Same Question, Different Prices
If you check the price of a political outcome across Polymarket, Kalshi, and PredictIt at the same time, they rarely agree exactly. One platform might price a candidate at 61 cents while another has them at 57 cents. Four cents might not sound like much, but on a binary contract that pays $1.00 or $0.00, a four-cent edge is significant.
These gaps exist because capital is not freely mobile between platforms. Moving money from Kalshi to Polymarket takes time and has friction costs. Different regulatory constraints and user demographics create natural price divergence. And sometimes platforms define resolution criteria slightly differently, which justifies some of the gap.
Research analyzing Polymarket transactions from April 2024 to April 2025 found that arbitrageurs extracted over $40 million in effectively risk-free profits. The top three arbitrage wallets alone captured $4.2 million. This is not theoretical. Real traders are actively profiting from these discrepancies.
The Mechanics of Executing an Arbitrage
A clean arbitrage requires buying YES on one platform and the opposite position on another, such that your total outlay is less than the guaranteed $1.00 payout. If Platform A prices "Will X happen?" at 55 cents YES, and Platform B prices the same event at 40 cents NO, your total cost is 95 cents for a guaranteed $1.00 payout. That is a 5.3% gross return regardless of the outcome.
In practice, you need to account for platform fees (typically 2-10% on profits), withdrawal costs, and the time value of locked capital. Polymarket charged no fees through most of 2025, only introducing taker fees in high-frequency crypto markets in January 2026. Kalshi has its own fee schedule. After all costs, a 5% gross arbitrage might net you 1-3%, and your capital is locked until the event resolves.
Automated Detection Is the Edge
The challenge with manual arbitrage scanning is speed. By the time you check six platforms, compare contract definitions, calculate net returns after fees, and fund your accounts, the opportunity may have closed. Most arbitrage windows last hours or days, not weeks.
Automated cross-platform matching that monitors prices across all major platforms simultaneously can flag discrepancies in real time, calculate net returns after platform-specific fee schedules, and surface opportunities before they close. The difference between scanning manually and having an automated system is the difference between occasionally catching an opportunity and systematically capturing them.
Risk Factors Worth Knowing
Even seemingly risk-free arbitrage carries risks that new traders overlook. Platform counterparty risk is real. If a platform becomes insolvent or freezes withdrawals, your locked capital is at risk regardless of your bet. Resolution disputes can leave one leg of your trade paying out while the other is contested.
There is also execution risk. By the time you buy on one platform and move to buy the other leg on a second platform, the price may have moved. In fast-moving markets, this slippage can turn a positive-expectation trade into a losing one.
The best approach treats prediction market arbitrage as a low-risk strategy, not a no-risk one. Size your positions accordingly, diversify across many small opportunities rather than concentrating on a few large ones, and always read the resolution criteria carefully before assuming two contracts on different platforms are truly equivalent.
Explore these tools on Blockcircle: Prediction Markets Mispricing Engine