[beagle3]

No, you aren't parsing this correctly.

Cabanac has reproducible research that shows people losing weight on (IIRC) 2500 calories per day and NO physical activity whatsoever, because they were bed-ridden.

MGU had a research once that had people gaining weight on 1500 calories/day and losing weight on 3000 calories/day with comparable physical activity (this one included running, etc.)

The "calories in / calories out" argument is very far from science. If you trace its origins, you'll see it's not much more than a folk tale. It's about as accurate as "eat less to lose weight" in terms of being predictive.

Gary Taubes "Good Calories, Bad Calories" has all the explanations and references you could need, if that interests you.