That's true. If you can detect fake data patterns, you can generate fake data that doesn't follow those patterns. That wouldn't detect the researchers who were

That's true. If you can detect fake data patterns, you can generate fake data that doesn't follow those patterns. That wouldn't detect the researchers who were lazy and published fake studies because they are expected to. It would, however, foil the researchers who are creating supporting data to arrive at a particular conclusion