The original research supporting "carbs are good, fats are bad" made such trivial statistical errors it's kind of crazy.