How we compare venues and match markets

Comparing prices across venues only means something if the two contracts are genuinely the same question, and comparing venues only means something if every claim can be checked. This page is how we decide both, and, just as important, the list of things we will not publish until we can.

The criteria, and why there are no ratings yet

Venues are compared on verifiable attributes: regulator and registration status, funding rails, whether machine-readable prices are published, fee model where the venue publishes one, and state-level legal posture with a citation per claim. They are not scored. A score asserts a weighting, and a weighting that has not been published is not a method; it is an opinion wearing a number. The venues page lists venues without a score for exactly that reason, and will until the criteria and weights are written down here.

Matching two contracts

Candidate contracts are shortlisted by semantic similarity against the full catalog of contract families on each venue, then compared at the individual market level. Similarity alone is never enough: questions in the same family read almost identically and differ only in the parts that decide the answer. So every pairing must pass four checks as well as the similarity threshold, the year named in the question, the direction (above versus below), the strike, and the scope (winning a party's nomination is not winning the election). A candidate that fails any one is discarded, however similar it reads.

Settlement wording can still differ between two venues pricing the same question, so every comparison page shows each venue's resolution rules side by side. Read them before trading on a difference in price.

What we do not score, or publish

A consensus price from a single venue. One quote is not a comparison, and those pages say so.
Market share across venues measured differently. Where one venue's coverage is complete and another's partial, no share figure is given.
Worked fee amounts where the venue publishes no rate to work them from.
Venue ratings, until the weights are published here.
State legality without a cited ruling or regulator action.
Promotional codes that are not live.

Cadence and corrections

Cross-venue prices refresh within a minute. Venue facts, legality data and fee schedules change when a venue, court or regulator changes them, and every legality page shows the date it was last updated. Every page shows when its content, not its prices, last changed. Corrections are made in place and the change date moves; a claim is never rewritten without moving its date.

Who does it

Cross-venue pairings are matched by the method above as markets change. Legality pages link a public ruling, docket, regulator action or report for each claim and carry the date it was last updated; fee figures come from the venue's own API or published schedule. The whale and smart-money signals are computed from on-chain data by CoinStats and describe one venue only. The people who write and review these pages: Jeremy Bragin (author), Vardan Khachatryan (legal author). Our position on risk