The laws of thermodynamics have not changed, but try examining a really-really-really-really-really high dimensional phase without getting tendrils/'tubes' that appear essentially random w.r.t. prior conditions. How would you know the direction of a volume of phase space with fairly significant perturbation across all the variables that can change?
A flippant counter-example: eat only carbohydrates with a 1000 calorie deficit, drink 5 liters of water daily- 0 additional calories. How would you not gain water weight? Doesn't that count also?
The laws of thermodynamics tell you your caloric equations have to be balanced, but the internal state of the body can alter behaviors such as feeding and moving, so appealing to the laws of thermodynamics is pretty much off-topic. It's a common fallacy, and you're not alone in committing it. Obesity is increasingly seen as a metabolic disease with behavioral consequences rather than the other way around.
Here are two arbitrary papers I've blindly pulled out of the google-scholar hat, both of which discuss aspects of the issue at hand. Much has been written on the subject, and I respectfully urge you to investigate further.
A flippant counter-example: eat only carbohydrates with a 1000 calorie deficit, drink 5 liters of water daily- 0 additional calories. How would you not gain water weight? Doesn't that count also?