“evidence markets reveal not just what the crowd believes, but why”
Hossain, Andrade, Zang, and Chen introduce evidence markets, a new mechanism that generalizes prediction markets to handle two limitations: prices reveal what the crowd believes but not why, and markets require external resolution events with known dates. The system uses a Logarithmic Market Scoring Rule with dynamic liquidity that responds to evidence quality. The authors prove bounded platform loss, show evidence rewards scale with uncertainty, and demonstrate epsilon-dominant strategy incentive compatibility for truthful reporting. An LLM-as-a-Judge verification layer with staking handles operational deployment.
Extensive technical background assumed