Prediction Market Arbitrage

Every contract on Polymarket and Kalshi settles at $1 if the event happens and $0 if it does not. YES and NO are therefore two halves of the same dollar: buy both and you are guaranteed exactly $1 back. The trade only makes sense if the two halves cost less than a dollar together, which happens when the venues disagree.

Take a real shape from this scanner. If Kalshi offers YES at 46¢ and Polymarket's book means NO costs 48.8¢, the pair costs 94.8¢ and returns $1 — a gross return locked the moment both legs fill. Fees take a bite out of that, which is why the figures here are net.

This is a different trade from buying whichever venue is cheaper. That is price shopping: useful if you already have a view, but you still lose when the event goes the other way. Arbitrage has no view at all — you hold both sides and collect the difference. Full mechanics are in our guide to arbitraging prediction markets.

In the scanner, cost is what both legs together take out of your pocket with fees included — anything under 100¢ is an edge. Return is the profit on that committed capital, and annualised scales it to a year, since capital is locked until the market resolves.

In the comparison below, each row is one outcome priced on both venues in cents, where 52.0¢ means a 52% implied chance. The cheaper cell is outlined in that venue's colour. A blank difference means Kalshi's bid–ask is wider than five points, and a midpoint from a book that thin would invent a disagreement rather than report one. New to the mechanics? How prediction markets work starts from the beginning.

Execution risk is the big one. An arbitrage is only locked once both legs fill. Get one and miss the other and you are holding a directional position you never wanted, at a price you chose for reasons that no longer apply. On thin books this is the normal outcome, not the exception.

Resolution risk is the one that actually costs people money. The two venues write their rules separately, and a question that reads identically can settle differently — different sources, different cutoff times, different treatment of an ambiguous outcome. When that happens you do not break even, you lose both legs.

Fees are deducted using Kalshi's published formula and Polymarket's taker rate; both are detailed in our Kalshi fees explainer and Polymarket fees explainer.