Can a Strategy Be Good Even If It Has Long Periods of Underperformance?

Every approach may experience unfavourable periods, but persistent losses can also indicate that its edge has weakened. How would you distinguish a normal drawdown from a reason to stop using the strategy?