Mifflin-St Jeor vs Every Other BMR Formula: Which One Is Actually Right For You

Five Equations, One Question

Type your height, weight, age, and sex into any calorie calculator on the internet and a number comes back. The number looks authoritative. It is usually presented to the kilocalorie, sometimes with a confident note about how many days until you hit some target weight. What the interface almost never tells you is which equation it used to produce that number, and the answer matters more than the interface design suggests. There are at least five BMR equations in serious use in 2026, they disagree by hundreds of calories per day for the same person, and the right one for you depends on facts about your body the calculator did not ask for.

This post is the working field guide to that choice. The goal is to leave you able to read any calorie calculator's output, know roughly which equation produced it, understand where that equation breaks, and either pick a better one or correct for the error in your head. None of this requires being a sports scientist. It requires about fifteen minutes and a willingness to take the numbers less literally than the interface invites you to.

What BMR Actually Is (And Why The Number Is Slippery)

Basal metabolic rate is the energy your body spends doing nothing — keeping your heart beating, your kidneys filtering, your brain ticking over at idle, your core temperature where it needs to be. Measured under laboratory conditions, BMR is the calorie cost of being alive for twenty-four hours in a thermoneutral room, post-absorptive, completely at rest, awake but unstressed. It is the floor of your daily expenditure. Everything above the floor — walking to the kitchen, digesting lunch, holding tension in your shoulders during a Zoom call, the thermic effect of food — gets added on top to produce TDEE, your total daily energy expenditure.

The gold standard for measuring BMR is indirect calorimetry: you breathe into a metabolic cart for thirty or forty minutes, the cart measures the oxygen you consume and the carbon dioxide you produce, and the ratio (your respiratory quotient) plus the gas exchange volumes get translated into a real calorie burn. Done correctly, indirect calorimetry has a stated precision in the single-digit-percent range. It is also expensive, requires a clinic or sports lab, and is not happening for the average person trying to recompose.

Everything else — including every calorie calculator you have ever used — is an equation that takes anthropometric inputs (the numbers your body presents from the outside: height, weight, age, sex, sometimes body fat percentage) and produces a population-average estimate of what indirect calorimetry would have measured. The equations are fit on real measurement data, but the fit is imperfect by construction. You are getting the average of a distribution, not the truth about you specifically.

Harris-Benedict (1919, Revised 1984)

The original. James Arthur Harris and Francis Gano Benedict published their equations in 1919 based on indirect calorimetry measurements of 239 subjects. Their formula was the standard for most of a century and is still the default in a surprising number of older textbooks and clinic intake forms. Roza and Shizgal published a revision in 1984 (sometimes called the Harris-Benedict Revised equation) that updated the constants against a larger and more contemporary sample.

The Harris-Benedict equations weight height, weight, and age separately, with sex-specific intercepts. The 1984 revision for men, in metric units, looks like BMR = 13.397 × weight + 4.799 × height − 5.677 × age + 88.362, with a different set of constants for women. The numbers are oddly precise for a 1919 data set, and that precision is part of the problem.

The honest critique of Harris-Benedict is that it overestimates BMR for the modern population, often by ten to fifteen percent at the high end. The 1919 cohort was leaner, more active, and demographically narrow compared with the population using the equation today. If you have ever wondered why a calorie calculator told you that you maintain on 2,800 calories and you have been gaining weight at 2,400, Harris-Benedict is the most common culprit.

Mifflin-St Jeor (1990)

The current default, and the one most reputable modern calculators — including ours at our TDEE calculator — use under the hood. M.D. Mifflin and S.T. St Jeor published the equation in the American Journal of Clinical Nutrition in 1990, fit on a sample that was more contemporary, more demographically representative, and larger than Harris and Benedict's. The American Dietetic Association reviewed the literature in 2005 and concluded that Mifflin-St Jeor was the most accurate of the common anthropometric equations for non-obese and obese adults alike, predicting measured BMR within roughly ten percent for the majority of subjects.

The formula is clean enough to keep in your head. For men: BMR = 10 × weight (kg) + 6.25 × height (cm) − 5 × age − 5 + 10 (often written as +5 after combining constants). For women: BMR = 10 × weight + 6.25 × height − 5 × age − 161. The sex term is large — about 166 calories per day — and reflects the average difference in lean body mass and organ size between men and women at the same height, weight, and age.

Mifflin-St Jeor is the right default for most people for the same reason Goodhart's law applies to other defaults: it is what the equation was fit on. If you are a sedentary-to-moderately-active adult of average body composition, between roughly nineteen and seventy years old, not at extremes of weight or height, the equation will be approximately right. It will not be exactly right, and the residual error is structured in ways the next equations address.

Where Mifflin-St Jeor Breaks: The Lean Body Mass Problem

The deepest assumption in any height-weight-age equation is that two people with the same height, weight, age, and sex have roughly the same metabolic rate. They do not. A 75-kilogram man who is 12 percent body fat has substantially more lean tissue — muscle, organ mass, the metabolically active stuff — than a 75-kilogram man who is 28 percent body fat. Lean tissue burns more calories at rest than fat tissue does. Mifflin-St Jeor, by ignoring body composition, treats the two men identically and gets one of them meaningfully wrong.

For most people the error is small enough to ignore because most people are within a normal range of body composition. For two populations the error gets large enough to matter: trained athletes (more lean tissue than the equation expects) and people with substantial excess adipose tissue (less metabolically active mass per kilogram of body weight than the equation expects). In the first case the equation under-predicts. In the second it over-predicts. Both errors push people toward dieting strategies that backfire.

Katch-McArdle and Cunningham: When Body Composition Is Known

The cleanest fix for the lean body mass problem is to use an equation that takes lean body mass as an input. Two are in common use: Katch-McArdle and Cunningham. Both were published around 1980 and both are functions only of lean mass, with no separate term for sex, age, or height — the logic being that lean mass already encodes most of what those terms were proxying for.

The Katch-McArdle equation is BMR = 370 + 21.6 × lean body mass (kg). The Cunningham equation is BMR = 500 + 22 × lean body mass (kg). They are close enough numerically that the choice between them is mostly a matter of which sports-nutrition tradition you grew up in; the difference between the two is in the noise of measurement error for lean body mass itself.

The catch is that you need an honest lean body mass number, which means you need a body composition measurement. DEXA is the gold standard, hydrostatic weighing and air-displacement plethysmography (BodPod) are close, and bioimpedance is the most accessible but the least reliable, often by ten percentage points in either direction. A skinfold caliper measurement done by a trained practitioner can be more accurate than a cheap bioimpedance scale. If your lean body mass number is wrong, you will get a confidently wrong BMR, and the confidence is the dangerous part.

If you are an athlete with a recent DEXA scan, Katch-McArdle or Cunningham is the right equation. If you are guessing at your body fat percentage from a mirror, Mifflin-St Jeor with its known error bars is the more honest choice.

Schofield (1985) and the WHO/FAO Equations

The Schofield equations were derived in 1985 from a large international dataset and adopted by the WHO and FAO as their reference equations for energy requirements. There are separate Schofield equations for each of six age brackets (under 3, 3 to 10, 10 to 18, 18 to 30, 30 to 60, over 60), and they take only weight as the primary input, with optional height as a secondary term.

Schofield is the standard in clinical nutrition and dietetics outside the United States and in pediatric populations. The age stratification matters more than the equation's other properties: BMR per kilogram of body weight changes substantially across the life course, and an equation fit on a single age-bracket sample will mispredict for ages outside that bracket. If you are working with a child or an older adult, or designing a meal plan for a clinical context, Schofield or one of its successors is the conservative choice.

From BMR to TDEE: The Activity Multiplier Is Where The Real Error Lives

Every BMR equation has measurable error in the single-digit-percent range. The activity multiplier you apply to get TDEE has error several times larger. The standard Harris-Benedict-derived multipliers in wide use — sedentary 1.2, light 1.375, moderate 1.55, active 1.725, very active 1.9 — are buckets, not measurements, and they hide an enormous amount of behavioral variation.

Two people who both report "moderate exercise three to five times a week" can differ by 500 calories a day in non-exercise activity thermogenesis (NEAT) alone — the energy spent fidgeting, standing instead of sitting, pacing during phone calls, taking the stairs. NEAT is genuinely variable across individuals and is the single largest source of error in any TDEE estimate. The activity multiplier captures none of it.

The practical implication: any TDEE number you get from a calculator should be treated as a starting point, not a setpoint. Eat at the predicted maintenance number for ten to fourteen days, weigh yourself at the same time of day under the same conditions, and adjust based on what your body weight actually does. The calculator gives you the hypothesis. Your scale gives you the data.

How To Pick The Right Equation For You

The short version. If you have no body composition data and you are an adult of roughly average build between nineteen and seventy, use Mifflin-St Jeor and move on. If you are an athlete with a recent DEXA or BodPod measurement, use Katch-McArdle or Cunningham and trust them more than the anthropometric equations. If you have substantial excess body fat, expect Mifflin-St Jeor to over-predict and start a few hundred calories below the number it gives you. If you are working in a clinical context or with a child or older adult, use Schofield. If you are using a calculator that defaults to Harris-Benedict in 2026, switch calculators.

Above all, treat the number as a hypothesis. The equations are decades of careful work and they are still population averages applied to an individual. The honest calorie calculator gives you a starting point and tells you to verify it against your own body. The dishonest one gives you a precise number and implies it is the truth.