Prediction Markets Explained: How Event Contracts Work in 2026
What is a prediction market?
A prediction market is a marketplace for event contracts: contracts whose payout depends on whether something happens. "Will the unemployment rate be above 4.5% in the next jobs report?" "Will this bill pass the Senate by June 30?" Each question becomes a contract with two sides, Yes and No.
The design is simple on purpose. A Yes contract pays $1 if the event happens and $0 if it doesn't. A No contract pays the opposite. Nobody gets paid out more than $1 per contract, so the price you pay is also the most you can lose.
Because traders put money behind their views, prices tend to move fast when news arrives. That is the reason economists, newsrooms and traders watch them: a liquid market is a live, money-weighted forecast.
How prices become probabilities: a worked example
Take an illustrative market: "Will the Fed cut rates at its next meeting?" The numbers below are made up to show the math, not a real quote.
- The market prices Yes at 62 centsThat means traders, on balance, put the chance of a cut at about 62%. No trades near 38 cents, because one Yes plus one No always pays exactly $1.
- You buy 100 Yes contracts for $62That is your maximum loss. Fees come on top and vary by venue.
- News moves the priceA soft inflation report comes out and Yes rises to 80 cents. You can sell now for $80 and lock in $18 before fees, without waiting for the meeting.
- Or you hold to resolutionIf the Fed cuts, your 100 contracts pay $100: a $38 profit. If it doesn't, they pay $0 and you lose the $62.
Two details matter for anyone building or reading these markets:
- The quoted price is not always a trade price. Polymarket's documentation, for example, shows the midpoint between the best bid and the best ask, and switches to the last traded price when the spread is wider than 10 cents. A thin market can show a "probability" nobody is actually willing to trade at.
- Yes and No are two halves of one dollar. On Polymarket, a buy order for Yes at 60 cents can match a buy order for No at 40 cents: together they put up $1 of collateral, which creates one Yes and one No contract. That is how a market starts trading without anyone holding inventory first.
The life of an event contract
Every prediction market, regulated or on-chain, runs the same five stages. Most of the hard engineering and most of the disputes sit in the first and the last two.
Order books vs automated market makers
Someone has to be on the other side of every trade. Prediction markets solve this in two ways.
An order book (also called a central limit order book) collects buy and sell orders at stated prices and matches them. Kalshi and Polymarket both run order books, and Polymarket's documentation describes trading as peer-to-peer: users trade with each other, not against the platform. Order books need market makers, firms or traders that keep quotes on both sides, or prices get thin and jumpy.
An automated market maker (AMM) is a formula that always quotes a price. The best-known one for prediction markets is the logarithmic market scoring rule (LMSR), designed by economist Robin Hanson. The operator sets a liquidity parameter, and the formula moves the price a little with every purchase. The operator's worst-case loss is capped and known in advance, which makes LMSR popular for internal corporate forecasting markets and new markets with few traders.
| Order book | AMM (LMSR and similar) | |
|---|---|---|
| Who takes the other side | Other users and market makers | The operator's formula |
| Always has a price | Only if someone is quoting | Yes |
| Who pays for liquidity | Market makers, through incentives or spread | The operator, up to a known maximum loss per market |
| Best for | Popular markets with many traders | Long-tail questions, internal forecasting, new venues |
| Engineering effort | Matching engine, risk checks, market data feeds | Simpler pricing, but careful math on rounding and limits |
Many venues end up hybrid: an AMM or a paid market maker seeds new markets, and the order book takes over once there is real interest.
Resolution and oracles: who decides what happened
Resolution is the step where a prediction market earns or loses trust. There are two main models.
The exchange resolves. A regulated US exchange publishes the terms of each contract, including the source that decides the outcome, such as an official statistics release or a league's final score. When the event ends, the exchange settles by those terms. Disputes go through the exchange's rulebook, under CFTC oversight.
An oracle resolves. On-chain markets need an oracle: a system that brings a real-world fact onto the blockchain. According to Polymarket's documentation, its international prediction markets use the UMA Optimistic Oracle:
- Anyone can propose an outcome by posting a bond (typically $750, per the docs). An honest, undisputed proposer gets the bond back plus a reward.
- Anyone can dispute within a 2-hour challenge window by posting the same bond. Without a dispute, the market resolves after about 2 hours.
- If a second proposal is also disputed, UMA token holders vote on the answer. A disputed resolution takes roughly 4–6 days, and the losing side loses its bond, half of which goes to the winning side.
- If the answer can't be determined, the market can resolve 50-50: each contract pays 50 cents.
Polymarket's documentation also notes that some short-term price markets ("up or down") don't use UMA at all. They resolve automatically from Chainlink price data, comparing the price at the start and at the end of the window. For builders, the lesson is the same either way: pick the source before you list the market, and write down what happens if the source is late, revised or silent.
Who runs regulated prediction markets in the US
In the US, real-money event contracts are regulated as derivatives by the Commodity Futures Trading Commission (CFTC). They trade on designated contract markets (DCMs): exchanges the CFTC has approved. According to a CFTC court filing from October 2026, 30 exchanges hold active DCM designations, including newer event contract venues and long-established ones such as CME Group and Nadex. The main names as of October 2026:
| Venue | Status (from official sources) |
|---|---|
| Kalshi | KalshiEX LLC, designated as a contract market by the CFTC on November 3, 2020. |
| Polymarket US | Runs on QCX LLC, designated by the CFTC on July 9, 2025 and now operating as Polymarket US. The CFTC later amended the order so that brokers (futures commission merchants) can carry customer accounts. |
| Polymarket (international) | The on-chain venue on Polygon. Its documentation lists the US as a restricted jurisdiction: US users can only close existing positions. In 2022 the CFTC fined Polymarket's operator and ordered it to wind down unregistered US operations. |
| Robinhood | Offers prediction markets to its customers. A Robinhood-controlled joint venture with Susquehanna closed its purchase of MIAXdx, a CFTC-licensed exchange and clearinghouse, on January 20, 2026, according to Robinhood. |
| CME Group, Nadex | Long-standing DCMs that the CFTC names among event contract providers. |
Each exchange also needs a clearinghouse (a derivatives clearing organization) that holds collateral and guarantees settlement. Brokerages can offer event contracts to their customers by routing orders to a DCM rather than running one.
The legal fight over sports contracts (as of October 2026)
Contracts on economic data and weather draw few legal challenges on DCMs. Sports contracts draw the most, and some states also challenge election contracts. Many states see them as sports betting that needs a state license. The CFTC says federal law gives it exclusive jurisdiction over contracts traded on its exchanges. Where things stand on October 29, 2026:
- The appeals courts are split. The Third Circuit sided with Kalshi against New Jersey in April 2026 and held that sports event contracts are swaps under federal law. The Ninth Circuit (Nevada, August 2026) and the Sixth Circuit (Ohio and Tennessee, September 2026) went the other way, reasoning that these contracts are, in substance, sports gambling that states can regulate. The CFTC summarizes this split in its own October 2026 rulemaking.
- The Supreme Court has been asked to decide. New Jersey filed a petition on September 2, 2026 (Flaherty v. KalshiEX, No. 26-299). Kalshi's response is due November 9, 2026, and the Court has not yet decided whether to hear the case.
- The CFTC is suing states. On October 27, 2026 the CFTC announced a lawsuit against Utah, which had sued Kalshi and two Susquehanna firms in state court on October 20. It was the tenth state the CFTC has taken to court, after Arizona, Connecticut, Illinois, Kentucky, Minnesota, New Mexico, New York, Rhode Island and Wisconsin. Courts have granted preliminary injunctions against state enforcement in Arizona, Illinois and Minnesota.
- The CFTC is rewriting the rules. In February 2026 it withdrew a 2024 proposal that would have restricted event contracts. In June 2026 it proposed rules on how it would decide whether contracts involving activities such as gaming are contrary to the public interest. On October 9, 2026 it announced an interim final rule saying casino-style gambling, including sportsbook wagers, is not a swap (published in the Federal Register and effective October 19), and proposed a rule that would expressly include sports, political, cultural and weather event contracts in the swap definition.
This area changes month to month, and the outcome may differ by state until the Supreme Court or Congress settles it. If you plan to list or trade event contracts, especially on sports, confirm the current position with counsel who practices CFTC and gaming law.
Risks: manipulation, insiders, liquidity and disputes
- Insider trading. Some contracts are about events a few people control or know early. In February 2026 the CFTC's enforcement division published an advisory after Kalshi disciplined a political candidate who traded on his own race and a YouTube channel editor who likely traded on advance knowledge of its videos. The CFTC has since brought its own cases, including one against a Google employee accused of trading on nonpublic information about Google's Year in Search list.
- Manipulation. Thin markets are cheap to move. A trader can push a price to create a headline, or trade with themselves to fake volume (wash trading). The CFTC advisory lists pre-arranged and wash trades among prohibited practices on DCMs.
- Sports integrity. The CFTC has signed information-sharing agreements with Major League Baseball, the National Hockey League and the National Basketball Association so it can react to suspicious activity around games.
- Liquidity. A market with a wide spread shows a price that nobody will trade at. Read volume and depth before reading the probability.
- Resolution disputes. Ambiguous questions, revised official data and late sources all lead to fights. On-chain, token-holder votes add another party with its own incentives.
- Regulatory change. A contract type that is allowed today can be challenged tomorrow, state by state. Geofencing, the ability to block users by location, is a feature you should plan for from day one.
Built by Gilzor
Results we’ve shipped
Talk to the people who build it. Tell us about your project and get a free estimate of scope, timeline and cost.
How a company builds a prediction market
The first decision is not the tech stack. It is the venue model: who holds the license, who holds customer money, and who you can legally serve.
- What it is: you apply to the CFTC for designation and arrange clearing through a clearinghouse you own or contract with.
- You build: the matching engine, market surveillance, audit trails, rulebook, member onboarding, and the systems that protect customer funds, all of which the Commodity Exchange Act's core principles require.
- Watch out: the longest and most expensive path. Several companies bought an existing licensed exchange instead (Polymarket with QCX, Robinhood's joint venture with MIAXdx).
- What it is: you build the customer experience and route orders to an existing DCM, usually through a registered broker relationship.
- You build: the app, onboarding, market discovery, portfolio views, notifications and the integration with the exchange's APIs.
- Watch out: the exchange decides which contracts exist and how they resolve. Your margin depends on the partner's fee split.
- What it is: smart contracts hold collateral (usually a stablecoin) and pay out after an oracle resolves the market.
- You build: outcome token contracts, an order book or AMM, oracle integration, wallet login, and geoblocking for restricted countries.
- Watch out: serving US users without registration is what led to the CFTC's 2022 order against Polymarket's operator. Plan audits as for any DeFi app.
- What it is: play-money or points markets, or internal forecasting markets for employees on launch dates, sales or project risk.
- You build: an LMSR-style market maker, simple accounts and dashboards.
- Watch out: once real money or prizes of value enter, the rules change. Check with counsel before adding them.
What the platform needs
| Component | What it does |
|---|---|
| Market creation | Templates for contract terms: question, resolution source, deadlines, edge cases. On a DCM, new contracts are typically self-certified with the CFTC before listing. |
| Identity and compliance | KYC (identity checks), sanctions screening, location checks, and blocks for people who may have inside information, such as athletes or campaign staff. |
| Trading engine | Order book or AMM, pre-trade risk checks (does the user have the funds?), and market data feeds. |
| Ledger and custody | A double-entry ledger of every position and dollar, reconciled with the clearinghouse or the chain every day. Our fintech app guide covers the pattern. |
| Surveillance | Alerts for wash trading, spoofing, unusual volume before news and linked accounts. Regulated venues must keep audit trails and act on them. |
| Resolution operations | A workflow to record the source, the decision, who approved it and how disputes were handled. Oracle integration for on-chain markets. |
| Settlement and payouts | Pay winners, release collateral, handle 50-50 and voided markets, then withdrawals. |
Under the hood, much of this looks like an exchange. Our crypto exchange development cost guide shows how matching engines, wallets and back offices drive budget, and our blockchain app development cost guide covers the on-chain pieces.
Readiness checklist
Tick what is already true for your product.
Prediction market readiness
FAQ
How do prediction markets work?
Are prediction markets legal in the US?
Is Polymarket legal in the US?
What is the difference between a prediction market and sports betting?
Who decides the outcome of a prediction market?
How do prediction market platforms make money?
Where Gilzor fits
We build Web3 and dApp products, smart contracts and fintech software: trading apps, wallets, onboarding with KYC providers, back offices and the APIs that connect them to exchanges and partners. For KickEX, a crypto exchange, our team refactored the app architecture and rewrote unstable code so users could trade, transfer and withdraw reliably. We work from Poland and Cyprus, with a few shared hours a day with the US East Coast.
If you are planning an event contract product, tell us your venue model. We'll help you map the architecture, the compliance questions for your counsel and a first version you can launch.
No sales pitch
Get a straight answer for your project
Tell us what you’re building. We’ll reply with options, a rough cost and timeline. If we’re not the right fit, we’ll say so.
Co-Founder of Gilzor. Works with founders and product companies on how to staff and run engineering: team extension, dedicated teams, and getting stalled projects moving again.
Gilzor · Web Development partner
Need a team for your web product?
Services
Web DevelopmentCustom websites and web apps — front-end, back-end, launch and support.→By company type
Selected projects
The team behind them
