What is arbitrage on Polymarket?

Arbitrage on Polymarket means exploiting pricing inconsistencies between related markets or outcomes to lock in a guaranteed profit regardless of the outcome. For example, if a multi-outcome market's probabilities sum to more than 100%, you can sell all outcomes and pocket the difference.

Are Polymarket arbitrage opportunities common?

Pure arbitrage opportunities on Polymarket are rare and short-lived. Most are eliminated within seconds by automated bots. When they do appear, the profit margins are typically small (0.5-3%). They occur most frequently during high-volatility events when manual traders create temporary mispricings.

How much capital do I need for Polymarket arbitrage?

Because individual arbitrage profits are small (often under 2%), you need significant capital to make meaningful returns. Most successful arbitrage bots on Polymarket deploy $10,000-$100,000. With smaller capital, the absolute dollar returns may not justify the development and operational costs.

Can arbitrage bots lose money on Polymarket?

True arbitrage is risk-free by definition, but execution risk can cause losses. If only one leg of a multi-leg arbitrage trade fills and the other doesn't, you're left with an unhedged position. Network congestion, API latency, and order book changes between legs are the primary risks.

Polymarket Arbitrage Bot Guide: Finding Risk-Free Profits

How arbitrage bots exploit pricing inefficiencies on Polymarket — strategies, implementation, and honest limitations.

What Is Prediction Market Arbitrage?

Arbitrage is the simultaneous purchase and sale of related assets to profit from a price discrepancy. In prediction markets, arbitrage opportunities arise when the prices of related outcomes are inconsistent with mathematical certainty.

The simplest example: a binary market has YES and NO outcomes. If YES trades at $0.55 and NO trades at $0.48, the combined cost is $1.03. Since exactly one outcome will pay $1.00, buying both guarantees a $0.03 loss. But if you sell both (if you already hold them), you collect $1.03 for something worth exactly $1.00 — a risk-free $0.03 profit per share.

More commonly, arbitrage appears in multi-outcome markets where the sum of all outcome prices deviates from 100%.

Types of Polymarket Arbitrage

1. Multi-Outcome Overround Arbitrage

Multi-outcome markets (e.g., "Who will win the Oscar for Best Picture?") have multiple possible outcomes that should sum to 100%. When they sum to more than 100% (overround), you can sell all outcomes and lock in a profit.

Example: Overround Arbitrage

A market with 5 outcomes has these YES prices:

Movie A: $0.35
Movie B: $0.28
Movie C: $0.22
Movie D: $0.12
Movie E: $0.08

Sum: $1.05 — This exceeds $1.00 by $0.05. By selling YES on every outcome (or equivalently, buying NO on every outcome), you collect $1.05 and will pay out exactly $1.00 when one outcome wins. Guaranteed $0.05 profit per share set.

2. Underround Arbitrage

When outcome prices sum to less than 100%, you can buy all outcomes cheaply. This is rarer because market makers typically ensure slight overround, but it appears during rapid price movements when some outcomes haven't adjusted yet.

3. Cross-Market Arbitrage

Related markets sometimes imply contradictory probabilities. For example, if "Candidate X wins the primary" is priced at 40% but "Candidate X wins the general election" is priced at 45%, there's an inconsistency — winning the general requires winning the primary, so the general price should be lower. Exploiting this requires careful position sizing across both markets.

4. YES/NO Spread Arbitrage

In binary markets, the YES and NO prices should sum to approximately $1.00 (minus fees). When the spread widens beyond the fee structure, a risk-free trade exists. This is the most common and smallest arbitrage type — typically just fractions of a cent.

Building an Arbitrage Bot

An effective arbitrage bot has four core components:

Component 1: Market Scanner

Continuously scan all active markets for arbitrage opportunities. For multi-outcome markets, sum all YES prices and flag any market where the sum deviates from $1.00 by more than your minimum profit threshold (accounting for fees).

Poll the CLOB API for all market prices every 1-5 seconds
Or maintain WebSocket connections for real-time price updates
Calculate the theoretical arbitrage profit after fees for each opportunity
Filter out opportunities below your minimum profit threshold

Component 2: Opportunity Validator

Not every price discrepancy is a real opportunity. Before executing, validate:

Liquidity check — Is there enough volume at the quoted prices to fill your desired size? Thin order books mean the actual execution price may differ from the quoted price.
Fee calculation — Polymarket charges fees on trades. Your arbitrage profit must exceed total fees across all legs.
Execution feasibility — Can all legs execute simultaneously? If one leg fills but another doesn't, you're exposed to directional risk.
Market resolution risk — Is the market about to resolve? Arbitrage in markets resolving within hours carries additional settlement risk.

Component 3: Execution Engine

Speed is critical. Arbitrage opportunities disappear within seconds as other bots compete for the same mispricing. Your execution engine must:

Submit all legs of the arbitrage trade as close to simultaneously as possible
Use limit orders at the identified prices (not market orders, which may fill at worse prices)
Monitor fill status for all legs and handle partial fills
Implement rollback logic — if one leg fails, attempt to cancel or unwind the other legs

Component 4: Risk Manager

Even "risk-free" arbitrage has execution risk. Your risk manager should:

Track net exposure across all open arbitrage positions
Set maximum capital allocation per arbitrage trade
Monitor for stuck positions (one leg filled, other pending) and alert immediately
Calculate realized P&L per trade and cumulative performance

Real-World Limitations

Arbitrage sounds like free money, but several factors limit its profitability in practice:

Why Pure Arbitrage Is Hard

Competition — Dozens of sophisticated bots scan for the same opportunities. The fastest bot wins, and latency differences of milliseconds matter.
Fees eat profits — Polymarket's fee structure means you need price discrepancies larger than the round-trip fee to profit. Many apparent opportunities vanish after fees.
Execution risk — Multi-leg trades can partially fill, leaving you with unintended directional exposure. This is the biggest real-world risk.
Capital efficiency — Your capital is locked in positions until market resolution, which can be weeks or months. The annualized return on locked capital may be low.
Diminishing opportunities — As more bots enter the market, arbitrage opportunities become rarer and smaller. The market becomes more efficient over time.

Expected Returns

Be realistic about arbitrage returns on Polymarket:

Per-trade profit — Typically 0.5-3% before fees, 0.1-1.5% after fees
Frequency — A well-tuned bot might find 5-20 opportunities per day across all markets
Capital requirement — $10,000+ to generate meaningful absolute returns
Monthly return — Realistic expectation is 1-5% monthly on deployed capital, not the 50%+ returns sometimes claimed online

For most traders, combining arbitrage with other strategies (like market making) produces better risk-adjusted returns than pure arbitrage alone.

Getting Started

If you want to build an arbitrage bot, start with the API tutorial to understand the CLOB interface, then implement a scanner that monitors multi-outcome markets for overround conditions. Paper trade for at least two weeks before deploying real capital.

Want to Copy Top Polymarket Traders Automatically?

Polycool lets you follow the best wallets and copy their trades in one tap. No manual tracking needed.

Try Polycool Free →