What Makes a Prediction Market Resolvable?

​

Quick Answer

​

Key Insights

• A popular question is not automatically a good market.
• Resolvability is created before trading, not repaired after an event ends.
• The headline should attract attention; the detailed rule must function like a contract.
• Permissionless creation needs stronger rule validation because topic volume can scale faster than human review.
• AI can help standardize questions and inspect gaps, but difficult factual and policy choices still need transparent oversight.

​

Why Resolvability Matters

Every trader should be buying the same contract. If one person thinks “win” includes penalties while another thinks it means regulation time, they are not pricing the same event even if they trade in the same order book. Weak rules create three costs:

1. Pricing noise: traders price different interpretations.
2. Resolution delay: reviewers must reconstruct intent after the event.
3. Dispute risk: a losing side can reasonably argue that another interpretation was possible.

A well-written rule does not guarantee agreement, but it narrows disagreement to verifiable facts.

​

Seven Elements of a Resolvable Market

​

1. One objective event

The question should identify an observable event, publication or official decision.

• Weak: “Will the AI boom continue?”
• Stronger: “Will Company X report more than $Y in data-center revenue for Q4 2026?”

Words such as successful, major, significant or recession need definitions unless a named authority supplies one.

​

2. Complete and exclusive outcomes

The allowed outcomes should cover the possible result without overlapping. For a binary market, the rule must specify what produces Yes and what produces No. If cancellation, invalidation or another outcome is possible, the rule should say how it is handled.

​

3. A deadline and time zone

“By year-end” is incomplete when participants operate globally. A strong rule specifies the date, time and zone, and distinguishes event time from publication time.

​

4. A primary source or source hierarchy

The rule should name the evidence that controls:

1. primary official source;
2. fallback source if the primary source is unavailable;
3. treatment of conflicting or later-corrected reports.

For macro data, this may be a named government agency and a specific release table. For sports, it may be the governing body’s final result.

​

5. A revision policy

Many official numbers change. A market must state whether it uses:

• the first published value;
• the latest value available at a cutoff;
• a later finalized revision.

Without this rule, two accurate sources from different dates can produce different outcomes.

​

6. Edge-case rules

The market should address plausible exceptions, including:

• postponement or cancellation;
• ties, overtime or penalty shootouts;
• renamed entities or replacement candidates;
• partial completion;
• source outage;
• legal appeal or recount;
• an event that occurs after the deadline but is announced later.

The goal is not to anticipate every imaginable event. It is to cover cases that a reasonable participant would price differently.

​

7. A proposal, review and dispute path

The rule should tell users how evidence becomes a proposed result, when that result becomes final and how a factual error can be challenged. Read What Is a Prediction Market Oracle? for the main resolution models.

​

A Resolvability Scorecard

If a market fails one of the first five tests, it may need rewriting before publication.

​

Worked Example: Inflation Data

Weak version:

Will US inflation fall below 3% next month?

Questions left unanswered:

• CPI or PCE?
• headline or core?
• year-over-year or month-over-month?
• seasonally adjusted?
• which release date?
• first publication or revision?

Stronger structure:

Will the year-over-year change in the US Bureau of Labor Statistics all-items CPI for July 2026, as first published in the August 2026 CPI release, be below 3.0%?

The detailed rule should then identify the BLS table, treatment of an unavailable release and the exact comparison method.

​

Worked Example: Football

Weak version:

Will Team A beat Team B?

Stronger versions depend on the intended contract:

• regulation-time result only;
• team to advance, including extra time and penalties;
• official final result after disciplinary review;
• match played by a stated deadline, otherwise invalid.

These are different markets. A clear title and rule prevent traders from silently choosing different versions.

​

What Opinion Learned From Testnet

Opinion co-founder Jeff publicly described a testnet in which whitelisted users created hundreds of permissionless markets and community processes handled rules and outcomes. The experiment exposed inconsistent rule quality and markets that could not be reliably resolved. Opinion’s stated response is AI-assisted market creation: help users draft objective rules, define conditions, specify sources and evaluate whether a topic is suitable for publication. Its whitepaper similarly describes Opinion AI as checking whether a topic meets resolvability standards. This is a product direction, not permission to hide the rule behind AI. Users still need to see and evaluate the final contract.

​

What AI Can and Cannot Do

AI can help:

• detect missing dates, sources and definitions;
• suggest common edge cases;
• compare a proposed rule with known event formats;
• normalize wording across many markets;
• summarize unstructured evidence during resolution.

AI cannot by itself guarantee:

• that the chosen source is legitimate or politically neutral;
• that every future edge case was anticipated;
• that an ambiguous social concept has one objective definition;
• that its evidence was not incomplete or manipulated;
• that the final judgment is beyond dispute.

This is why Opinion has publicly described a hybrid approach combining AI efficiency with multiple agents, secure execution, human oversight and a dispute layer. See How Opinion AI Resolves Markets.

​

Sources Used

1. OPINION Whitepaper v1.0 — Opinion AI’s market-creation and resolvability role.
2. Jeff / Opinion: Resolution — Opinion AI — testnet lessons and publicly described hybrid direction.
3. OPINION Docs: Resolution — market-specific resolution requirement.
4. Kalshi Help Center: Market Rules — independent example of outcome and verification-source requirements.

​

FAQ