Why TDEE Calculators Are Estimates (and How to Dial Yours In)

By Jesse · Last updated August 20, 2026

A TDEE calculator gives you a precise-looking number — say, 2,437 calories. It is tempting to read that as your body's exact fuel requirement. It is a well-educated estimate with a margin of error you can put a size to, and once you know where the error comes from you can replace most of it with measurement.

This is the guide that owns accuracy and calibration. What to do with the number once you trust it — how big a deficit to run, what to do when the scale stalls — is covered in how many calories to eat to lose weight.

Where the number comes from

Almost every calculator, including our TDEE calculator, works in two steps. First it estimates Basal Metabolic Rate — the energy you burn at rest — from your sex, age, height, and weight, using the Mifflin-St Jeor equation. Then it multiplies that by an activity factor.

That second step deserves more suspicion than it gets. It is a single number covering everything you do all day, and because it multiplies rather than adds, its error scales with the whole estimate. Here is the same body — a 35-year-old man, 178 cm, 80 kg, with a BMR of 1,743 kcal — at every setting the calculator offers:

A single notch — the difference between calling yourself lightly and moderately active — is worth 305kcal a day for this body. That is larger than most people's intended deficit. And the descriptions are genuinely ambiguous: someone with a desk job who trains hard four times a week can talk themselves into either bucket.

The formula argument is a distraction

Online discussion tends to fixate on which BMR equation a calculator uses. We tested that directly, running Mifflin-St Jeor and Harris-Benedict across 1,778 adult body combinations. The median difference between them is +57 kcal per day, or 3.8% — and choosing the wrong activity level is the larger error in 98% of cases, by a factor of about 2.9.

So the practical advice is unglamorous: stop shopping for calculators, and spend the attention on the activity setting instead. The full analysis, including which bodies the formulas disagree about most, is in where the two standard BMR formulas disagree.

Why it can only ever be an estimate

A prediction equation is fitted to a population. It returns the average resting energy of people who share your sex, age, height, and weight — not yours. The review that compared the common equations found Mifflin-St Jeor the most likely of them to land within 10% of measured resting energy, which is a real endorsement and also an admission: a substantial minority of people fall outside that band, and the equation cannot tell you whether you are one of them. Four things it cannot see:

• Body composition. Two people identical on all four inputs can carry very different amounts of muscle, which is more metabolically active at rest than fat. The equation sees only total weight. This is also why it tends to under-read for very muscular people and over-read for people with little muscle.
• NEAT — everything that is not exercise. Standing, walking, fidgeting, and general restlessness vary by hundreds of calories a day between people of the same size, and are the single largest source of unexplained variation in daily expenditure.
• The activity multiplier is coarse. Five buckets to describe every possible life, applied multiplicatively. The table above shows what the granularity costs.
• Adaptation. Expenditure drifts as you gain or lose weight, partly because a smaller body costs less and partly because people move less when eating less. So even a correct estimate has a shelf life.

How to measure your own maintenance calories

You do not have to accept the estimate, because your own body is a better instrument than any equation. Energy balance runs in reverse perfectly well: if you know what you ate and what your weight did, you can solve for what you burned.

actual maintenance = average daily intake
                     + (weight change in kg x 7700 / days)

Weight loss makes the second term positive — you burned more than you ate — and weight gain makes it negative. Worked through: you averaged 2,000 kcal a day for 21 days and lost 1 kg. That kilogram was worth about 7,700 calories, or 367 per day, so your actual maintenance at that body weight was near 2,367 kcal. If the calculator had said 2,200, it was low by about 8% — unremarkable for a prediction equation, and now corrected.

The protocol, in order:

1. Eat at the estimate from the TDEE calculator and hold it steady. Changing the target mid-experiment wastes the experiment.
2. Log intake for two to three weeks — weighed, not eyeballed, at least at the start. Two weeks is the minimum for water shifts to average out; three is better.
3. Weigh yourself most mornings under the same conditions, and compare the average of the first week with the average of the last. Never compare single days.
4. Apply the formula above, then use that number in place of the estimate.

Two ways this misleads, both worth knowing. Logging error is usually larger than formula error and it runs one way: under-recording is the norm, not the exception, so a suspiciously low computed maintenance usually means missed calories rather than a slow metabolism. And a fortnight is short enough that a change in sodium, carbohydrate, or training volume can move scale weight by a kilogram for reasons that have nothing to do with fat — which is precisely the error the three-week window is there to dilute.

Re-estimate as you change

In Mifflin-St Jeor each kilogram of body weight is worth 10 kcal of resting energy, and the activity multiplier scales that up — so at a moderately active setting, maintenance falls by roughly 16 kcal per kilogram lost. Ten kilograms down and the target you set at the start is delivering a noticeably smaller deficit than you think. Re-estimate every 5 to 10 kg (10 to 20 lb) of change, and re-run the calibration above whenever progress stalls for more than two or three weeks. To turn a trustworthy maintenance figure into a fat-loss target, use the calorie deficit calculator.

Common questions

These are general estimates for healthy adults, not medical advice. If you are pregnant, under 18, have a medical condition, or a history of disordered eating, talk to a doctor or registered dietitian before changing your intake. See our methodology for the sources behind these figures.