Comment by fdej

There's no a priori reason why the expected success rate of research projects should be 50% and not, say, 1% or 99%.