a high win rate overrated?

A high win rate looks impressive, but it can hide large average losses. Which matters more to you when evaluating a strategy: win rate, expectancy, or drawdown?