What is your approach to evaluating a strategy that performs well on one instrument but poorly on another?