How Do Prediction Markets Settle Contracts?
Prediction markets settle contracts by determining the outcome of an event, then paying $1 per share to winning positions and $0 to losing ones. Kalshi resolves off-chain through its CFTC-regulated process; Polymarket resolves on-chain through the UMA optimistic oracle.
Prediction markets settle contracts in two distinct steps: resolution (deciding what actually happened) and settlement (moving funds to winning positions). A binary contract pays $1 per share to the winning side and $0 to the losing side. The mechanics differ sharply between regulated venues like Kalshi and on-chain venues like Polymarket, and those differences matter for any operator integrating prediction-market data or trading.
The Two-Step Anatomy of Settlement
Every prediction-market trade ends the same way — winners get paid, losers don't — but the path has two distinct stages:
- Resolution is the determination of which outcome occurred. It requires a trusted source, a defined window, and a dispute mechanism.
- Settlement is the financial movement that follows. Funds held in escrow flow to winning shareholders; losing shares become worthless.
Conflating these two is the most common source of confusion. Resolution can be slow even when settlement is fast, and vice versa.
How Kalshi Settles
Kalshi operates as a CFTC-regulated Designated Contract Market (DCM) and clears trades through Kalshi Klear, its own CFTC-registered Derivatives Clearing Organization (DCO) since 2024. Settlement looks like a traditional derivatives flow:
- Resolution criteria are written into the contract specification and published before the market opens. Each contract names a primary data source (e.g., the Bureau of Labor Statistics for an inflation print).
- Resolution timing is typically immediate once the named source publishes the data.
- Settlement occurs within a few hours of resolution, with USD moving between member accounts inside Kalshi Klear.
- On-chain mirroring is increasingly used: finalized results are published through oracle networks like Pyth and RedStone so other systems can consume them.
Disputes are handled through Kalshi's internal market-integrity team and, ultimately, CFTC oversight. This is the same model used for any regulated futures product.
How Polymarket Settles
Polymarket runs on the Polygon blockchain and outsources resolution to UMA's optimistic oracle:
- Proposal. Anyone can propose an outcome by posting a USDC bond. If no one disputes within the challenge window (often two hours), the proposal stands and the market resolves.
- Dispute. A challenger posts their own bond to dispute. The dispute escalates to a vote among UMA token holders. The losing side forfeits their bond.
- Settlement. Once the oracle returns a final answer, smart contracts release USDC to winning share holders. There is no separate clearinghouse — settlement is the on-chain transfer.
- Timing. Two hours minimum if uncontested; days or weeks for disputed markets.
This model is transparent but slower in adversarial cases, and the dispute outcome depends on a tokenholder vote rather than a regulatory body.
Side-by-Side Comparison
| Dimension | Kalshi | Polymarket |
|---|---|---|
| Resolution venue | Internal market integrity team + named data sources | UMA optimistic oracle |
| Custody during the market | Kalshi Klear (CFTC-registered DCO) | Smart-contract escrow on Polygon |
| Settlement timing | Hours after resolution | 2 hours minimum, longer if disputed |
| Dispute process | Internal review, CFTC backstop | Bonded challenge + UMA tokenholder vote |
| Currency | USD | USDC |
| Transparency | Audited financials, regulatory reporting | Fully on-chain |
| Failure mode | Operational error, regulatory action | Oracle capture, contested vote |
Why Settlement Mechanics Matter for Operators
For B2B operators using prediction markets as a content feed, an intelligence layer, or a trading product, settlement shapes both UX and risk profile.
- Display semantics. A market that says "resolved" on Polymarket may still be in a challenge window. You need to know when it closes, resolves, and pays.
- Hedging windows. Operators laying off risk upstream must model the gap between their own customer payout and upstream settlement. Polymarket-style settlement can leave operators carrying risk for days on contested events.
- Reconciliation. Settlement data is the canonical record. Pulling it cleanly — through Kalshi's API or by indexing on-chain events — is what lets an operator close its books daily.
- Dispute exposure. When a market is contested, downstream operators who already paid users can be left with no upstream payout. Reserve policies have to be designed in, not bolted on.
What's Changing in 2026
Settlement is one of the areas the CFTC's 2026 prediction-market rulemaking is targeting. Expect harder requirements around named data sources, dispute timelines, and clearinghouse standards. On-chain venues are likely to be pushed toward stronger oracle guarantees — or toward partnering with registered DCOs — to serve U.S. customers under a federal framework.
The practical takeaway: resolution and settlement belong on the diligence checklist alongside liquidity and category coverage. A prediction-market integration without a clear answer to "how does this resolve and when does the money move?" is not production-ready.
Last verified: May 2026