Our Methodology
Every report, chart, and dataset is produced under a published methodology — sources, refresh frequency, aggregation rules, and known limitations.
Updated 24 October 2026.
How to read our research
Prediction Frontier studies prediction market activity, market structure and regulation. This page sets out our general approach. The sources, definitions and notes accompanying each report or chart explain how its particular figures were produced.
Sources and scope
Our sector research has drawn on Dune, DeFi Rate and exchange APIs. The source mix varies by project. Analysis should identify the venues, products, dates and categories covered, as well as any exclusions or missing data. A figure from a third-party dashboard should be attributed to that provider, with its underlying definition made clear where available.
For regulatory reporting, our approach is to distinguish legislation, court decisions and regulator statements from proposals, company announcements and commentary. Primary documents should be linked where available.
Measures and comparisons
Notional contract volume measures trading using the contract-value convention of the cited source. It is not interchangeable with the amount paid to acquire those contracts. Low-priced contracts, including some combos and parlays, can create a large gap between these measures.
Dollars paid measures trading value based on transaction prices under the source’s counting method. It is not a measure of unique deposits, net capital inflows or money remaining in a market.
Open interest describes positions outstanding at a point in time. Its valuation method, timestamp and venue coverage should be stated. Trading volume measures activity over a period; open interest is a snapshot.
Liquidity analysis should identify the measure used, such as bid–ask spread or visible depth at a specified price. Historical trading volume alone does not establish the price or size at which a new order could execute.
Cross-venue comparisons should explain differences in counting conventions, market categories, product coverage and observation periods. Similar event titles do not establish identical settlement rules. Wallet counts should not be presented as counts of unique people.
Dates, updates and limitations
Figures should carry an observation date or period. Report figures are historical snapshots; dashboard data may update on a different schedule and can lag its source. Readers should use the displayed timestamp and source notes when assessing freshness.
Available data can omit venues, products or fields. Source classifications and historical figures may also be revised. Estimates should be identified, and conclusions kept within the limits of the observed data. Market prices express implied probabilities at a particular time and can be affected by liquidity and market structure.
Sources readers can examine
Reports and charts should link to their sources, queries or supporting materials where access and permissions allow. Public availability of a data source does not mean that Prediction Frontier has independently audited it. See our Q3 2026 Sector Report for an example of our sector analysis.
Corrections and questions
Our policy is to explain material corrections in a dated note on the affected report or article. Send questions, supporting evidence or requests for source materials to info@predictionfrontier.com. See our editorial and research standards for our approach to disclosures and editorial responsibility.