> I'd bet a fair amount that your Mass General study used self reported calorie intake.
99% of nutrition studies are, but this one wasn't. I can't find the reference now, though.
Personally, I've been on a ~1200kcal/day diet for years (for some of that time, I did very detailed tracking), and stayed at my 220lbs. Which, of course, makes no sense, and it didn't to my girlfriend at the time who was an MD - so she decided she'd show me how wrong I am by eating the same as me. She lost weight quickly and started blacking out (apparently some form of malnutrition) within a few days, and stopped after a week with a SEP field resolution ("Contradicts everything I know, so I'm just going to ignore it").
Then, 12 month ago, on essentially the same diet, I started gaining weight - slowly but surely. And then I decided to cut away wheat, and lost 30 pounds within a month. (In retrospect, I also noticed that my gain weight coincided with going from pure-egg protein powder to egg-and-wheat protein powder).
There are about a thousand more variables than calories, and the body can change its efficiency.
Ignore the rat studies if you like, but cabanac has similar experiments with humans -- basically, people fed through a nose tube lose weight almost independently of the amount of calories you put in their stomach. Body just doesn't use the food unless proper signaling (apparently, scent related) happens.