An Approximation of a Group's CEV — LessWrong