People vary widely. Without finding a large enough group, it seems a comparison of before and after of the same group of people is actually more accurate than the classic "experimental" group and the "control" group.
With a control group, would you have matched pairs, with the same excess pounds, blood numbers, and blood pressure? And what are the chances of finding them?
Or you'd have to go for the difficult task of having enough people for your study so that these differences would wash out, which would be a rather high number, once you include people who are losing ten pounds, losing a hundred pounds, don't need to lose weight but have bad blood numbers...etc.

Do you see how control groups are actually not as effective as you would think?

Where different experimental groups could come in handy is comparing same calories/different diet composition. There are such studies out there and they disprove your "calorie is a calorie" hypothesis. The same amount of calories (in the same people) create different amounts of weight loss depending on the ratios of carbohydrate/fat/protein. I would check out these studies before I go out on a limb with it as a criticism. Such studies show that the "calorie is a calorie" hypothesis is not sufficiently supported.