I meant that the gap between % of placebo deaths and % of treatment deaths doesn't need to be as large as you said to get statistical significance.

20% in placebo, and 10% in treatment gets us there. Obviously we hope the difference is larger.