How do nutritionists calculate calorie requirements for weight loss?

I hear very often that 3500 calories is equal to one pound of fat. I can provide a cite if anyone needs this. Using this information, weight loss sites say that to lose one pound of fat per week, you need to burn 500 more calories than you eat. It’s generally reduced down to a simple mathematic equation, that I hear pretty much everywhere. They also make it seem like calories, regardless of source, are equal in terms of weight loss.

However, as I understand it, 3500 calories is literally the amount of energy that one would get if one burned a source of pure fat, like oil, in a bomb calorimeter, although I may be wrong about this. Is that really equivalent to calculating how a human would gain one pound of adipose tissue? What about the idea that protein and carbs are each 4 calories and fat is 9 calories/gram? Isn’t this also from a bomb calorimeter? But wouldn’t the human body process amino acids vs. glucose vs. fatty acids differently? For example, I think that amino acids are metabolized in the Krebs Cycle and not in glycolysis, and so one molecule of amino acid would produce two less ATP than a molecule of glucose, and so effectively have less net calories. Another example . . . while glucose and sucrose might create the same amount of energy in a calorimeter, the body must break down sucrose into its component monosacchrides before metabolizing it, which I would think requires some ATP.

Could anyone shed any light on the validity of the frequently-tossed-around numbers? Or if I’m wrong in my analysis, show me where?

I can tell you this: ‘3500 calories = 1 pound of fat’ is not based on any science.

And yes, macronutrients are processed by the body very, very differently. PM me if you want book recommendations, I’m sure everyone here is sick of my harping.