3% on order-book markets that gets cheaper as an outcome becomes more certain. None of that matters for most US readers, though: Limitless isn't available to US residents, since real-money event contract platforms need CFTC registration to operate there, and Limitless hasn't pursued it. Trade on Limitless

Prediction Markets vs. Polls vs. Sports Betting

A poll costs a respondent nothing to answer. A prediction market costs a trader real money to be wrong, and that difference shows up in the results. A pollster reports what people say. A market prices what people are willing to risk being right about, and it reprices the moment new information lands, not on the next survey's schedule. Historically, that's made prediction markets a strong complement to polling, not a replacement for it. Full breakdown: how accurate are prediction markets.

The gambling comparison runs on different logic. A sportsbook sets a price and profits when bettors lose, with its advantage built into that price. A prediction market has no house. Traders take positions against each other, and platforms earn revenue from trading fees, not from your loss. That's also why the language is different: positions and contracts, not bets and winnings. More detail: prediction markets vs. sports betting.

Are Prediction Markets Legal in the US?

Kalshi is a CFTC-regulated derivatives exchange, which makes it legal to trade in the US under federal commodities law. That federal status doesn't end the question, though. Sports-related event contracts specifically are caught in an active jurisdictional fight: in April 2026, the Third Circuit Court of Appeals ruled the CFTC likely has exclusive jurisdiction over Kalshi's sports contracts, rejecting New Jersey's attempt to apply state gambling law. Three months earlier, a Massachusetts state court reached the opposite conclusion for that state, blocking Kalshi's sports contracts there without a license. The CFTC and DOJ have since sued Arizona, Connecticut, and Illinois directly over their state bans.

Polymarket's US status runs through a separate product. The original global Polymarket exchange has been geoblocked for US users since a 2022 CFTC settlement over unregistered event contracts. Polymarket US, operated by QCX LLC (a CFTC-licensed exchange Polymarket acquired for roughly $112 million in July 2025), launched December 3, 2025 as the only Polymarket product US residents can legally use. It's still blocked in eight states as of mid-2026: Arizona, Illinois, Massachusetts, Maryland, Michigan, Montana, Nevada, and Ohio.

Limitless is not available to US residents. Predict.Fun and Opinion don't publicly document a US regulatory status.

This is a fast-moving, actively litigated area. Treat every claim above as needing a same-day check before publishing.

How Do Prediction Markets Make Money?

Fee structures vary by platform and change often. As of July 2026:

• Kalshi charges a per-contract taker fee, roughly 7% of price times (1 minus price), peaking on contracts near $0.50 and shrinking toward the extremes. No settlement, membership, or ACH-transfer fees. Debit card deposits and withdrawals carry a 2% processing fee. Kalshi fee schedule

• Polymarket charges category-based taker fees: up to $0.75 per 100 shares on sports, $1.00 on politics and finance, $1.25 on economics and culture, $1.80 on crypto, peaking near 50% probability. Makers pay nothing and receive a rebate. Geopolitical and world-event markets carry no fees at all. Polymarket Help Center

• Opinion uses a dynamic fee between 0% and 1% that rises as a market nears 50% probability, with a $5 minimum order and $0.25 minimum fee. Limit orders pay nothing. Opinion fee docs

• Limitless runs two fee structures depending on market type. AMM markets charge a flat 0.25% on every trade, no curve. Order-book markets charge between 0.03% and 3%, moving inversely with probability: a near-certain outcome costs less to trade than a coin flip, and trades returning more than 2x carry a slightly higher fee. Either way, makers pay nothing; only takers pay. Limitless fee docs

• Predict.Fun charges takers only, with fees finalized at settlement and collected in USDT. The exact percentage isn't publicly disclosed.

None of the five profit when a trader loses. Revenue comes from trading activity itself, the core distinction from a sportsbook's built-in advantage.

A Brief History and Track Record

The Iowa Electronic Markets launched in 1988 as the Iowa Political Stock Market, built by University of Iowa economists George Neumann, Robert Forsythe, and Forrest Nelson to trade on that year's presidential election. It's the oldest continuously operating prediction market and still runs as an academic research project today.

Its track record holds up: research comparing IEM prices to 964 polls across five presidential elections from 1988 to 2004 found the market closer to the eventual result than the polls in 74% of comparisons, with an average error of 1.33 percentage points on election-eve forecasts. (MarketsWiki, citing Berg et al.)

The modern picture is more mixed. A 2025 analysis of Polymarket's 2024 election forecasts found it correctly called the presidential winner but didn't clearly beat statistical models from The Economist, Silver Bulletin, and FiveThirtyEight on the popular vote or Electoral College margin, and underperformed those models on individual Congressional races, likely due to thinner trading volume down-ballot.

Glossary of Terms

AMM (automated market maker): a pricing algorithm that sets contract prices from a formula instead of matching individual buy and sell orders. Limitless runs AMM markets at a flat 0.25% fee, alongside a separate order-book market with variable fees.

Arbitrage: trading the same event on two platforms at different prices to lock in a profit no matter which side wins. The Rubio example above, 50% on Limitless versus 21% on Opinion, is exactly the kind of gap arbitrageurs look for.

CFTC: the Commodity Futures Trading Commission, the federal regulator with jurisdiction over event contracts listed in the US. Kalshi and Polymarket US both operate under CFTC oversight.

Crypto-native platform: a platform that settles trades on a blockchain and funds accounts with crypto rather than a bank transfer. Polymarket, Limitless, and Predict.Fun all fall into this category; Kalshi doesn't.

Designated Contract Market (DCM): the CFTC's formal license for an exchange to list event contracts in the US. Kalshi has operated as a DCM since it launched.

Event contract: a financial instrument tied to a yes-or-no outcome, priced between $0 and $1.

Geoblocking: restricting access to a platform based on a trader's location. Polymarket's original global exchange geoblocks US users; Limitless blocks US residents entirely.

Implied probability: a contract's price read directly as a percentage chance. A $0.60 Yes contract implies a 60% chance the event happens, which is where the pricing examples throughout this guide come from.

Liquidity: how much money is available to trade a market without moving its price significantly. The Rubio market above had $48 million in live liquidity behind it.

Maker: a trader whose order sits on the book waiting to be filled rather than matching immediately. Polymarket and Limitless charge makers nothing, and Polymarket pays them a rebate.

Order book: the live list of buy and sell orders waiting to be matched.

Resolution: the point at which a contract's outcome is determined and it settles to $1 or $0.

Settlement: the payout that follows resolution, based on a pre-defined source: an official election result, an economic data release, or a game's final score.

Slippage: the gap between the price you expected to pay and the price you actually got, usually because your order size moved the market before it filled.

Spread: the gap between the highest price a buyer will pay and the lowest a seller will accept.

Taker: a trader whose order fills immediately against an order already on the book. Taker fees are how most of these platforms make money — see "How Do Prediction Markets Make Money?" above.

Volume: the total dollar value traded on a market or platform over a given period. Distinct from liquidity, which measures what's available to trade right now, not what's already traded.

Yes / No: the two sides of a contract. Yes pays $1 if the event happens; No pays $1 if it doesn't.

How to Get Started

1. Compare live odds across all five platforms on PredictionHero before committing to one. If Kalshi and Polymarket disagree by double digits, that's information too.

2. Create an account on the platform that carries the market you want, and fund it. Make sure to read the fine print as some platforms accept crypto and others only accept normal fiat transfers.

3. Pick a market you actually understand, not just one with a headline that catches your eye.

4. Size a position you're fine holding to resolution. You can sell early if the price moves before then.

5. Track it. Prices move continuously, and nothing locks you in until the event resolves.

PredictionHero aggregates publicly available prediction market data for informational purposes only. This is not financial advice. Prediction markets may not be available in all jurisdictions.