How price discrepancies between Polymarket, Kalshi, and other platforms arise — and why "risk-free" arbitrage is rarer than it looks.
Test a gap yourself: the Kalshi vs Polymarket arbitrage calculator subtracts both fee bases, and the EV calculator shows what an edge is worth.
Arbitrage in prediction markets means buying an outcome on one platform and selling the same outcome on another platform at a different price — capturing a risk-free profit from the price discrepancy. The same event can trade at materially different prices across Polymarket, Kalshi, Manifold Markets, and other platforms.
For binary markets, arbitrage also exists within a single market when YES + NO prices don't sum to $1.00 (rare on Polymarket, more common on lower-liquidity platforms).
The main prediction markets where arb opportunities exist:
Arb between Polymarket and Kalshi is the most profitable real-money opportunity, as both platforms cover many overlapping US political and economic events but often at different prices due to different user bases and liquidity pools.
Track how prices move across related markets. When a correlated market moves significantly and a related market hasn't updated, an arb-like opportunity often emerges.
On Polymarket, market makers set the YES and NO prices. In efficient markets, YES + NO = $1.00 exactly (before fees). But on smaller or newly launched markets, the spread sometimes allows:
Such opportunities are rare and small on Polymarket's well-traded markets, but can appear on newly created markets or during high volatility when liquidity providers widen spreads.
⚠️ Resolution risk is the most dangerous. The exact wording of a market's resolution criteria matters enormously. A 1% price difference on a 6-month market may not be worth the risk that the two platforms interpret the resolution differently.
Execution risk is listed above as one bullet among five, but it is the one that actually decides whether an arbitrageur makes money. Arb edges close in seconds. The gap between seeing an opportunity and having both legs filled is where the profit disappears.
| Demand on you | Manual reality | What a written rule changes |
|---|---|---|
| Both legs at once | One leg fills, the other reprices while you click | Legs are triggered on your condition, not on your reaction time |
| Watching YES + NO drift below $1.00 | You are not watching at 3am | The equation is checked continuously |
| Exiting a locked position | You leave it because you missed the moment | The exit rule fires when the spread closes |
This is not a Polymarket-specific problem and there is no tool that makes Polymarket arb risk-free — resolution risk stays yours. What automation removes is only the speed problem: if you also trade crypto on an exchange, CoinRule runs written if-this-then-that rules on the exchange side without code. It does not place Polymarket orders.
A related but distinct strategy is narrative arbitrage: exploiting the lag between real-world news and prediction market prices. When a significant development occurs (e.g., a central bank policy shift, a geopolitical event, an earnings report), markets that are closely related to that event often update slowly.
These cascading price movements across correlated Polymarket markets are the core arbitrage signal to watch for.
When one market moves, related markets often lag. Ripple Effect identifies cross-market pricing discrepancies in real time.
Read: Liquidity & Spread →To arbitrage across Polymarket and Kalshi you need funded accounts on both. For Polymarket, the cheapest funding path is an exchange with native Polygon withdrawals → buy USDC via ACH (free) → withdraw on Polygon (~$0.90 flat fee). The comparison sets out every option and its fee.