Informational only. Not financial advice. Prediction markets involve risk and you can lose your entire stake. Check that any platform is legal and available where you live.
The short version: Resolution is the moment a prediction market decides which outcome was correct and pays out. Someone, or something, has to say “Yes” or “No”, and the way a platform does that matters as much as the price you traded at. Most disputes don’t come from fraud. They come from vague wording, unclear sources and edge cases nobody thought about when the market was created.
If you’re new to this world, start with our beginner’s guide to prediction markets. If you want to understand what prices mean before you worry about payouts, read how to read prediction market odds first.
Why resolution matters more than most beginners think
Most new traders focus on the question: “Will this happen?” Experienced traders ask a second question: “Who decides whether it happened, and how?”
A contract is only as good as its resolution. If the rules are clear and the source is reliable, you are betting on the real world. If the rules are fuzzy, you are also betting on how someone will interpret them. That second bet is often invisible in the price, but it is there.
Think of a football match. If the rules say “the team that scores more goals wins,” that’s easy. If the rules say “the best team wins,” you’ll be arguing for days.
The basic life of a market
Almost every market follows the same steps, whatever technology runs behind it:
- Creation: A question is written, with an end date and resolution rules.
- Trading: People buy and sell shares while the outcome is still unknown.
- Event or deadline: The event happens, or the deadline passes.
- Proposal: An outcome is proposed by the operator, a data feed, or a participant.
- Challenge window: In some systems, there’s a short period where others can object.
- Dispute, if any: A contested outcome goes to a higher authority, such as a committee, a vote, or a court-like system.
- Settlement: Winning shares are redeemed for the full payout, and losing shares become worthless.
Steps 5 and 6 only matter when people disagree. In the vast majority of markets, the answer is obvious and nobody objects.
The main resolution models
Platforms use different designs. Here is a plain-language map.
| Model | Who decides | Typical strength | Typical weakness |
|---|---|---|---|
| Operator-run | The exchange, following a published rulebook | Clear accountability, often with regulatory oversight | You must trust one company’s judgment |
| Committee or admin | A team or panel, sometimes with community input | Can handle messy, human situations | Possible bias, slower, less transparent |
| Optimistic oracle | Anyone proposes an answer with a deposit; if no one objects, it stands | Cheap and fast when the outcome is clear | Disputes can be slow, and votes can be influenced |
| Staked reporter system | Token holders stake funds on outcomes, with escalating challenges | Strong economic incentives to be honest | Complex, slow and expensive when contested |
| Crowd answer with escalating bonds | Anyone answers; challengers must post bigger bonds each round | Open and self-correcting in theory | Needs a backup arbitrator for hard cases |
| Data-feed oracle | An automated feed that reports objective numbers | Fast and objective for prices and scores | Only works for machine-readable outcomes |
| Hybrid | Several methods combined | Flexibility | More moving parts to understand |
Operator-run resolution
On regulated exchanges, the rulebook usually names the contract, the data source and the exact payout condition. The exchange then applies those rules. It’s simple and predictable, and there is often a formal process to contest a decision. The trade-off is trust: you rely on the operator and its compliance process.
Optimistic oracles
The word “optimistic” means the system assumes the answer is right unless someone challenges it. A typical flow looks like this:
- A person proposes an outcome and posts a deposit (a bond).
- A short challenge window opens.
- If nobody disputes, the answer is accepted and the proposer gets the deposit back.
- If someone disputes, the case escalates, often to a vote by token holders.
This approach is popular in crypto because it’s cheap and fast for the 95% of markets with obvious answers. Some well-known platforms have used oracle designs like this. The risk shows up in the contested 5%: votes can be slow, and a small number of large holders can have outsized influence. Publicly reported disputes have shown how a contested outcome can turn into a debate about incentives rather than facts.
Staked reporter and bond-escalation systems
Older on-chain designs use token staking. Reporters put money behind the outcome they say is correct. If someone disagrees, they stake more, and the stakes escalate. In some systems, the final backstop is a fork, where the system splits and the market decides which version to follow. These designs are intellectually elegant but can be slow, costly and hard for newcomers to follow.
Other systems let anyone submit an answer and require each challenger to double the bond. If the dispute gets too expensive or complicated, an outside arbitrator, such as a decentralized court of jurors, makes the final call.
Data-feed oracles
For outcomes that are pure numbers, such as an asset price at a specific time, an automated oracle can report the result without human involvement. This is very efficient. But it only works when the outcome is a clean data point. It can’t decide whether a company “officially announced” something.
Newer ideas
Some newer designs experiment with AI-assisted resolution or hybrid human-and-machine review. These are still young. Treat them as unproven and read how they handle appeals.
The edge cases that cause most disputes
You can learn most of what you need about resolution by studying where it breaks. These are the patterns that come up again and again.
1. Ambiguous wording
“Will the game launch by December 31?” Does early access count? A closed beta? A launch in one region only? If the question doesn’t say, traders will argue.
2. Disagreeing or changing sources
Two reputable sources report different numbers. Or a page is edited after the fact. Or official data is later revised. A good rule names a single primary source and says what happens if it changes.
3. Timing and time zones
“By midnight” in which time zone? Does “before the deadline” include the last second? Is the cut-off based on when the event happened or when it was reported?
4. Fuzzy definitions
Words like announce, release, win, confirm and official sound clear but are not. Did the company “announce” something if a leaked trailer appeared first? Did a team “win” if the result was later overturned?
5. The event changes shape
An event is postponed, split in two, renamed or canceled. Markets differ here: some resolve to “No,” some void the market and refund, and some split the payout.
6. Ties and partial outcomes
Dead heats, shared victories, forfeits and disqualifications are classic problem cases in sports, esports and elections alike.
7. Subjective questions
“Will X be considered a success?” has no hard number behind it. Unless the rules define success, the outcome depends on who is judging.
8. Early or late-breaking information
A market resolves quickly, then new information appears. Or traders already know the outcome but the source hasn’t published yet, so the price sits at 95% for days.
9. Incentives and manipulation
People who hold positions may also be among those voting or proposing. In small systems, a single large participant can swing the result. Good designs try to make honesty the most profitable strategy, but no design removes this risk completely.
10. Rule changes after trading starts
Some platforms reserve the right to clarify or edit rules. A “clarification” can flip the outcome for traders who bought under the original wording.
What happens when a market can’t be resolved cleanly?
Platforms usually have a fallback, and it’s worth knowing which one before you trade:
- Void and refund: the market is canceled and traders get their money back (often at the original cost, not the current price).
- 50/50 payout: each side receives half the payout.
- Resolve to “No”: the default if the event didn’t clearly happen.
- Extend the deadline: wait for more information.
These fallbacks have different financial consequences. If you bought “Yes” at $0.90, a 50/50 payout costs you 40 cents per share, while a refund leaves you whole. The fallback is part of the real price.
How to read a market’s rules: a practical checklist
Before you buy, spend two minutes on the fine print. Look for:
- The exact question. What precisely needs to happen?
- The resolution source. Is a specific, named source given? Is there a backup?
- The timing. Start date, end date, time zone, and whether the event must occur before, by or on a date.
- Definitions. Are key terms (launch, win, announce) defined?
- The edge-case clause. What happens if the event is canceled, postponed, tied or ambiguous?
- Who decides. Operator, committee, oracle or vote? Is there an appeal?
- The dispute process. How long is the challenge window? How long can a dispute take? Is your money locked in the meantime?
- Rule changes. Can the platform edit the rules after trading begins?
- Past examples. Have similar markets on the same platform been resolved smoothly or contested?
A worked example
Weak wording:
“Will Studio X release Game Y by December 31?”
Possible arguments: Does early access count? Does a console release count if the PC release is delayed? Which time zone?
Better wording:
“Will Game Y be available for purchase on the official PC storefront page, in a full (non-early-access) release, by 23:59 UTC on December 31? Resolution source: the official storefront listing and the studio’s official announcement. If the release date is postponed beyond this date, the market resolves to No.”
The second version names the version of the game, the platform, the time zone, the sources and the fallback. It leaves much less room for argument.
Red flags to watch for
- Vague words such as “significant”, “successful” or “widely reported”
- No named resolution source
- No stated end time or time zone
- No mention of what happens if the event is canceled or changed
- Very low liquidity combined with a complicated question
- A history of unclear or disputed outcomes on the same platform
- Rules that can be changed at the operator’s discretion without notice
What resolution risk does to prices
Here is a trick that connects this topic to everything else. Suppose an event looks almost certain, yet the “Yes” share trades at $0.93 instead of $0.99. Part of that gap may be resolution risk: the market is pricing in the chance of a dispute, a strange interpretation, or a long delay before payout.
So when a price looks too low for something that seems obvious, don’t assume it’s free money. Ask what the market might be worried about. Read the rules, and see whether the answer is in the wording. This ties directly to the idea that a price is an estimate of both the outcome and how the outcome will be judged.
Practical habits that protect you
- Read first, trade second. Make rule-reading a habit, even for markets that look simple.
- Size positions to the clarity of the rules. Put less money into markets with fuzzy wording.
- Don’t treat “certain” as paid. Until the market is settled, your money is still at risk and may be locked.
- Be careful with last-minute trades. Late bets on “obvious” outcomes can be hit by technicalities.
- Keep records. Save the market rules and screenshots in case of disputes.
- Avoid one big bet on one oracle design. If you’re trading on-chain, understand who can influence the result.
FAQ
What does “resolution” mean in a prediction market?
It’s the process of deciding the final outcome and paying out. Once resolved, winning shares can be redeemed for the full payout and losing shares become worthless.
What is an oracle in a prediction market?
An oracle is a system that brings real-world information into a smart contract, so that the contract can settle automatically. It can be an automated data feed, a vote, a group of stakers, or a mix.
What is an optimistic oracle?
It’s a design that assumes a proposed answer is correct unless someone challenges it within a set time. If nobody objects, the answer is accepted. If someone does, the dispute escalates to a vote or another authority.
What happens if a prediction market is disputed?
It depends on the platform. The case may go to a committee, a token-holder vote, or an arbitrator. Disputes can take days, and in many systems your funds stay locked until the final decision.
Can a prediction market be canceled?
Yes. Some platforms void markets that can’t be resolved fairly and refund traders, while others use a 50/50 payout or default to “No.” Check the fallback rule before you trade.
Why do some obviously safe markets trade below $1?
There may be a time cost (your money is tied up), fees, or resolution risk, which means the chance of a dispute or unusual interpretation. A discount can be a sign to read the rules more carefully.
Who is most at risk from resolution problems?
Newer traders who skip the rules, and anyone trading in thin markets with vague wording. Experienced traders treat resolution rules as part of the price.
