Prediction market arbitrage sounds simple. Buy both sides of the same event on two venues for less than a dollar combined, and you keep the difference no matter how the event resolves. In practice, the math is easy and the execution is not. This guide walks through the logic, an honest example, and why pure arbitrage is hard for retail traders in 2026. This is not financial advice.
What is prediction market arbitrage
Arbitrage means taking both sides of the same event on different venues to lock a profit regardless of the outcome. It works because of how these contracts settle. In a single market, the YES share and the NO share settle to $1 between them. If the event happens, YES pays $1 and NO pays nothing. If it does not, NO pays $1 and YES pays nothing.
So the trick is price. If you can buy YES on one platform and NO on another for less than $1 combined, you hold both outcomes and one of them is guaranteed to pay $1. The gap between what you paid and that $1 is your profit. If the two sides cost more than $1 combined, you lock a loss instead. That single rule, YES plus NO equals $1, is the whole basis of prediction market arbitrage.
How it works (with an example)
Say the same event is listed on two venues. Kalshi shows 65 cents for YES. Polymarket shows 42 cents for NO. If you buy both, you pay 107 cents to hold a position that only ever returns 100 cents. That is a 7 cent loss per pair, before any fees. This is the losing case, and it is common. Most of the time the two sides add up to more than a dollar, which is why quoted prices alone do not mean an opportunity exists.
The trade only works when the two sides total under 100 cents. If Kalshi showed 55 cents for YES and Polymarket showed 43 cents for NO, the pair costs 98 cents and returns 100 cents. That is a 2 cent gain per pair, locked regardless of how the event resolves. The table below shows both cases side by side.
| Platform | Side | Price |
|---|---|---|
| Kalshi | YES | 65c |
| Polymarket | NO | 42c |
| Combined (loss) | Both | 107c |
| Kalshi | YES | 55c |
| Polymarket | NO | 43c |
| Combined (gain) | Both | 98c |
Notice how small the winning edge is. A 2 cent gap on a dollar is a 2 percent gross return, and that is before you account for fees and the effort of moving money between venues.
Cross-platform arbitrage: Polymarket vs Kalshi
Polymarket and Kalshi are the two names that come up most for prediction market arbitrage, but they are built differently, and that difference is where the friction lives. Polymarket is international and settles in crypto. Kalshi is US based, regulated, and settles in dollars. To run a trade across both, you are holding capital in two currencies on two systems that do not talk to each other.
That adds real cost and delay. You need KYC on each platform. You need funds sitting on both sides before an opportunity appears, because you cannot wait for a transfer to clear when a window lasts seconds. Moving money in or out takes time, and the currency mismatch means the price you saw is not always the price you get once conversion is factored in. For a deeper split of the two venues, see our Polymarket vs Kalshi comparison.
Why it is hard in 2026
The idea is old and the competition has caught up. On liquid markets, manual arbitrage is essentially impossible now. Here is what stands in the way.
- Windows close in seconds to minutes, and automated bots dominate the fast ones.
- Fees eat the edge. A 6 percent paper spread can net 1 to 2 percent after fees, sometimes less.
- Capital is a barrier. You need roughly $2,000 to $5,000 split across venues for the returns to be worth the effort.
- Spreads compress as more bots compete, so the gaps that do appear get smaller over time.
- Cross-venue settlement, KYC, and transfer time slow you down at exactly the moment speed matters most.
Put together, the clean arbitrage that looks good on paper is usually gone by the time a person can act on it.
A more realistic approach
For most retail traders, directional trading on events is more realistic than pure arbitrage. Instead of trying to hold both sides for a locked gain, you take a view on where an event is heading and enter one side when the price looks wrong. That does carry outcome risk, but it does not require two funded accounts, split capital, and sub-second speed.
The useful skill there is spotting dislocations, moments where price has not caught up to new information or where large traders are moving. SmartX is an AI terminal on the Polymarket ecosystem that surfaces fast moves and smart-money flow, which is what you watch for when you want to act on a mispriced event before the crowd does. It runs a 0.5 percent fee. If you want to see the flow yourself, you can open SmartX or start on Polymarket directly. For the broader picture, read how to make money on prediction markets.
Watch fast moves and smart-money flow on the Polymarket ecosystem in one place.
Open SmartX →Frequently asked questions
Is prediction market arbitrage legal?
Trading on regulated venues like Kalshi is legal in the US where the platform is permitted to operate, and Polymarket operates internationally. Arbitrage itself is a normal trading activity, not a loophole. What matters is that you meet each platform's rules and your own local regulations, including KYC. This is not financial advice, and access varies by region, so check what applies to you.
Is it really risk-free?
No. The locked profit only holds if both sides fill at the prices you saw, the event settles the same way on both venues, and fees do not swallow the gap. In practice prices move while you act, fills are partial, and transfers take time. Treat any claim of risk-free returns with suspicion. This is not financial advice.
How much money do I need?
To make the numbers worth the effort you need roughly $2,000 to $5,000 split across venues, since capital has to sit ready on both sides before an opportunity appears. Smaller balances get eaten by fees and the tiny per-pair edge. This is not financial advice.
Can I arbitrage Polymarket and Kalshi?
In theory yes, because they list overlapping events. In practice the two settle differently, one in crypto and one in dollars, which adds currency conversion, separate KYC, and transfer delay. Those frictions plus bot competition make clean cross-platform arbitrage very hard for a person doing it by hand. This is not financial advice.