How much energy you use in a day — and how wrong the estimate is
Step 2 of 6. "How many calories do I need?" has an answer, but the answer is a range with a model behind it, and the model is worth ten minutes because it explains why two calculators disagree and why neither is lying.
The model, in four layers
Resting energy. Even asleep, your body spends energy keeping you alive — most of what you use in a day. Nobody measures this for you; it is estimated from weight, height, age and sex with a population equation. The one used on this site and in the app is the Mifflin–St Jeor equation, fitted in 1990 to a few hundred adults. In a systematic comparison of such equations it came out the most reliable, and "most reliable" still means that for a meaningful share of individuals the estimate is off by ten percent or more in either direction.
The ordinary day. On top of resting energy comes everything you do without calling it exercise: sleeping, sitting at work, standing, commuting, walking around. The method the app uses — the same one the FAO/WHO/UNU expert report describes — expresses each part of the day as a multiple of resting energy, hour by hour, and adds them up. A day at a desk comes out around 1.4 times resting energy; a day of heavy physical work, more than 2. These multiples are population averages for categories of activity. They do not know how you specifically sit, stand or walk.
Deliberate sessions. A run, a squash match, a gym session is added separately, using a standardised value for that activity from a compendium that scores hundreds of activities. Those values are standardised for comparison between studies; a fair number of them are estimated rather than measured, and the adult edition covers ages 19 to 59.
The total for that day. Add the three and you have your expenditure for one specific day — which is why the number should move when the day does. A day with a long walk and a day on the sofa are different days, and a good tool treats them differently.
How far off it can be
Each layer carries its own error, and they do not cancel:
- The resting-energy equation is a population fit. For an individual it is a starting guess.
- The activity multiples are category averages. Your "desk job" may involve more or less movement than the category assumes.
- The session values are standardised, not measured on you.
Put together, a daily figure is honestly a range, not a point. A tool that shows you one number to the calorie is showing you false precision. The app shows the number it plans with and says what it is.
What to do with the error
This is the useful part. Because the estimate is only a starting point, the method is:
- Start from the estimate.
- Eat to it for two or three weeks, logging what you actually eat.
- Watch the trend of the scale — not any single morning.
- If the trend does not match the plan, the estimate was off for you; adjust the target, not your self-respect.
The scale over weeks is the only measurement of your expenditure you will ever get without a metabolic ward. The estimate exists to make the first weeks reasonable, not to be believed.
What this page does not claim
It does not claim that the number from any calculator is "your metabolism". It does not claim that a wearable, a smarter formula or a DNA test removes the error — they trade one population model for another. And it says nothing about body-fat percentage: the app estimates that separately where it needs it, and shows the error of that estimate too.
The app runs the same formulas as this page, but from the actual day — the sleep you logged, the sessions you planned or did — instead of a typical one, and re-issues the target when the day changes. The public calculator gives the same number for the same input; that is a test in the build, not a promise.