
In recent years, prediction markets have grown from a niche toy into a notable segment of the crypto market. Wherever there are many platforms and different participants, prices inevitably diverge, which means arbitrage appears.
In this guide, we look at how yes/no contracts work, where to look for price gaps and why a scheme that is simple on paper requires reading the rules carefully.
Each market asks a question with a clear outcome: will an event happen by a certain date. Two contracts are issued for the question, "yes" and "no".
Once the market resolves, the contract for the correct outcome pays one dollar and the incorrect one pays zero. Until then, both contracts trade freely, and their price floats between zero and one.
If "yes" costs $0.63, the market on average puts the chance of the event at about 63%. The logic is simple: nobody will pay more than the expected payout.
This gives the main rule. In an ideal market, the price of "yes" plus the price of "no" equals exactly one dollar, since one of the two contracts is bound to pay out.
In practice, the sum wanders. Because of the order book spread, fees and uneven demand, it is sometimes slightly above one and sometimes below, and below one is where things get interesting.
Price gaps appear in three places, and each requires its own approach.
From what we see, windows within a single platform close the fastest. Bots catch them, so it is more realistic to work manually with gaps between platforms.
Bookmakers have their own math. The probability from decimal odds is calculated as one divided by the odds, but the line already includes the bookmaker's margin, and it has to be subtracted before comparing.
Type of arbitrage | Where to look | Window speed | Main risk |
Within a platform | One market, yes + no < 1 | Seconds | Bot competition, fees |
Between platforms | The same question on two markets | Minutes to hours | Different resolution rules |
Against a bookmaker | Sports and political events | Hours | Bookmaker margin, limits |
Linked markets | Questions whose outcomes depend on each other | Hours to days | A flaw in the logic of the link |
Take a textbook situation; all numbers are hypothetical. The question is the same, on platform A "yes" costs $0.46, and on platform B "no" is offered at $0.51.
We buy 1,000 contracts on each side and spend $460 + $510 = $970. Whatever the outcome, one of the bundles will pay $1,000, so the gross difference is $30, or about 3.1% on the capital invested.
Next, we subtract from that $30 the fees on both platforms and the cost of transferring and withdrawing funds. If together they eat, say, $12, you are left with $18.
The last part of the calculation is time. If the event resolves in four months, the capital is frozen the whole time, and the annualized return is more modest than it looks at first glance.
The main trap is different rules for resolving the outcome. Two platforms can ask a similar question but define the source of truth, the exact time or the interpretation of an edge case differently.
In our view, this is where the most money is lost. If one platform resolves "yes" and the other "no", both legs of the trade lose at the same time.
Start by observing. Write down a few popular questions that are listed on two platforms at once, and check the sum of "yes" and "no" between them every day.
Before your first trade, read the resolution rules of both markets in full. If the wording differs even slightly, treat the trade as risky and reduce your size.
We believe it is sensible to choose events with a near resolution date. Capital comes back faster, which means less time spent waiting and less chance that the rules change in the meantime.
The skill of finding the same thing at different prices on different venues is just as useful in prediction markets as on crypto exchanges, where all quotes are easy to see at once. Our arbitrage screener keeps dozens of venues in one window, refreshes quotes every second and shows each gap together with the volume actually behind it. The spread calculator helps you check what survives fees, network costs and slippage at your size. The bot is fully manual. It never connects to your exchange API keys.
To test the tools on a live market, ArbitrageScanner offers one day of free access to the whole ecosystem.
1. What is prediction market arbitrage?
It is buying opposite outcomes of the same event so that the combined price is less than the one-dollar payout. The difference remains whatever the outcome, as long as the resolution rules match.
2. How is the price on Polymarket related to probability?
The price of a "yes" contract in dollars roughly equals the probability of the event priced in by the market. A contract at $0.63 means an estimated chance of about 63%.
3. Why don't "yes" and "no" add up to a dollar?
The order book spread, fees and uneven demand get in the way. When the sum drops below a dollar, an arbitrage window opens.
4. Can you arbitrage between Polymarket and a bookmaker?
You can compare the contract price with the probability implied by the bookmaker's odds. But first you need to subtract the bookmaker's margin and account for its limits.
5. What is the main risk of prediction market arbitrage?
Different rules for resolving the outcome. If the platforms interpret the event differently, both legs of the trade can lose at the same time.
6. Is trading on prediction markets legal?
It depends on the country. Check the rules in your jurisdiction and the terms of the platform itself before depositing money.
Prediction market arbitrage rests on a simple rule: "yes" and "no" together should cost a dollar. When the sum drops below that, within a platform, between platforms or against a bookmaker line, a window opens.
But the real profit is calculated only after fees, withdrawals and the time capital is frozen. And before any calculation, compare the outcome resolution rules: they are what most often turns a seemingly risk-free trade into a double loss.
IMPORTANT! We are software developers. We do not give recommendations or promises of earnings and we do not advise you to invest your money anywhere. Our software is fully manual, all your money stays under your own control. We show examples of how our clients have earned on arbitrage in the past, but we do not advise repeating those actions one to one. Your earnings depend solely on your own actions and on market factors.
Get a subscription and access the best tool on the market for arbitrage on Spot, Futures, CEX, and DEX exchanges.
