>The example you point to is dubious to me. The study seems to be working as intended: if you have a drug that doesn't work better than olive oil - you don't have a useful drug.

Why not?