x
Optimizing Repeated Correlations — LessWrong