Scott analyzes play money and reputation systems in prediction markets, discussing their designs, challenges, and potential improvements.
Longer summary
Scott Alexander discusses the design and effectiveness of play money and reputation systems in prediction markets, focusing on Metaculus and Manifold. He explores the trade-offs between absolute and relative accuracy, zero-sum and positive-sum scoring, and the challenges these systems face. Scott points out that reputation systems often fail to provide actual reputation benefits, and play money systems struggle with market inefficiencies due to limited incentives. He suggests potential improvements, such as offering interest-free loans for specific markets, to address these issues.
Shorter summary