Right. If you're going to accept a 50% failure risk, your average return from the companies that don't fail has to be better than 2x. If 70% of them fail, even 3x isn't enough.