Strategy

Prediction Market Arbitrage: Trade the Spread Between Kalshi and Polymarket

Jay MalaviaJuly 21, 20266 min read

The same event. Two prices. One profitable trade. Prediction market arbitrage is the practice of buying a contract at a discount on one venue and simultaneously selling the equivalent position on another, locking in a spread that pays out regardless of how the event resolves.

Kalshi, Polymarket, and Predict.fun list contracts on many of the same events. They do not share a price. That is the entire opportunity.

What is prediction market arbitrage?

A binary event contract settles at $1 (Yes wins) or $0 (Yes loses). If the same question trades at 35 cents Yes on Kalshi and 32 cents Yes on Polymarket, those two contracts should converge. They describe the same outcome. They will resolve identically. The price difference is inefficiency, and inefficiency is the raw material of arbitrage.

The trade: buy the cheap Yes on Polymarket at 32 cents. Buy No on Kalshi at 65 cents (No is the inverse: 1 minus 35 cents). Spend 97 cents total per pair. At settlement, one of those two positions pays $1. Net gross profit per pair: 3 cents. At 1,000 contract pairs, that is $30 before fees.

Two types of opportunity exist. Cross-venue arbitrage captures the spread between Kalshi and Polymarket on the same question. Yes-No arbitrage captures gaps within a single venue when Yes and No prices do not sum cleanly to $1. Cross-venue is the more active space. If prediction markets are new to you, what is a prediction market covers the foundations.

Where do spreads come from?

Fragmented liquidity. Kalshi is CFTC-regulated and restricted to U.S. residents who can pass KYC. Polymarket runs on the Polygon blockchain and draws a global, crypto-native audience. Two capital pools with different compositions, different market hours, and no direct flow between them. That structural asymmetry keeps prices misaligned longer than they would stay misaligned on a unified market.

Information latency. When a breaking news item reprices a contract, it travels faster on one venue than the other. The window where both prices exist simultaneously can be seconds. Those seconds are the trade.

Fee structures also keep prices sticky. Both venues charge per trade, but with different structures. Participants price in the cost of exiting and hold positions longer than they otherwise would. That patience creates the friction that keeps spreads alive. For a detailed look at how the venues stack up, read Kalshi vs Polymarket.

What does a real trade look like?

A Fed rate decision market: Yes trades at 38 cents on Kalshi and 34 cents on Polymarket. The gap is 4 cents gross. On 500 contract pairs, that is $20 gross before fees. After fees on both venues, a trade at this size might net $8 to $12. Not a windfall.

Scale changes the math. Serious prediction market arbitrageurs work at high volume across many events simultaneously. Reddit threads document traders turning systematic cross-venue arbitrage into seven-figure returns. Oddpool runs a free public scanner that surfaces live spreads between Kalshi and Polymarket in real time. The edge is not in knowing the opportunity exists. The edge is in executing faster and cheaper than everyone else who also knows.

Event contracts carry real market risk. A trade that looks like arbitrage but resolves differently on two venues because the resolution criteria differ is not arbitrage. It is an open position. Read the market rules before you size up.

The real limits of prediction market arbitrage

Speed kills most retail arbitrage. Open-source prediction market arbitrage bots exist on GitHub and execute on cloud infrastructure faster than any manual trader. By the time you see a spread in two browser tabs, a bot may have already closed it. The spreads that remain for manual traders are the ones too small for automated systems to bother with, or the ones that just appeared.

Basis risk is the subtler limit. Kalshi and Polymarket sometimes list different versions of the same event with slightly different resolution criteria. A Fed rate decision market on Kalshi might resolve on the scheduled FOMC meeting while the Polymarket version resolves on any rate action in the calendar month. Same headline. Different contract. Not arbitrage. Know the exact resolution terms before treating a price difference as riskless.

The opportunity is real. The edge belongs to faster infrastructure and tighter execution.

How serious arbitrageurs execute

Serious arbitrageurs use one of three approaches. The Kalshi API and Polymarket API let programmers build automated strategies that scan and execute at machine speed. Dedicated scanners like Oddpool surface the live spread without custom code. Trading terminals give manual traders the unified book view they need to act the moment a spread opens.

Kairos is the terminal for this. One unified book across Kalshi, Polymarket, and Predict.fun. Sub-second data from every venue. Global best bid and best ask rendered in a single view. Advanced order types. Low-latency execution. What used to require two browsers, two logins, and a race against your own lag now happens from one screen.

Prediction markets are fragmenting across more venues every year. The edge in cross-venue arbitrage compounds with every new market added to your view. Kairos adds them. Start trading on Kairos.

Prediction markets are becoming a real asset class. The tooling has to match. Kairos.

See you in the order books.

Frequently asked questions

What is prediction market arbitrage?

Prediction market arbitrage exploits price differences for the same event contract across venues like Kalshi and Polymarket. Buy where the contract trades cheap. Sell the equivalent position where it trades expensive. At settlement, both sides resolve identically. You capture the spread, minus fees.

Is prediction market arbitrage risk-free?

Cross-venue arbitrage has low directional risk but is not risk-free. Basis risk exists if two venues resolve the same event with different criteria. Execution risk is real: if one leg of a trade fills and the other does not, you hold an open position. Event contracts carry real market risk on both sides.

Why do price differences exist between Kalshi and Polymarket?

Kalshi is CFTC-regulated and U.S.-only. Polymarket runs on blockchain and draws a global audience. Different capital pools, different fee structures, and different information speeds create persistent price gaps on identical events. That gap is the source of cross-venue arbitrage.

How do I find arbitrage opportunities in prediction markets?

Scanners like Oddpool aggregate live spreads across venues in real time. Programmatic traders use the Kalshi API and Polymarket API to automate the search. A trading terminal like Kairos shows the unified book across Kalshi, Polymarket, and Predict.fun in one view with sub-second data from every venue.

How big are prediction market arbitrage spreads?

Spreads between Kalshi and Polymarket on active markets range from a few cents to 10 cents or more on the same event. After fees on both venues, the net spread is smaller. Speed and size determine whether a spread is worth trading. Systematic traders have documented significant returns from high-volume arbitrage across many events.

Can automated bots do prediction market arbitrage?

Yes. Open-source prediction market arbitrage bots are publicly available on GitHub. They run on cloud infrastructure and close spreads faster than manual traders can. For manual traders, the best edge is a terminal with unified real-time data across all venues, fast order entry, and the low latency to execute the moment a spread opens.

By Jay Malavia

Start Trading

More from the blog