What is a prediction-market resolution-quality audit?

A prediction-market resolution-quality audit is an independent review of the step where a market settles: whether it resolves to the correct outcome, on time, and strictly per its written rules, without ambiguity, disputes, or manipulation. It scores the trustworthiness of the resolution step for a market or a cohort of markets, which is a different question from whether the price direction was right. Convexly treats this as a market-quality problem and does not take custody, broker trades, or recommend copying any trader or wallet.

What resolution quality means

A prediction market is a claim plus a settlement rule. The price is a live estimate of the claim; the resolution is the moment the claim is graded true or false and traders are paid. Resolution quality is about that second step. A high-quality resolution is unambiguous (the rule maps cleanly onto what actually happened), timely (it settles close to when the real-world outcome is knowable), rule faithful (it follows the written criteria rather than an ad hoc reading), and clean (no oracle dispute, no retroactive rewording, no manipulation of the settlement source).

These properties can be assessed after the fact on resolved markets and estimated in advance from a market's rule text, liquidity, and settlement source. A resolution-quality assessment is descriptive: it tells you how much weight the eventual settlement deserves, not what the outcome will be.

Why it matters

A price is only informative if the resolution behind it can be trusted. If a market's rule is ambiguous, its oracle can be contested, it settles late, or its liquidity is too thin to move the price toward the true probability, then the number on the screen carries less signal than it appears to. Anyone citing a prediction-market price as evidence inherits the quality of the settlement it points to.

A market-quality lens turns that into a usable judgment: whether a given market is strong enough to cite, usable with caveats, too weak to rely on, or too thin to use. That is a distinct discipline from forecasting, and it is where an independent layer adds value, because a venue grading its own resolution quality has an obvious conflict of interest.

Where Convexly Market Trust fits, and its limits

Convexly Market Trust is an experimental canary_preview research surface for market-quality scoring. It is a preview of an approach, not a finished product. Specifically, it is not a calibrated public rating and not procurement-ready. It is published so the method is inspectable and so a resolved-outcome ledger can accumulate evidence, and it stays in canary_preview until that ledger calibrates it. Do not treat any preview value as a decision-ready score.

The reason for the caveat is honest sample size. Reliable calibration of resolution-quality scoring needs a large base of resolved markets with known outcomes, and that ledger matures on time and outcomes alone. Until it does, the preview shows the shape of the analysis, not a validated grade. This is the same discipline Convexly applies to wallet skill, which is reported as distinguishable from chance (or not) at a frozen bar rather than as a signal to copy any wallet.

The enterprise market-quality lane

Beyond the free research preview, Convexly runs a market-quality lane for enterprise engagements: an independent, reproducible review of a defined cohort of prediction markets, delivered with the methodology and receipts attached rather than as a black-box number. It is a sales-led engagement, scoped per customer.

The discipline is the same in both surfaces. Every figure in a deliverable travels with its 95% confidence interval and the resolved-market or resolved-position count it was computed on, so a reader can see the precision behind each number instead of a bare point estimate. Convexly does not take custody, broker trades, route investment advice, or recommend any position.

Frequently asked questions