> Approaches that are wrong from a statistics point of view are more likely to generate impressive seeming results.

This is to be expected from an information theory point of view. It's why "fake news" will always be a thing.