Combinatorial Information Market Design

Information markets are markets created to aggregate information. Such markets usually estimate a probability distribution over the values of certain variables, via bets on those values. Combinatorial information markets would aggregate information on the entire joint probability distribution over many variables, by allowing bets on all variable value combinations. To achieve this, we want to overcome the thin market and irrational participation problems that plague standard information markets. Scoring rules avoid these problems, but instead suffer from opinion pooling problems in the thick market case. Market scoring rules avoid all these problems, by becoming automated market makers in the thick market case and simple scoring rules in the thin market case. Logarithmic versions have cost and modularity advantages. After introducing market scoring rules, we consider several design issues, including how to represent variables to support both conditional and unconditional estimates, how to avoid becoming a money pump via errors in calculating probabilities, and how to ensure that users can cover their bets, without needlessly preventing them from using previous bets as collateral for future bets.