Polymarket Arbitrage: Finding Mispriced Markets

What Is Arbitrage on Prediction Markets?

Arbitrage is buying and selling related assets to lock in a guaranteed profit. On prediction markets, the classic form is: if Yes trades at 60¢ on one exchange and No trades at 35¢ on another exchange for the same outcome, you buy both for 95¢ total and are guaranteed $1 payout — a risk-free 5¢ profit.

On Polymarket itself, the Yes + No prices for a binary market always sum to approximately $1.00, so this type of direct arbitrage doesn’t exist within the platform. However, several other mispricing opportunities do.

Type 1: Cross-Platform Arbitrage

The same event is often traded on both Polymarket and Kalshi (or other platforms). Prices can diverge by 3–5¢ due to different liquidity pools and participant bases. Buy the cheaper side on one platform and sell the other side on the other platform to lock in the spread.

The challenge: you need accounts on both platforms, capital tied up on both, and fast execution. The opportunity is most common around news events when one platform’s price adjusts faster than the other.

Type 2: Correlated Market Mispricings

Polymarket often has multiple related markets for the same event. For an NBA game, there might be: “Lakers win” (moneyline), “Lakers -5.5” (spread), and “Total over 225.5” (totals). These markets are mathematically related, but they’re priced by different sets of traders.

A mispricing occurs when the implied probabilities from correlated markets are inconsistent. For example, if “Lakers win” is at 70¢ but “Lakers -5.5” is at 60¢, and historical data shows that teams who cover -5.5 win 95% of the time, then the moneyline might be overpriced relative to the spread.

Exploiting this requires buying the underpriced market and (if possible) selling the overpriced one. Smart money wallets are particularly good at spotting these cross-market inconsistencies, which is why following their trades via whale alerts is valuable.

Type 3: Information Arbitrage

The most profitable form of “arbitrage” on Polymarket isn’t technically risk-free, but it’s the closest thing: trading on information that hasn’t been fully priced in yet.

When a smart money wallet with $3.7M in profit buys $25K of a market at 55¢, they likely have information or analysis suggesting the true probability is 70%+. The 15¢ gap between the market price and the true probability is the “arbitrage” — it’s not risk-free, but over many such trades, it’s highly profitable.

This is what Edge Radar’s daily picks systematize. By aggregating smart money positions and using AI to estimate true probabilities, the system identifies markets where the current price is significantly below the likely true probability. The 60.8% win rate over 102 picks confirms that this information arbitrage is real and exploitable.

Type 4: Time-Based Mispricings

Markets are often mispriced at specific times:

Post-news overreaction — When breaking news moves a market 10¢ in 5 minutes, it often overshoots. Smart money buys the reversion.
Low-liquidity hours — During US nighttime, European sports markets may have stale pricing because fewer traders are active.
Pre-resolution convergence — Markets approaching resolution often have small mispricings as they converge toward 0¢ or 100¢. A market at 96¢ with certain resolution at $1 offers 4% return in hours.

Using Smart Money Data for Arbitrage

The most practical approach for individual traders isn’t pure arbitrage — it’s using smart money signals to identify markets where the current price is wrong: