The Mifflin-St Jeor equation: the formula, and how accurate it is
The formula behind most calorie calculators: what it estimates, where it is less accurate than its reputation, and how to test a target against your trend.
For men, the Mifflin-St Jeor equation is 10 x weight (kg) + 6.25 x height (cm) - 5 x age (years) + 5. For women the last term is -161. It estimates resting energy expenditure: the calories your body uses at rest, over a day, doing nothing.
That is a narrower number than most men assume. It is not what you burn in a day, and it is not what you should eat. Everything between the equation's output and a daily calorie target is added afterward by conventions the 1990 study never tested.
The formula, original and simplified
The published regression used unrounded coefficients and one equation for both sexes: REE = 9.99 x weight + 6.25 x height - 4.92 x age + 166 x sex (men 1, women 0) - 161. The authors then reported the rounded, sex-separated version above and stated that the simplification did not affect its predictive value (Am J Clin Nutr, 1990). Calculators use the rounded form, which is why you may see two slightly different results for the same inputs.
Take a 35-year-old man at 85 kg and 180 cm. That gives 850 from weight, 1,125 from height, minus 175 for age, plus 5. His resting energy is about 1,805 calories a day.
What the coefficients are, and are not
They are the weights a multiple regression assigned when fitting resting energy to weight, height and age across 498 people, and together those three variables explained 71% of the variation between individuals. They are not a ranking of physiological importance, and cannot be: a kilogram and a centimeter are not comparable quantities.
The age term is the one most often misread. It lowers the estimate by 5 calories per additional year, holding weight and height constant, which describes how the sample differed across ages. It does not follow that one man's resting energy drops by exactly 5 calories a year as he ages.
Which measurement this is: BMR, RMR or REE
Basal metabolic rate is the strictest of the three: overnight fast, full rest, measured before you get up and move. Resting metabolic rate and resting energy expenditure allow looser conditions and come out a few percent higher. Mifflin-St Jeor was built against measured resting energy expenditure, so a calculator printing "your BMR" is using the word loosely.
Turning the estimate into a daily target
Resting energy is not a calorie goal, so consumer calculators multiply it by conventional activity factors:
| Activity | Multiplier | Our 1,805-calorie example |
|---|---|---|
| Desk job, little exercise | 1.2 | 2,166 |
| Light, 1 to 3 sessions a week | 1.375 | 2,482 |
| Moderate, 3 to 5 sessions | 1.55 | 2,798 |
| Hard, 6 to 7 sessions | 1.725 | 3,114 |
The gap between the first row and the last is nearly a thousand calories, and nothing in the Mifflin-St Jeor study validates these labels or decides which row you belong to. They are a convention that spread through fitness software.
Research uses a defined version of the same idea: the physical activity level, total daily energy expenditure divided by resting metabolic rate. Even that is usually estimated rather than measured, and a 2026 analysis argues the usual method inflates it, because the conventional 1 MET value overstates true resting metabolic rate by 10% to 30% (BMJ Open Sport Exerc Med, 2026). Our calorie calculator uses the consumer multipliers to produce a starting estimate rather than a measured daily total.
How accurate is the resting number
Reasonable, and dependent on who you are.
A 2005 systematic review compared the four equations then used most in clinical practice and found Mifflin-St Jeor predicted resting metabolic rate within 10% of the measured value in more people, obese and not, than Harris-Benedict, Owen or the WHO equation, with the narrowest error range (J Am Diet Assoc, 2005). That review is the source of the equation's reputation.
Later work complicates it. A 2023 appraisal pooled 353 individual measurements from 14 studies and found Harris-Benedict came closest in white adults, within 10% for more than 70% of them. Its authors also flag a problem underneath the whole literature: a 12 to 14 hour overnight fast may not reach the post-absorptive state these equations were validated against (Eur J Clin Nutr, 2023).
"Within 10%" is an accuracy criterion across participants, not a personal confidence interval. For the example above, that band spans roughly 1,625 to 1,985 calories.
Mifflin-St Jeor against Harris-Benedict
Two studies, two answers, and the difference is the population. The 1990 paper found the 1919 Harris-Benedict equations overestimated measured resting energy by about 5% in its sample; the 2023 appraisal found Harris-Benedict closer in white adults. The honest summary is that Mifflin-St Jeor is a sensible default rather than the most accurate equation everywhere.
Is it accurate for athletes
Less so. A 2023 meta-analysis covering 29 studies, 1,430 athletes and 100 different equations found five that did not differ statistically from measured values, and Mifflin-St Jeor was not among them: its predictions differed statistically from the measured values. For precision, the Ten-Haaf equation landed within 10% for 80.2% of participants, against 40.7% to 63.7% for the rest (Sports Med, 2023).
Age matters as well. Among adults over 65 with a body mass index of 30 or more, only 43.1% of Mifflin-St Jeor estimates fell within 10% of indirect calorimetry, and every equation tested managed under 60% (J Nutr Gerontol Geriatr, 2022).
Where uncertainty enters
The multiplier you chose. Picking "moderate" over "light" adds 316 calories to our example without anything changing about the man, and the error can run either way.
How much you move without noticing. When 16 volunteers were overfed 1,000 calories a day for eight weeks, about two-thirds of the increase in their daily energy expenditure came from non-exercise activity: fidgeting, standing, walking around. Differences in that one variable tracked a tenfold spread in how much fat they stored (Science, 1999). It was a study of overfeeding, so it shows how widely unlogged movement varies rather than measuring the gap between any two men.
What you are eating. Researchers screened 224 consecutive patients seeking obesity treatment and studied ten of them, chosen for a history of resistance to dieting, over 14 days. Their measured energy expenditure sat within 5% of prediction. Their food records did not hold up: intake was under-reported by an average of 47%, and exercise over-reported by about half (N Engl J Med, 1992).
Ten selected people cannot explain every stalled diet. What the study establishes is that self-reported intake is unreliable enough to check before the equation gets the blame.
What the equation cannot know
| It uses | It has no idea about |
|---|---|
| Body weight | How much of that weight is muscle |
| Height | How much you move outside the gym |
| Age | What your body did after months of dieting |
| Sex | Medication, thyroid problems, sleep debt |
Body composition is the substantial omission. The 1990 authors found fat-free mass alone predicted resting energy better than any other single measure, and still kept weight, height, age and sex for a usable equation. That explains why two men with the same inputs can differ, without establishing the direction or size of the error for any one muscular man. Our guide to body fat percentage covers the uncertainty in common estimates.
How to use the number properly
- Calculate resting energy, then pick the activity factor that fits your current routine. Treat it as a hypothesis.
- If tracking intake suits you, hold that number for two to four weeks. Weighing some foods for a short calibration period reduces portion error without turning every meal into a test.
- Track weight as a weekly average. Fluid and gut contents move daily weight around, often by more than the day's change in fat mass.
- Adjust from what happened rather than from a new calculation. A stable trend means a modest change, and 200 to 300 calories is a practical example rather than a validated threshold.
- Recalculate after a real change in weight, routine or health.
Step four does the work. A multi-week trend calibrates the target you are working from, which arithmetic cannot. It will not tell you which part of the estimate was wrong: resting energy, unlogged movement, training or your food records.
When the number is not the problem
If you have eaten at a genuine deficit with weighed food for a month and the trend has not moved, book an appointment rather than cutting further. A real energy deficit does reduce stored energy, so a flat month usually means intake or expenditure is not what you think it is. Thyroid problems, some medications and untreated sleep apnea all affect appetite, fluid balance and how much you move, which changes both sides of that sum. Unintentional weight loss deserves the same appointment, faster.
Two related reads: how much protein you need, which is one of the things that influences how much of a deficit comes off muscle rather than fat, and building muscle, which adds tissue that raises resting energy a little.
Frequently asked questions
What is the Mifflin-St Jeor equation?
A formula that estimates resting energy expenditure, the calories your body uses at rest. In its simplified form for men it is 10 x weight in kg + 6.25 x height in cm - 5 x age in years + 5; for women the final term is -161. It was published in 1990 from indirect calorimetry in 498 adults, and it is the equation most consumer calorie calculators run on.
Is Mifflin-St Jeor more accurate than Harris-Benedict?
It depends on who is being measured. A 2005 systematic review of healthy adults found Mifflin-St Jeor landed within 10% of measured resting metabolic rate more often than Harris-Benedict, Owen or the WHO equation. A 2023 appraisal pooling 353 individual measurements found the opposite in white adults, with Harris-Benedict closest. Treat Mifflin-St Jeor as a reasonable default rather than a proven winner.
What is the difference between BMR, RMR and REE?
Basal metabolic rate is measured under strict conditions, after an overnight fast and rest, usually on waking and without moving beforehand. Resting metabolic rate and resting energy expenditure are measured under slightly relaxed conditions and run a few percent higher. Mifflin-St Jeor was built against resting energy expenditure, so a calculator that labels its output BMR is using the word loosely.
Does the equation give me my daily calorie target?
No. It gives resting energy only. To reach a daily total you multiply it by an activity factor, and those factors are a convention rather than something the 1990 study measured. Treat the result as a starting point and correct it from what your weight trend does over a few weeks.
Is it accurate for athletes?
Less than for the general population. A 2023 meta-analysis of 29 studies in athletes found Mifflin-St Jeor among the equations whose predictions differed statistically from measured resting metabolic rate, while the Ten-Haaf equation was the most precise, landing within 10% for 80.2% of participants.
References
- A new predictive equation for resting energy expenditure in healthy individuals · American Journal of Clinical Nutrition, 1990
- Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults: a systematic review · Journal of the American Dietetic Association, 2005
- Predicting resting energy expenditure: a critical appraisal · European Journal of Clinical Nutrition, 2023
- Accuracy of Resting Metabolic Rate Prediction Equations in Athletes: A Systematic Review with Meta-analysis · Sports Medicine, 2023
- Validation of Resting Energy Expenditure Equations in Older Adults with Obesity · Journal of Nutrition in Gerontology and Geriatrics, 2022
- Revisiting the calculation of PAL from physical activity data · BMJ Open Sport & Exercise Medicine, 2026
- Discrepancy between self-reported and actual caloric intake and exercise in obese subjects · New England Journal of Medicine, 1992
- Role of nonexercise activity thermogenesis in resistance to fat gain in humans · Science, 1999
Read next
Foods and testosterone: what your diet really changes
No single food reliably boosts testosterone. See what evidence says about excess weight, dietary fat, alcohol, soy, zinc and vitamin D.
How much protein do men actually need?
The official minimum, the amount that supports muscle, and what changes after 50. Real numbers for your body weight, plus how to spread protein across the day.