Prediction Markets: Kalshi, Polymarket, and PredictIt Explained
In short: Prediction markets trade binary contracts on event outcomes. Each contract pays $1 if the event happens and $0 if it doesn’t, with current prices reflecting the crowd’s estimated probability. Kalshi covers sports, politics, economics, and weather. Polymarket has deep liquidity on political events. PredictIt focuses on politics. They’re priced independently from traditional sportsbooks, which creates a uniquely valuable category of arbitrage.
If sportsbooks are betting venues, prediction markets are something more like financial exchanges. The mechanics are different, the regulatory framing is different, and most importantly, the prices are different. That last point is what makes them valuable to anyone hunting for edges.
How Prediction Markets Work
The core unit on a prediction market is a contract. Each contract pays $1 if a specified event occurs, $0 otherwise.
If you think the Lakers will win the NBA Championship, you buy a YES contract. If a YES contract is currently trading at 15 cents, you pay 15 cents per contract. If the Lakers win the championship, you get $1 per contract — a profit of 85 cents on a 15-cent investment, or about +567 in American odds equivalent.
Prices fluctuate based on supply and demand. If lots of people start buying YES, the price rises. If sellers flood in, it falls. The current price represents the market’s aggregate estimate of the probability of the event.
This is fundamentally different from a sportsbook. There’s no bookmaker setting a line based on a model. The “line” is whatever price the market has settled on through trading.
The Major Platforms
Kalshi
Regulation: CFTC-licensed designated contract market. Fully compliant with US derivatives law.
Markets: Sports (game outcomes, championships), politics (elections, policy), economics (inflation, jobs reports, Fed decisions), weather, entertainment.
How it trades: Order book model. Place limit orders at your desired price; trades execute when matched.
Access: Available to US users in most states. Also integrated into Robinhood, which gives 100M+ users prediction market exposure inside an app they already use.
Kalshi has been the major recent expansion in the US prediction market space. Their sports markets are growing quickly, with full coverage of NFL, NBA, MLB, and NHL games, plus championships and player awards. Their help documentation describes the contract mechanics in detail.
Polymarket
Regulation: After being blocked from the US since a 2022 CFTC settlement, Polymarket re-entered the US market in late 2025 by acquiring a CFTC-licensed exchange. US access is available through a regulated on-ramp with KYC requirements, though it’s currently in a phased rollout.
Markets: Politics, current events, sports (lighter coverage than Kalshi), pop culture, cryptocurrency.
How it trades: Order book on a custom blockchain (Polygon). Trades settle on-chain.
Access: US access available to verified users, though some states (Massachusetts, Nevada) have challenged prediction markets through state gambling laws. Availability varies.
Polymarket has the deepest liquidity in the prediction market space, particularly on political events. During major elections and breaking news cycles, billions of dollars in volume flow through. For cross-platform arbitrage, Polymarket prices are often the most reliable reference for non-sports events.
PredictIt
Regulation: Operates under a CFTC no-action letter, with academic affiliation through Victoria University of Wellington.
Markets: Primarily US and international politics — elections, primaries, party leadership, policy outcomes.
How it trades: Order book model. Notable feature: PredictIt charges a 10% fee on profits (not on losses).
Access: Available to US users with traditional banking. The platform has been around since 2014, making it the longest-running US-accessible prediction market.
After winning a lengthy legal battle with the CFTC in mid-2025, PredictIt expanded its market caps significantly. The former $850 position cap was raised to $3,500, and the per-market user limit was removed.
The 10% profit fee meaningfully changes arbitrage math. A 5% gross arb between PredictIt and another platform isn’t a 5% net arb — once you net out the fee, it could be 2-3% or even negative depending on which side wins.
Why Prediction Market Pricing Differs From Sportsbooks
Three structural reasons that create persistent pricing gaps:
Different user bases. Sportsbook users are primarily sports bettors. Prediction market users include forecasters, hedge funds, crypto traders, political wonks, and academic researchers. The collective opinion that sets prices is genuinely different.
Different mechanics. Sportsbooks set lines via traders + models + risk management. Prediction markets set prices purely through order book trading. These produce different pricing dynamics, especially in fast-moving markets.
Different liquidity profiles. Sportsbook lines are backed by the house, which is willing to take essentially unlimited action up to its risk limits. Prediction market liquidity depends on whoever happens to be posting orders. This matters when sizing positions — even with a great price, you may not be able to fill your full intended bet at it.
The result: when a sportsbook offers odds on an event that’s also priced on a prediction market, the implied probabilities frequently disagree. Sometimes by a few percentage points; occasionally by 5-10% or more.
A Real Example
In April 2026, a notable cross-platform spread:
- Kalshi offered a contract on whether the US would agree to a new Iranian nuclear deal: YES priced at 65¢ (implied 65% probability)
- Polymarket offered a roughly equivalent contract: NO priced at 73¢ (implying about 27% probability the deal happens)
Combined implied probability: 65% + 27% = 92%. Below 100% means an arbitrage exists. By buying YES on Kalshi and NO on Polymarket, you guaranteed a profit regardless of whether the deal happened.
Cross-platform arbs of this size happen because almost no traditional arbitrage tools scan prediction markets alongside each other. The prediction market and sportsbook ecosystems have historically been separate worlds, and most arbitrage scanners stay in one or the other.
What Prediction Markets Are Good For
Cross-platform arbitrage. When a sportsbook and a prediction market price the same outcome differently — championship futures, player awards, election outcomes — the gap is often arbitrageable.
Non-sports markets. Prediction markets cover events sportsbooks don’t: elections, economic data releases, policy outcomes, weather, even celebrity events. For exposure to those topics, prediction markets are the only game in town.
Lower-vig pricing on overlapping events. Like exchanges, prediction markets have very low effective margin on liquid markets. For some events, they offer better prices than sportsbooks.
Transparency. Order books are visible. You can see exactly how much volume is offered at each price, which gives you better information about liquidity than sportsbook markets typically provide.
Tradeoffs to Keep in Mind
Liquidity is uneven. Major events (presidential elections, championship futures, popular sports games) have deep order books. Niche markets often have thin order books, where placing even a modest order can move the price against you.
PredictIt’s profit fee is real. Always factor the 10% fee into PredictIt arbitrage calculations. It’s the most common source of “obvious arb” mistakes for new users.
Settlement timing differs. Sportsbook bets settle minutes after the event ends. Prediction market contracts may not settle for hours or days, depending on the platform’s resolution process. For positions you want to hedge or close before resolution, this matters.
Programmatic access varies. Kalshi has a robust API. Polymarket trades on-chain (with all the friction of crypto wallets). PredictIt’s UI is the primary interface. For automated arbitrage, the platforms differ substantially in how easy they are to integrate with.
How to Get Started
If prediction markets are new to you:
- Open a Kalshi account first. It’s the most accessible US-regulated platform with the broadest market coverage.
- Start with markets you have an opinion on. Sports outcomes, elections, economic events. Place small positions to learn the mechanics.
- Compare prices to sportsbooks when you can. Get a feel for where the markets agree and disagree.
- Watch the order book. Liquidity reveals itself in the bid/ask spread and depth at each price.
- For arbitrage specifically, use a scanner that covers both prediction markets and sportsbooks. Manual cross-platform arbitrage is impractical at scale.
Frequently Asked Questions
Are prediction markets legal in the US?
Kalshi is fully CFTC-regulated as a designated contract market. PredictIt operates under a CFTC no-action letter. Polymarket operates with US access through a CFTC-licensed acquisition, though some states have challenged prediction markets under state gambling law. Federal status is increasingly settled; state-level uncertainty remains.
Are these the same as sports betting?
Mechanically similar but legally distinct. Sports betting is regulated state-by-state under gambling law. Prediction markets are regulated federally under derivatives law (CFTC). The functional difference for users is that prediction markets cover non-sports events sportsbooks can’t (politics, economics) and operate on order books rather than house-set lines.
Why does Kalshi seem to have lower limits than sportsbooks?
Liquidity on Kalshi (and prediction markets generally) is set by user order books, not by the platform itself. On heavily-traded markets, you can move significant volume. On thin markets, even small orders can move the price. This is fundamentally different from a sportsbook, where the house provides liquidity up to its risk limits.
Can I do automated trading on prediction markets?
Yes, all three major platforms have APIs (Kalshi’s is the most mature). This makes them better candidates for systematic strategies than most sportsbooks. The catch is that the API ecosystems are less developed than financial exchanges, and integration work is required.
Parcae is a sports betting intelligence tool currently in development. Learn more about what we’re building →
Sports betting involves risk. If you or someone you know has a gambling problem, contact the National Council on Problem Gambling at 1-800-522-4700.