Three Equations, Three TDEEs: We Ran Mifflin, Harris–Benedict, and NASEM on Six Bodies
Type the same body into three different TDEE equations and you get three different answers — not because one is broken, but because each one is a different population average wearing a lab coat. We ran all three on six personas through the same engine that powers our calculators, so you can see the disagreement in kilocalories instead of adjectives.
· 9 min read
Every TDEE calculator picks one equation, multiplies by an activity factor, and prints the answer to the nearest calorie. This site runs three equations and shows you where they disagree — so for this post we took six personas, ran each one through Mifflin-St Jeor, the revised Harris–Benedict, and the NASEM 2023 Estimated Energy Requirement equations using the exact engine behind the TDEE calculator, and tabulated the damage. All figures below are build-time outputs of that engine, not hand-typed numbers.
The six personas, computed three ways
| Persona | Mifflin BMR | Harris–Benedict BMR | Mifflin TDEE (mid) | NASEM 2023 EER | Mifflin ↔ NASEM spread |
|---|---|---|---|---|---|
| Man, 30, 180 cm, 80 kg activity ×1.55 |
1,780 | 1,854 | 2,759 | 3,126 | 367 |
| Man, 45, 175 cm, 95 kg activity ×1.2 |
1,824 | 1,945 | 2,189 | 2,743 | 554 |
| Man, 60, 172 cm, 78 kg activity ×1.375 |
1,560 | 1,618 | 2,145 | 2,525 | 380 |
| Woman, 28, 165 cm, 60 kg activity ×1.375 |
1,330 | 1,392 | 1,829 | 2,197 | 368 |
| Woman, 40, 160 cm, 75 kg activity ×1.2 |
1,389 | 1,464 | 1,667 | 2,098 | 431 |
| Woman, 55, 163 cm, 68 kg activity ×1.55 |
1,263 | 1,343 | 1,958 | 2,230 | 272 |
Mifflin TDEE uses the site's standard multipliers (sedentary 1.2 / light 1.375 / moderate 1.55); each persona's activity tier maps to the closest NASEM PAL class per the official Table 5-4 bands. BMR figures are resting-only; TDEE/EER include activity.
What the table actually shows
Three patterns fall out. First, the revised Harris–Benedict runs hotter than Mifflin at the BMR level in 6 of 6 personas — by 58–121 kcal/day, smallest for the man, 60, 172 cm, 78 kg persona and largest for the heaviest one (man, 45, 175 cm, 95 kg). That is the Frankenfield 2005 finding reproduced on our own small sample: Harris–Benedict overestimates, and the bias grows with obesity (PMID 15883556). Second, the Mifflin↔NASEM TDEE spread averages 395 kcal/day across the six, ranging from 272 (woman, 55, 163 cm, 68 kg) to 554 kcal/day for the man, 45, 175 cm, 95 kg persona. Third, NASEM reads higher than the Mifflin-based TDEE in 6 of 6 personas here — not because NASEM is generous, but because its coefficients weight age, height, and weight differently than a BMR × multiplier stack, and at these activity tiers the PAL-class terms land above the classic multipliers.
Put bluntly: 395 kcal/day is not a rounding difference. It is the difference between "maintain on a small lunch" and "maintain with a second coffee and a snack" — decided entirely by which population average you borrowed.
Why the equations disagree: a one-paragraph history
Mifflin-St Jeor (1990, PMID 2305711) fitted resting expenditure to several hundred healthy adults spanning normal weight and obesity. The revised Harris–Benedict (Roza & Shizgal 1984) did not measure anyone new — it re-fitted the 1919 Harris & Benedict dataset with modern statistics, inheriting a cohort that predates most of modern nutrition science. The NASEM 2023 EER equations (DOI 10.17226/26818) come from the National Academies update behind current US and Canadian dietary reference intakes, and model total energy needs directly with explicit sex, age, body-size, and PAL-class terms instead of bolting a multiplier onto a resting prediction. Different cohorts, different decades, different statistics — different answers for the same body.
The honest way to use a number like this
None of the three is "the truth"; each is a population average you borrow for an individual, and the borrowing error is a few hundred kilocalories per day for every equation in the literature. The practical protocol, which the TDEE formulas compared guide and the calibration guide spell out in full:
- Start from Mifflin (best overall in Frankenfield 2005) and read NASEM as a second opinion — the TDEE calculator shows both side by side for exactly this reason.
- Treat the gap between them as your uncertainty band. A 554 kcal gap means any target inside that band is defensible and any target outside it needs evidence.
- Let your own data overrule both: four weeks of logged intake and daily weight settles the question every equation leaves open, because the residual error after calibration is mostly yours to close, not the equation's.
The one thing not to do
Do not pick the equation that prints the number you wanted. If you would "rather" maintain on 2,189 than 2,743 kcal, that preference is not evidence — it is the bias this table exists to expose. The spread between equations is real information about your uncertainty; spending it on wishful thinking is the most expensive calorie accounting there is.