How prediction markets resolve and handle disputes

Resolution turns contract wording into a payout

Resolution is the step where a traded contract becomes a payout. For ordinary binary contracts, the winning outcome pays 1 and the losing outcome pays 0. Treat that 1/0 pairing as a default assumption, not a law. Some contracts allow a rare unknown or 50-50 outcome that may pay 0.50 per share, which is not the same as a blanket refund or cancellation policy. What actually decides the outcome is the rules text: the named source, the deadline, the qualifying threshold and the edge cases.

A short title hides most of that. A question like whether something will happen by December does not tell you which source counts, which timezone the deadline uses, or how a revised report is treated. 2 markets on the same headline can resolve differently if their rules differ. On Polymarket, ordinary markets resolve through UMA, the oracle that proposes and settles outcomes, while up/down markets resolve through a Chainlink time-weighted average price (TWAP), meaning an average price over a window rather than a single snapshot. Those are separate paths, so do not apply the UMA flow to every market type.

An illustrative contract: official certification by a deadline

Take a hypothetical contract worded like this: the market resolves Yes if the official certifying body publishes a certification that the threshold was met by 2026-12-31 23:59 UTC, based on that body's published report. An early projection, forecast or unofficial statement does not qualify. 4 things are fixed here. The source is the certifying body's published report. The deadline is the UTC timestamp. The qualifying threshold is whatever the rules state, such as a named percentage or count. And the wording explicitly excludes an early projection, so an October forecast that later proves right still fails.

This example is illustrative only, not an empirical study of any real market. Its purpose is to show how wording creates a decision procedure. If the certification appears before the deadline, Yes can resolve to 1 and No to 0. If nothing appears by the deadline, the ordinary binary path makes No the winner. If the rules include a provision for unknown outcomes, or if the certifying body's own status becomes ambiguous, a 50-50 or another defined method may apply instead. Only the actual rules text tells you which branch you are in.

UMA proposal and dispute flow

In ordinary Polymarket markets, UMA handles resolution. UMA has proposal and dispute flows. A proposed outcome may be disputed rather than accepted as final. Check the current proposal and dispute state instead of assuming settlement the moment a proposal appears, because the first proposal is not automatically final.

A dispute does not imply a refund. The contract rules and the resolution system define what happens when a proposal is challenged. An exceptional unknown or 50-50 outcome can exist, but it is not the default and it is not a universal cancellation policy. Until resolution is final, your position stays tied to the payout structure written into the contract.

Up/down and TWAP markets follow a different path

Up/down markets do not use the UMA proposal and dispute flow. They compare a time-weighted average price (TWAP) at the start of the period with the TWAP at the end, using the same named asset stream. If the final TWAP is greater than or equal to the start TWAP, UP wins. If it is lower, DOWN wins. Because different exchange candles can produce different averages, the exact stream and observation window named in the rules matter.

Do not stretch this to every duration or every price source. A one-hour market and a daily market can share the TWAP idea while using different windows and different comparison rules. Before trading, identify which resolution path applies and which source the rules name.

Reading market details without fooling yourself

Market details include the question, the outcomes, outcome prices, token identifiers, activity, closure, order acceptance, liquidity and volume. Some fields arrive as serialized arrays that need parsing before they are readable. Always note the observation date of any price you use, because prices move and a stale number is not a current one. Volume and liquidity measure different things, and either can be stale or missing; a missing value is not the same as zero.

Public market data and trading access are separate paths. Reading market details needs no authentication, but placing orders does. For resolution, the fields that matter most are the rules text and the current proposal state. Use the official Polymarket prediction markets documentation to find the rules, and read the prediction market risks overview before assuming a dispute protects your position. Whether a given country allows access is a separate question covered by the prediction market availability guide.

Worked illustrative decision from the rules

The case below is a hypothetical worked decision, not an empirical study and not a record of any real market. It shows how the wording, source, timezone, qualifying threshold and dispute status change what you should conclude. It does not recommend a trade.

Limits and what to verify next

The worked case is hypothetical. Before taking a position, check the current rules and dispute state of that specific market, alongside the terms applicable to your location.

Before you trade or decide, open that market's own rules, decide whether it is an ordinary UMA-resolved market or an up/down TWAP market, check the current proposal and dispute state if shown, and confirm the source, deadline and qualifying threshold in the exact wording. If any of those is missing, treat the resolution as uncertain rather than assuming a 1/0 payout, a refund or a cancellation. The losing case is real: holding a position that the rules do not entitle you to, for example on an early projection that never becomes official certification, ends at 0 for that side.