A contract is only as clear as the rule that resolves it. Before you trade an outcome, the more useful question is who reports the deciding value and what criteria turn that value into a yes or a no. That rule, not the headline, is what pays you.
A settlement source is the named authority whose data decides the outcome. It is written into the contract before trading. On regulated exchanges it is a named source agency; on oracle based venues it is criteria the oracle verifies.
Per reporting in February 2026, the Super Bowl halftime market settled at the last traded price on Kalshi but resolved yes at one dollar on Polymarket. Same event, different rules, different result.
No resolution model is flawless, and disputes happen on both. The rule that resolves a contract deserves the same attention as the price. Read the criteria before you hold to settlement.
A settlement source is the named authority whose data decides which side of a contract wins, and it is chosen before trading and written into the contract terms. On a regulated exchange like Kalshi the source is a named source agency, a league, a benchmark provider, or a government body, filed with the CFTC, and an internal team applies it. On Polymarket the criteria are stated in the market and verified through the UMA Optimistic Oracle, which can escalate a disputed result to a token holder vote. Because the rule, not the headline, decides the payout, the same event can settle differently on two venues, so the criteria deserve as much attention as the price.
A settlement rule first names who reports the deciding value. It might be a sports league, a news agency, a benchmark index provider, or a government statistics office. Naming the source in advance is what stops the outcome from turning on opinion after the event, and it is why two markets on the same question can still resolve differently if they name different sources.
Beyond the source, the rule states precisely what counts. A market is not asking the loose version of its title; it is asking the specific condition written into the terms, down to the timing window and the edge cases. The difference between a casual reading and the exact criteria is where most settlement surprises live, because real events rarely fit a yes or no as neatly as the headline suggests.
Many sources produce a value at a specific moment, and the rule says which moment counts. A crypto price market may take a short observation window at expiry rather than a single tick, and trim the extremes to resist manipulation. An economics market waits for a scheduled release. Knowing the window tells you when a market can actually resolve and what data it will read.
A good rule also says what happens when the criteria cannot cleanly resolve. On a regulated exchange that can mean an internal review committee, or a defined fallback such as settling at the last traded price. On an oracle based venue it can mean escalation to a vote. A fallback is not a prediction of the outcome; it is a defined way to close a market the criteria cannot settle.
A general settlement flow for a binary event contract. The step that decides who is paid is the source reporting a value and the criteria applying it, not the headline of the market. Illustrative, as of June 2026, not a depiction of a specific platform's interface.
| Market category | Typical named source | How the value is taken |
|---|---|---|
| Sports | The governing league, the Associated Press, ESPN, and similar outlets named per contract. | The official final result as reported by the named source after the event concludes. |
| Crypto prices | A benchmark index provider, for example CF Benchmarks real time indices. | A short observation window at expiry, with a trimmed average that excludes the most extreme observations. |
| Economics | Government statistics bodies such as the Bureau of Labor Statistics, the Bureau of Economic Analysis, and the Federal Reserve. | The official data release at the scheduled publication time. |
| Weather | Official meteorological bodies such as NOAA and the National Weather Service. | The official recorded measurement at a specified station and time. |
Methodology: categories and sources summarised from reporting collected by DeFi Rate, as of April 2026, and from Kalshi contract terms filed with the CFTC. Sources are named per individual contract and can vary within a category, so always read the specific market's terms rather than relying on the category. This is general information, not advice.
On Kalshi, the source is not improvised at resolution time. Per reporting collected by DeFi Rate in April 2026 and Kalshi's own contract terms, every market names one or more source agencies in terms filed with the CFTC as part of the exchange's self certification process. The expiration value, in the language of the terms, is the value of the underlying as documented by the source agency at the stated expiration time. The naming happens before you can trade, which is the point: the deciding authority is fixed in advance and visible in the contract.
Applying that source is a human step. Reporting indicates Kalshi's internal markets team decides when the resolution criteria have been met, and that a trader can submit a request to settle that functions as a prompt rather than a binding instruction. When an outcome is ambiguous or disputed, the rulebook provides for an Outcome Review Committee that can make binding determinations, and a fallback rule that allows a market judged unresolvable to settle at its last traded price. The mechanics of that filing process are covered in our guide to self certification of contracts.
The honest tension reporters and at least one state regulator have raised is structural: on a centralised exchange, the same entity writes the contract, names the source, and grades the result, with no independent appeal for the trader. Reporting in 2025 and 2026 documented cases where corrections followed public pressure rather than a built in review. None of this means the model is unsound, and the named sources and CFTC oversight are real strengths, but it does mean the rule and the source are worth reading before you rely on them.
On Polymarket, the source still has to be named, but the verification is decentralised. Per Polymarket's documentation as of 2026, each market states its resolution source, end date, and edge cases in the market description. When the date arrives, anyone can propose an outcome by posting a bond, and a short challenge window opens. If no one disputes, the proposal is accepted and the market settles. If a proposal is disputed twice, the question escalates to UMA's Data Verification Mechanism, where token holders vote on the correct outcome, a process reporting describes as taking roughly 48 to 96 hours.
Reporting indicates the large majority of markets, on the order of 98.5 percent per one cited analysis, resolve at the oracle layer without ever reaching a token vote. The problems surface when outcomes require interpretation. Because the vote is weighted by token holdings rather than one person one vote, commentators have flagged the risk that a large holder can sway a contested result, and the bond required to dispute a proposal can discourage smaller traders from challenging a resolution they believe is wrong. Polymarket clarifications also function as binding context within the system, which keeps meaningful influence with the platform even under a decentralised process.
On a regulated exchange like Kalshi, an internal markets team applies the source agencies filed with the CFTC, with an Outcome Review Committee as a backstop and a last price fallback rule for the unresolvable case. The strength is named sources and regulatory oversight; the risk reporters note is that the exchange both writes the rule and grades it.
On Polymarket, the UMA Optimistic Oracle lets anyone propose an outcome against a bond, with disputes escalating to a vote by token holders. Per reporting, most markets resolve without a vote. The strength is an open, on chain process; the risks reporters note are token concentration and the cost of disputing a result.
A neutral summary of two resolution designs as described in platform documentation and reporting, as of June 2026. Neither model is endorsed here, and neither has a flawless record. This is general information, not advice.
The clearest reason to read the rule is that two venues looking at the same event can reach opposite conclusions. The reported example is the 2026 Super Bowl halftime market. Per reporting in February 2026, both Kalshi and Polymarket ran a market on whether a particular artist would perform during the halftime show. After the performer appeared on stage and the facts were genuinely ambiguous, Kalshi treated the outcome as unresolvable and invoked its last price fallback, settling yes holders at about twenty six cents and no holders at about seventy four cents. Polymarket resolved the same question yes at one dollar. The divergence was not caused by different information; both venues saw the same broadcast. It was caused by different criteria and different interpretation standards.
This is contested territory, and the point here is not that one venue was right. It is that the rule, not the event, produced the result, and a trader holding the same view on both venues could have been paid in full on one and partially on the other. Reporting has documented disputes on both models, from contested token votes on oracle based markets to graded results reversed only after public attention on centralised ones. We cover the broader pattern in our guide to resolution disputes and how they work.
Contested, and dated. The specific outcomes above are as reported in 2026 and the interpretation of ambiguous events is, by nature, disputed. Treat cross venue divergence as a real risk rather than a settled fact about either platform, and verify the live contract terms yourself before holding to settlement.
One more thing decides your settlement rule, and it is easy to miss: many consumer apps do not run their own exchange. Per reporting collected by DeFi Rate, as of April 2026, a number of well known apps route their trades through an underlying exchange and inherit that exchange's resolution criteria, source agencies, and dispute process. If an app routes through Kalshi, your market settles under Kalshi's rules and named sources, not the app's branding. Others route through different exchanges entirely, which means two apps that look similar can settle the same event under different rulebooks.
The practical lesson is to identify the exchange behind the app before you treat a market as settled by the name on the screen. The interface is the front door; the resolution rule lives with the exchange the order actually reaches. When in doubt, the contract terms will name the exchange and the source, which is the level at which settlement is genuinely decided.
If you intend to hold a contract through resolution rather than trade out beforehand, the resolution criteria deserve the same study as your entry price. Read the actual contract language, not the market title, because the title is a summary and the terms are the rule. Identify the named source and the timing window, and treat a vague or undefined source as a risk in its own right. Where the same market exists on two venues, read the resolution language side by side; subtle wording differences are exactly what produce divergent outcomes when an event does not fit cleanly into yes or no.
Know which process will resolve your market and what recourse exists if it is graded in a way you dispute. On a centralised exchange that generally means the internal team with a review committee as backstop, and informal support channels rather than formal arbitration. On an oracle based venue it means the proposal and dispute process, with a bond required to challenge and a token vote as the final step. Neither offers a guaranteed appeal, so the realistic protection is choosing markets whose criteria you understand. For how resolution timing fits the wider contract life, see our guides to how event contracts settle and the market categories where these rules apply.
Understanding how a market resolves does not make a position safe or a result certain. A contract can settle against you even when your read of the event felt right, and ambiguous outcomes can resolve in ways you did not expect. Prediction markets can lose you money. Stake only what you can afford to lose, never to chase a loss, and never on borrowed money. If it stops feeling like a free choice, step back. In the United States you can call or text the helpline on 1-800-GAMBLER or visit ncpgambling.org.
A settlement source is the named authority whose data decides which side of a contract wins. On a regulated exchange it is often a specific agency, for example a sports league, the Associated Press, a benchmark provider, or a government statistics office. The source is written into the contract terms before trading, so the question is settled by a rule agreed in advance rather than by opinion after the fact.
Per reporting collected by DeFi Rate in April 2026 and Kalshi's own contract terms, every Kalshi market names one or more source agencies in terms filed with the CFTC as part of the exchange's self certification. The expiration value is the value the source agency documents at the stated time. Kalshi's internal markets team applies the criteria, with an Outcome Review Committee as a backstop for disputed cases.
Polymarket uses the UMA Optimistic Oracle. Per Polymarket's documentation as of 2026, anyone can propose an outcome by posting a bond, a short challenge window opens, and if disputed twice the question escalates to a vote by UMA token holders through the Data Verification Mechanism. Reporting indicates the large majority of markets resolve at the oracle layer without ever reaching that vote.
Because the resolution criteria, not the headline, decide the outcome, and those criteria can differ. The clearest reported example is the 2026 Super Bowl halftime market: per reporting in February 2026, Kalshi treated the outcome as ambiguous and settled at the last traded price, while Polymarket resolved it yes at one dollar. Same event, different rules, different result. Read the contract language on each venue before assuming they match.
It is a rule a venue can invoke when an outcome is judged unresolvable under the stated criteria. Per reporting in 2026, Kalshi's Rule 6.3(c) allows a market to be settled at its last traded price in that situation. A fallback rule is not a prediction of the outcome; it is a defined way to close a market when the criteria cannot cleanly resolve it, and it can leave both sides with a partial result.
Read the actual resolution criteria, not the market title, and find the named source. On a regulated exchange the criteria sit in the contract terms; on an oracle based venue they sit in the market description and on chain data. Note which process resolves the market and what recourse exists if it is graded in a way you dispute. If the source or criteria are vague, treat that as a risk in itself.
The rules change fast. Get the changes that affect you, plain and current, not tips.
Independent. Every claim dated and sourced. No platform pays for its place.