It would be interesting to know which diets were being tested with exercise - it might not be the exercise, but rather the diet that was to blame.

I suspect the Atkins diet was not among those tested - it was out of favor in the scientific community for so long, I'd bet those diets were low-fat nightmares.

In such a diet, the exercise might reduce glucose levels, (meaning less of it would be stored as fat), but if ketosis never or rarely occurred, body fat would not decline - no weight loss.

But in ketosis, exercise means more FAT burning, and more muscle mass would lead to additional resting fat burning.