Why Every TDEE Calculator Gives You a Different Number

Put your details into three different TDEE calculators and you will get three different numbers, often several hundred calories apart. It is reasonable to conclude that one of them is broken. None of them is. The disagreement is structural, and understanding where it comes from tells you how much weight to put on any of them.

Total daily energy expenditure is built in two steps: estimate resting metabolic rate, then multiply it for activity. Both steps introduce error, and the second is much worse than the first.

Step one: the BMR equation

Basal metabolic rate is what you burn at complete rest. Several equations estimate it from height, weight, age and sex.

Harris-Benedict, published in 1919 and revised in 1984, is the oldest in common use. It was derived from a population that differs from most modern ones in body composition, and it tends to run high.

Mifflin-St Jeor, from 1990, was developed specifically because the older equations were drifting on contemporary populations. It is generally regarded as the better default for most people, and it is what a well-built calculator will use unless you tell it otherwise.

Katch-McArdle takes a different input: lean body mass rather than total weight. Since muscle is metabolically active and fat is much less so, this sidesteps the main weakness of the others — that two people of identical height, weight, age and sex can have very different body composition and therefore different resting rates. The catch is that it needs a body fat percentage, and a bad body fat estimate produces a worse answer than not using the equation at all.

Between equations, the spread on the same person is commonly 100 to 200 calories. That is real but manageable.

Step two: the activity multiplier, which is where it falls apart

The BMR figure is then multiplied by an activity factor, usually from a list like this:

Look at the gaps. On a BMR of 1,700, the step from "lightly active" to "moderately active" is 1.375 to 1.55 — a difference of about 300 calories, decided by which word you thought described you best.

There are five brackets covering the entire range of human activity, the descriptions are vague, and people are poor at placing themselves. Someone who trains four times a week but sits at a desk the rest of the time genuinely does not know whether they are lightly or moderately active. The honest answer is that they are somewhere between, and the list does not offer between.

This one choice introduces more error than every other part of the calculation combined.

What the multiplier is hiding

Activity is not only exercise. Two components matter and neither is on the list.

Exercise expenditure is the obvious one, and the easiest to estimate — though watch-reported calorie burn is itself a modelled figure with substantial error, not a measurement.

NEAT — non-exercise activity thermogenesis — covers everything else: walking, standing, fidgeting, gesturing, taking stairs, general restlessness. It varies enormously between individuals, and research on overfeeding has found differences of several hundred calories a day between people of similar size. A naturally fidgety person and a naturally still person with identical training loads have meaningfully different TDEEs, and no calculator can see this.

The thermic effect of food adds roughly 10% of intake, since digestion itself costs energy. Most calculators fold this into the multiplier rather than showing it.

Individual variation on top of all that

Even with a good equation and an honest multiplier, measured resting metabolic rate varies around the predicted value by roughly 10% in either direction for reasons that have nothing to do with the inputs — thyroid function, genetics, and factors not fully characterised.

Stack the equation choice, the multiplier bracket, NEAT, and individual variation, and a TDEE figure is not a measurement. It is a starting estimate with a plausible range of several hundred calories.

So what is it good for

A starting point, and that is not nothing. You have to begin somewhere, and beginning from a reasoned estimate beats guessing.

The method that actually works is to treat the number as a hypothesis and test it:

1. Take the estimate and eat at it for two to three weeks, tracking intake as honestly as you can manage.
2. Weigh yourself consistently — same time, same conditions — and use a weekly average rather than individual readings. Day-to-day weight is dominated by water, glycogen and gut contents, and can swing a couple of kilograms for reasons unrelated to energy balance.
3. If the weekly average is stable, the estimate was close and you now have a real maintenance figure measured on you.
4. If it is moving, adjust by 100 to 200 calories and repeat.

After a month you have something no calculator can produce: your actual maintenance intake, derived from your own response.

Why the number moves

One more thing worth knowing, because it explains a lot of stalled progress. TDEE is not fixed. It falls as you lose weight, for two reasons.

The obvious one is that a smaller body costs less to run. The less obvious one is adaptive thermogenesis: during sustained energy restriction, expenditure tends to fall by somewhat more than the size change alone accounts for, partly through reduced NEAT — you move less, largely without noticing.

The practical consequence is that a deficit calculated at the start of a diet is a smaller deficit by the middle of it, which is a common reason weight loss stalls while adherence has not changed. Recalculating periodically, or better, re-deriving maintenance from tracked data, handles this.

A note on precision

TDEE tools output a number like 2,347. That precision is an artefact of arithmetic, not knowledge — the underlying uncertainty is in the hundreds.

Read it as "roughly 2,300, probably somewhere between 2,100 and 2,600", set your intake from it, and then let a few weeks of your own data tell you where in that range you actually sit. That last step is the one that turns an estimate into information, and it is the step most people skip.