Thomas Sepulchre comments on K-complexity is silly; use cross-entropy instead