Even in well controlled clinical research, counting calories is extremely difficult.
This is why good research uses tools like diet records and food frequency questionnaires as ways to identify changes in eating patterns, and not to measure calorie intake.
The truth is, it is very difficult to accurately measure how many calories you eat in any day. Anybody who tells you they eat EXACTLY 3,125 calories in a day is really telling you they eat somewhere in the neighborhood of 2,000 to 4,000 calories per day (average margin of error in research trials is 30%, and the more calories a person eats, the bigger the error becomes!).
I'm not against calorie counting tools and software, but I do think you should use them to identify trends, and not to measure the exact number of calories you eat.
To learn more, watch the video below:
Calories are definitely important, but obsessive over exact numbers is a recipe for diet failure.