We curate structured data on real-world events — elections, conflicts, policy shifts — with verified outcomes and manipulation-resistant probabilities, built for researchers, newsrooms, and ML pipelines.
Stanford GSB and the Hoover Institution.
People learn what's happening in the world from journalists. Models learn from training data. Both need the same thing: verified facts about real events. We deliver that to each in the format they consume.
When journalists cite prediction markets, they need to know whether the price they're reporting is reliable or easily manipulated. We grade every outlet and every citation so readers know what to trust.
Media →AI models fail at predicting real-world events because they lack ground truth to learn from. We produce structured records — the question, the probability, and what actually happened — so they can get better.
Models →Every metric in the API is backed by research from the Hall Lab at Stanford GSB and the Hoover Institution.
Volume-weighted averaging and manipulation resistance scoring for prediction market prices.
Read →A methodological note on truncation timing — recent applied work increasingly relies on resolution-relative prices, often without explicit discussion of the implications.
Read →Examining the rapid growth of political prediction markets, their governance challenges, and what they mean for how we understand political information.
Read →The Tape API is live. If you're a newsroom, researcher, or builder working with prediction market data, we'd like to hear from you.