Following up on my "failed predictions" point, I point at the now widely-distributed article about the low-fat vs. low-carb diets.
The standard dietary orthodoxy predicts that the results of this study would be exactly the opposite of what happened. The predictions are wrong. Therefore, the standard model is wrong. I don't need to be a nutritionist to make this determination.
I don't know what the right model is. I think that's pretty sad, considering that if nutritionists hadn't gotten bogged down by a premature orthodoxy, we would probably be now in the position of refining a solid idea, instead of where we really are now, which is quite nearly square one.
Again, the value of post-facto corrections made to predict previous results is zero or less; if a nutritional theory can't predict it's just plain no good. Previous studies pointed in this direction years ago, but they were shorter-term studies, so the dietary orthodoxy was updated to say oh, sure, maybe higher fat diets work at first but in the long term they will fail. This study is important because it's the first big long-term study. Science would have been better off to take those first studies, which still completely violated the predictions of the models, more seriously.
(I recognize that when your models are off a little bit, it is better to try to fix them up than to just junk them. When your models are 100%, 180-degrees-opposed wrong on a grand scale, you're usually better off junking them. No matter how emotionally attached to them you are.)