Noise, even at the level of 10%, in actual calorie content of food compared to nutritional labels, is irrelevant (there are other, more significant sources of v

Noise, even at the level of 10%, in actual calorie content of food compared to nutritional labels, is irrelevant (there are other, more significant sources of variance in your own metabolism and activity in any case). If, over the course of a week, you haven't lost any weight, decrease your target calorie count by 50. Eventually the odds of predicted vs. actual noise giving you an "over maintenance" intake for a given week fall to epsilon. Counting calories is not even necessary.