Prediction Market Arbitrage: How to Find Risk-Free Profits

Arbitrage is the holy grail of trading: a risk-free profit generated by exploiting price differences between markets. In traditional finance, arbitrage opportunities are rare and fleeting because sophisticated algorithms close them in milliseconds. But prediction markets are a younger, less efficient asset class, and genuine arbitrage opportunities still appear regularly.

This guide covers the main types of arbitrage available in prediction markets, how to identify them, the math behind calculating guaranteed profits, the tools you can use to find opportunities, and the risks that make "risk-free" not always what it seems.

What Is Prediction Market Arbitrage?

Arbitrage occurs when you can buy and sell the same outcome (or complementary outcomes) at different prices across markets, locking in a guaranteed profit regardless of what actually happens. In prediction markets, this typically takes two forms:

1. Cross-platform arbitrage: The same event is priced differently on two different platforms (e.g., Polymarket vs. Kalshi)
2. Within-market arbitrage: The prices of all outcomes in a single market do not sum to 100%, creating a guaranteed profit opportunity by buying all sides

A Simple Example

Suppose the market "Will the Fed cut rates in June?" is priced at:

• Polymarket: Yes $0.72, No $0.28 (sums to $1.00)
• Kalshi: Yes $0.65, No $0.35 (sums to $1.00)

You could buy Yes on Kalshi at $0.65 and No on Polymarket at $0.28. Your total cost is $0.93. If the Fed cuts rates, your Kalshi Yes contract pays $1.00. If the Fed does not cut rates, your Polymarket No contract pays $1.00. Either way, you receive $1.00 for a $0.93 investment, netting a guaranteed $0.07 profit per contract pair (7.5% return).

That is arbitrage in its purest form: a locked-in profit with zero risk from the outcome itself.

Type 1: Cross-Platform Arbitrage

Cross-platform arbitrage exploits price differences between prediction market platforms. The same real-world event may be trading at different probabilities on Polymarket, Kalshi, and other platforms due to differences in trader composition, information flow, and liquidity.

How to Find Cross-Platform Arbs

1. Identify overlapping markets. Find events that are listed on multiple platforms. Common overlaps include:
   • U.S. election outcomes (Polymarket, Kalshi)
   • Fed interest rate decisions (Polymarket, Kalshi)
   • Crypto price milestones (Polymarket, various crypto prediction platforms)
   • Major political events (multiple platforms)
2. Compare prices in real time. Check the Yes and No prices on each platform for the same event.
3. Calculate the combined cost. Add the cheapest Yes price from any platform to the cheapest No price from any platform. If the total is less than $1.00, an arbitrage exists.
4. Execute both sides simultaneously. Buy the cheaper Yes on one platform and the cheaper No on the other.

Real-World Example with Math

Suppose "Will Bitcoin reach $200K by December 2026?" is priced as:

Platform	Yes Price	No Price
Polymarket	$0.38	$0.62
Kalshi	$0.33	$0.67

Arb calculation:

• Buy Yes on Kalshi: $0.33
• Buy No on Polymarket: $0.62
• Total cost: $0.95
• Guaranteed payout: $1.00 (one of the two will pay out)
• Guaranteed profit: $0.05 per contract pair (5.3% return)

If you deploy $1,000 on each side ($1,000 / $0.33 = 3,030 Yes contracts on Kalshi; 3,030 No contracts at $0.62 = $1,879 on Polymarket), your total investment is $2,879 and your guaranteed return is $3,030, netting $151 profit regardless of whether Bitcoin reaches $200K.

Frequency of Cross-Platform Arbs

Cross-platform arbs appear most frequently:

• After major news events: Different platforms absorb new information at different speeds. The first few minutes after breaking news often create temporary price dislocations.
• On less liquid platforms: Platforms with lower trading volume tend to have slower price discovery, creating arb windows against more liquid platforms like Polymarket.
• Around market close/resolution: As a market approaches its resolution date, prices can diverge if one platform's traders are more informed than another's.
• During off-hours: When trader activity is low (late night, weekends), prices may drift apart.

Type 2: Within-Market Arbitrage

Within-market arbitrage occurs when the prices of all outcomes in a single market do not sum to exactly $1.00. In a perfect market, if there are two outcomes (Yes/No), Yes + No = $1.00. If there are multiple outcomes, all prices should sum to $1.00.

When the sum is less than $1.00, you can buy all outcomes for less than the guaranteed $1.00 payout, creating a risk-free profit.

Example: Multi-Outcome Market

Suppose "Who will win the 2026 World Cup?" has the following prices:

Team	Contract Price
Brazil	$0.16
France	$0.14
Argentina	$0.13
England	$0.10
Spain	$0.09
Germany	$0.07
Other teams	$0.27
Total	$0.96

If you buy one contract for every outcome, you spend $0.96 and are guaranteed to receive $1.00 (since exactly one team will win). That is a risk-free $0.04 profit per contract set (4.2% return).

This type of arb is less common on well-functioning markets because market makers actively keep the sum close to $1.00. But it does appear, especially on markets with many outcomes or during periods of high volatility when prices shift rapidly.

The Overround Scenario

Conversely, if prices sum to more than $1.00, there is no arbitrage opportunity for buyers. Instead, there is an implicit cost (similar to a sportsbook's overround). This situation is actually the norm on some prediction platforms, where the sum of all outcome prices is $1.01-$1.03. On Polymarket, markets tend to be very efficient and typically sum to within $0.01 of $1.00.

Type 3: Related-Market Arbitrage

Sometimes, two markets are logically connected in a way that creates an arbitrage opportunity. This requires more analytical skill to identify but can be very profitable.

Example: Nested Markets

Consider two markets:

• "Will the Fed cut rates at least once in 2026?" trading at Yes $0.88
• "Will the Fed cut rates at least three times in 2026?" trading at Yes $0.52

If the Fed cuts rates three or more times, it has also cut at least once. So the probability of "at least three cuts" can never exceed the probability of "at least one cut." If the second market were priced at Yes $0.92 while the first was at $0.88, that would be logically impossible, and you could sell "at least three cuts" at $0.92 and buy "at least one cut" at $0.88 for a guaranteed profit in some scenarios.

Related-market arbs require careful analysis of the logical relationships between markets. They appear more frequently than cross-platform arbs but are harder to identify without systematic analysis.

Tools for Finding Arbitrage Opportunities

Manual Monitoring

The simplest approach is to manually check prices across platforms for overlapping markets. This is tedious but educational. Start by identifying 5-10 markets that appear on both Polymarket and Kalshi, and check the prices several times per day. You will quickly develop an intuition for when arbs appear and how quickly they close.

Spreadsheet Tracking

Build a spreadsheet that pulls prices from multiple platforms via their APIs. Set conditional formatting to highlight when the combined cost of all outcomes drops below $1.00. This is a step up from manual monitoring and can cover more markets simultaneously.

API-Based Bots

For serious arbitrage trading, build an automated system that:

1. Continuously polls prices from multiple platforms via APIs
2. Identifies arb opportunities in real time
3. Calculates optimal position sizes
4. Executes trades automatically on both platforms
5. Monitors for resolution and manages the lifecycle of arb positions

Both Polymarket and Kalshi offer REST and WebSocket APIs suitable for automated trading. Building a basic arb bot requires intermediate programming skills (Python is the most common choice) and a good understanding of the APIs.

Third-Party Tools

Several websites and tools aggregate prediction market prices across platforms, making it easier to spot discrepancies. These include:

• Prediction market aggregators that show side-by-side prices
• Discord and Telegram groups where traders share arb alerts
• Twitter/X accounts that post prediction market discrepancies

Be aware that publicly shared arb opportunities are often already stale by the time you see them. The most profitable approach is building your own monitoring system.

Calculating Arbitrage Returns

Simple Formula for Two-Outcome Arbs

For a binary market (Yes/No) with prices from two different platforms:

Arb Profit = $1.00 - (Cheapest Yes + Cheapest No)

Arb Return % = Arb Profit / (Cheapest Yes + Cheapest No) * 100

Capital Allocation Formula

To determine how much to allocate to each side of an arb:

If you want equal guaranteed profit regardless of outcome:

• Spend on Yes side: (Total Capital * Cheapest Yes) / (Cheapest Yes + Cheapest No)
• Spend on No side: (Total Capital * Cheapest No) / (Cheapest Yes + Cheapest No)

Example Calculation

Total capital: $5,000. Cheapest Yes: $0.40 (Platform A). Cheapest No: $0.55 (Platform B). Combined cost: $0.95.

• Spend on Yes: $5,000 * ($0.40 / $0.95) = $2,105
• Spend on No: $5,000 * ($0.55 / $0.95) = $2,895
• Total deployed: $5,000
• Yes contracts: 5,263 (at $0.40 each)
• No contracts: 5,263 (at $0.55 each)
• Guaranteed payout: 5,263 * $1.00 = $5,263
• Guaranteed profit: $263 (5.3% return)

Risks That Make "Risk-Free" Not Quite Free

While the math of arbitrage guarantees a profit in theory, real-world execution introduces several risks:

1. Execution Risk

You need to execute both sides of the arb before prices move. If you buy Yes on Platform A and the price moves on Platform B before you buy No, the arb may disappear. This is especially problematic during volatile periods when prices are moving rapidly.

2. Liquidity Risk

The arb price may only be available for a limited number of contracts. If you try to buy 10,000 contracts but only 500 are available at the arb price, you end up with an incomplete position. The remaining contracts may execute at worse prices, reducing or eliminating your profit.

3. Resolution Risk

Different platforms may resolve the same event differently. If the event's outcome is ambiguous or disputed, one platform might resolve Yes while another resolves No, resulting in a double loss instead of a guaranteed profit. Always read the resolution criteria carefully on both platforms.

4. Platform Risk

If one platform experiences technical issues, becomes insolvent, or changes its rules, your position on that platform may be at risk. Having capital on multiple platforms also means more exposure to smart contract risk (for decentralized platforms) or counterparty risk (for centralized platforms).

5. Capital Lock-Up

Arb positions tie up your capital until the market resolves, which could be weeks or months. A 5% return over 6 months is only about 10% annualized. If you could deploy that capital elsewhere at a higher return, the opportunity cost makes the arb less attractive.

6. Fee Risk

While Polymarket charges zero fees, other platforms charge significant fees that can eat into arb profits. Always factor in all fees (trading, settlement, deposit, withdrawal) when calculating arb returns. A 5% arb becomes a 1% arb (or worse) if one platform charges a 3-4% settlement fee.

7. Tax Complications

Arb profits across multiple platforms may create complex tax situations, especially if the platforms are in different jurisdictions or treat gains differently (gambling income vs. capital gains). Keep detailed records of all trades.

Advanced Arbitrage Strategies

Statistical Arbitrage

Instead of looking for guaranteed risk-free profits, statistical arbitrage identifies patterns where prices tend to converge over time. For example, if one platform consistently prices certain event types 2-3% lower than another, you can systematically buy on the cheap platform and sell (or buy the opposite side) on the expensive platform. Each individual trade carries risk, but over many trades, the expected value is positive.

Time-Based Arbitrage

Some markets on one platform resolve earlier than equivalent markets on another platform. If you can profit from the information revealed by an earlier resolution, you can trade on the second platform before it fully adjusts. This is more akin to fast information trading than pure arbitrage.

Market Making

While not arbitrage in the strict sense, providing liquidity by placing limit orders on both sides of a market and capturing the spread is a related strategy. On Polymarket's zero-fee platform, market making can be profitable if you manage your inventory risk effectively.

Frequently Asked Questions

How much capital do I need for prediction market arbitrage?

You can start with as little as a few hundred dollars, but arb profits are proportional to capital deployed. Most arb margins are 2-8%, so a $1,000 deployment generates $20-$80 in profit. To make arbitrage worth the effort and monitoring time, most serious arb traders deploy at least $5,000-$10,000 across platforms.

How often do arb opportunities appear?

On any given day, there are typically a few cross-platform arb opportunities between Polymarket and Kalshi, usually in the 1-5% range. Within-market arbs on multi-outcome markets appear several times per week. The best opportunities appear during and immediately after major news events when prices adjust at different speeds on different platforms.

Is prediction market arbitrage legal?

Arbitrage is a legitimate trading strategy and is not illegal in any jurisdiction we are aware of. However, the legality of trading on specific prediction market platforms varies by jurisdiction. Make sure you are legally allowed to trade on each platform you use before attempting arbitrage.

Can arbitrage opportunities be automated?

Yes, and many arb traders use automated bots. Both Polymarket and Kalshi offer APIs suitable for automated trading. Building a reliable arb bot requires programming skills, API integration, error handling, and careful testing. Start with manual arb identification and move to automation as you gain experience.

What happens if one side of my arb fails to execute?

If you only execute one side of an arb (e.g., you buy Yes on one platform but fail to buy No on another), you have an unhedged directional position. This is a real risk and is why execution speed matters. Always have both platform interfaces open and execute as simultaneously as possible. Some traders place the less liquid side first, since it is more likely to fail or get a worse price.

Does Polymarket's zero-fee structure help with arbitrage?

Enormously. Fees are one of the biggest profit killers for arb strategies. On Polymarket, every cent of the arb margin goes to you. On platforms with 3-7% settlement fees, many arb opportunities that exist in theory are unprofitable after fees. Polymarket's zero-fee model makes it the ideal anchor platform for any arb strategy.