To "count calories" and think that you are consuming EXACTLY the amount of calories you calculated is a rediculous idea anyway. i.e. "I ate 1562 calories today"...
1. How they determine how many calories are in the food you are eating.
The amount of food energy in a particular food could be measured by completely burning the dried food in a bomb calorimeter, a method known as direct calorimetry. However, the values given on food labels are not determined this way, because it overestimates the amount of energy that the human digestive system can extract, by also burning dietary fiber. Instead, standardized chemical tests and an analysis of the recipe are used to estimate the product's digestible constituents (protein, carbohydrate, fat, etc.). These results are then converted into an equivalent energy value based on a standardized table of energy densities
1a. The table of energy densiteis they refer to above is basically a unit of something (i.e. fat) is equal to X number of calories (kcal). These tables are also filled with estimates.. the number of calories per unit of material is not determined from ONE test, but from a bunch of tests, and the number used is the average of all the test. In truth, the tests have a statisitcal distribution to them, thus, the 9 kcal per gram (guessing) of fat is really 9 kcal +/- the standard deviation from the tests.
2. Portion sizes of items, especially if you are making a recipe, even more especially if you are using a volume measurement to determine the amount of a food you are eating.
Mass measurements allieviate this SOME, but you are still restricted to the accuracy of the scale being used.
Volume measurements.. how much are they talking about when the box/label says XX calories per cup? A cup of the product scooped into a cup, as it is sold, no messing with it after the scoop? A cup of the product scooped into a cup, then tapped/shaken a bit to improve the packing efficency inside the cup, then refilled back to the 1 cup level? Perhaps the product ground into a powder then filled to a cup?
If you are counting calories, you MIGHT be within 10%, and that is probably doing pretty good.
In my opinion, you should LOOSELY count calories for a short amount of time to get an idea of how much you consume, and more importantly learn the caloric density of the foods you tend to eat.