The three equations
All three return basal metabolic rate: what you burn at complete rest, before any activity. Maintenance calories are BMR multiplied by an activity factor.
Mifflin-St Jeor (1990)
Men: 10w + 6.25h − 5a + 5
Women: 10w + 6.25h − 5a − 161
w = kg, h = cm, a = years
Derived on 498 healthy adults and explicitly intended to replace Harris-Benedict, which was drawn from a much earlier and leaner population. It has the best validation record in the general population and is what this site leads with.
Harris-Benedict, Roza-Shizgal revision (1984)
Men: 88.362 + 13.397w + 4.799h − 5.677a
Women: 447.593 + 9.247w + 3.098h − 4.330a
The original is from 1919; this is the revision that corrected its known over-estimate. It still tends to read a little high in modern populations, which is exactly what you would expect from an equation fitted a century ago.
Katch-McArdle (1996)
370 + 21.6 × lean body mass in kg
No sex term, no height term, no age term — because lean mass already carries that information. It is the only one of the three that knows anything about body composition, and the only one that needs a measurement most people do not have.
The same person, three answers
A 30-year-old man, 80 kg, 180 cm, 18% body fat, training three to five times a week (activity factor 1.55):
| Formula | BMR | Maintenance | Needs |
|---|---|---|---|
| Mifflin-St Jeor | 1,780 | 2,759 | age, height, weight, sex |
| Harris-Benedict (revised) | 1,854 | 2,874 | age, height, weight, sex |
| Katch-McArdle | 1,787 | 2,770 | lean body mass |
A spread of 115 kcal, about 4%. That is the honest disagreement between three equations that all have a legitimate claim to being right, and it is worth roughly a small snack a day.
The multiplier that matters more
Now change nothing except the activity factor:
| Activity | Factor | Maintenance (Mifflin) |
|---|---|---|
| Sedentary | 1.2 | 2,136 |
| Light — 1 to 3 sessions | 1.375 | 2,448 |
| Moderate — 3 to 5 sessions | 1.55 | 2,759 |
| Very active — 6 to 7 sessions | 1.725 | 3,071 |
| Extremely active | 1.9 | 3,382 |
One step on that scale is worth about 310 kcal — nearly three times the entire disagreement between the three equations. Choosing the wrong activity level is a far bigger error than choosing the wrong formula, and the overwhelming majority of people choose one level too high. Four gym sessions a week is “moderate”, not “very active”, unless the rest of your day is also physical.
Which to use
- No body-fat measurement: Mifflin-St Jeor. It is the default here for a reason.
- A real body-fat measurement: Katch-McArdle. It is the only one that distinguishes 80 kg at 12% from 80 kg at 25%, and those two people genuinely differ.
- Very lean or very muscular: Katch-McArdle again. The other two will read low on you.
- Harris-Benedict: mainly when you need to match a figure someone else produced with it.
Then treat whichever number you get as a two-week hypothesis. Eat at it, weigh yourself consistently, and adjust from what actually happens. Your own trend data beats every equation on this page within a fortnight.