Run the same person through five published metabolism equations and you get five different answers. Take an 80-kilogram, 180-centimeter, 30-year-old man. The Mifflin-St Jeor equation puts his resting burn at 1,780 calories a day. The revised Harris-Benedict formula says 1,854. Katch-McArdle, if he carries 15 percent body fat, lands at 1,839. The original 1919 Harris-Benedict equation says 1,865, and the Cunningham equation climbs all the way to 1,996. Same body, same morning, a spread of more than 200 calories, and not one of those equations is broken.
This is the part of a calorie estimate almost nobody inspects. You pick a calculator, it reports a number, and the equation doing the work stays hidden behind the form. But the choice of equation is a real decision with a documented history, and understanding what it does (and, more usefully, what it doesn't do) tells you how much of your daily target you should actually trust.
What a BMR Equation Is Actually Doing
Basal metabolic rate is the energy your body spends doing nothing: keeping your heart going, holding your temperature, running the quiet chemistry that stays on while you sleep. It is a measurable quantity. In a lab it gets read with indirect calorimetry, which tracks how much oxygen you consume and how much carbon dioxide you produce over a set window under strict resting conditions. That measurement is the ground truth every equation is trying to approximate.
The catch is that indirect calorimetry needs a clinic, a fasted subject, and a technician. No calculator can do that from a web form. So what a BMR equation gives you instead is a regression: a formula fit to a large group of people whose resting burn was measured directly, tuned to predict that measured value from cheap inputs like age, height, weight, and sex. The output is a population average for someone shaped like you, not a reading of your own metabolism. Two people with identical stats can have genuinely different resting rates, and the equation reports the same figure for both. So keep the frame straight from the start: the number is a solid estimate of a group, applied to you.
A Hundred Years of Estimating Resting Burn
The oldest formula still in circulation is Harris-Benedict, published by J. Arthur Harris and Francis Benedict through the Carnegie Institution in 1919. They measured metabolism in a few hundred subjects and fit an equation to it, and for decades that equation was the default across dietetics and exercise science. In 1984 Roza and Shizgal reworked the coefficients against a larger, cleaner dataset, producing the "revised" Harris-Benedict most modern tools mean when they list it.
The equation that eventually took over came in 1990. Mifflin, St Jeor, and their colleagues published a new predictive equation in the *American Journal of Clinical Nutrition*, built from measurements on 498 healthy adults. They set out specifically to fix a drift: the older Harris-Benedict numbers were fit to an early-twentieth-century population and tended to read high against the heavier, more sedentary bodies of the late-twentieth-century sample.
A separate lineage took a different input entirely. Katch-McArdle and the Cunningham equation (Cunningham's dates to 1980) drop total body weight and work from lean body mass instead, on the logic that muscle and organ tissue drive nearly all resting burn while stored fat contributes very little. Those formulas can be sharper, but only if you feed them an accurate body-fat percentage, which is a measurement in its own right. Every one of these equations carries the fingerprint of the group it was fit to, which is exactly why they disagree.
What Swapping Equations Does to the Number
Here is the reference man run through each formula, holding his stats fixed:
| Equation | Year | Inputs it uses | BMR |
|---|---|---|---|
| Mifflin-St Jeor | 1990 | age, height, weight, sex | 1,780 |
| Katch-McArdle | ~1996 | lean body mass (needs body fat %) | 1,839 |
| Harris-Benedict (revised) | 1984 | age, height, weight, sex | 1,854 |
| Harris-Benedict (original) | 1919 | age, height, weight, sex | 1,865 |
| Cunningham | 1980 | lean body mass (needs body fat %) | 1,996 |
Cunningham was derived from lean, athletic subjects and is known to run high on the general public, so set it aside. What's left is telling: the three mainstream weight-based formulas plus Katch-McArdle all sit inside an 85-calorie band. For a typical person deciding which equation to believe, that is the entire size of the decision. Eighty-five calories is a slice of toast. It is not the reason your last three calculators disagreed by hundreds.
Why the Field Settled on Mifflin-St Jeor
The consolidation wasn't a matter of taste. In 2005, a systematic review by Frankenfield and colleagues in the *Journal of the American Dietetic Association* compared the common predictive equations against resting rates measured by calorimetry. Mifflin-St Jeor predicted within 10 percent of the measured value more often than Harris-Benedict, Owen, or the WHO/FAO/UNU equations, and it held that edge in both nonobese and obese adults. That is the practical bar for a formula meant to work on strangers, and Mifflin-St Jeor cleared it most consistently.
The body-fat formulas were the interesting near-miss. Katch-McArdle can beat Mifflin-St Jeor for a person whose lean mass is known, because it is measuring the tissue that actually burns. But "known" is doing heavy lifting. A DEXA scan or careful caliper reading gives a real number; a glance in the mirror gives a guess, and a wrong body-fat guess feeds straight into the result and reintroduces more error than the fancier equation removed. Mifflin-St Jeor won partly on accuracy and partly on humility: it asks only for inputs people can supply correctly. A tool that leads with it and treats the lean-mass equations as optional is making a defensible call rather than a lazy one.
The Variable That Dwarfs the Formula
Formula choice is where people spend their skepticism, and it is close to the smallest lever on the board. Multiply the reference man's 1,780-calorie baseline by the standard activity factors, which run from about 1.2 for a sedentary week to 1.9 for an extra-active one, and his total daily expenditure ranges from 2,136 to 3,382 calories. That is a 1,246-calorie spread produced by a single dropdown, and it is roughly fifteen times the gap between the mainstream formulas.
Those activity multipliers come from population research too, much of it grounded in doubly-labeled water studies, which are the reference standard for measuring real-world energy expenditure over days of ordinary life. But they are wide buckets by design, and everyone's real week straddles the line between two of them. The upshot is uncomfortable and clarifying at once: agonizing over Mifflin-St Jeor versus revised Harris-Benedict is optimizing an 85-calorie question while a 1,200-calorie question sits untouched one field away. The equation is not where your uncertainty lives.
When a Different Equation Earns Its Place
None of this makes Mifflin-St Jeor universal. It uses height and weight as a stand-in for body composition, and that proxy breaks at the edges. A heavily muscled lifter at a "high" weight gets treated as though the extra mass were average tissue, which overstates the fat he isn't carrying and can nudge his BMR estimate too low relative to how much lean tissue he actually runs. A very lean endurance athlete can land on the wrong side of the same approximation. For those bodies, and only when there's a real measured body-fat figure to feed it, Katch-McArdle or Cunningham track more honestly because they start from lean mass instead of inferring it.
That is the whole trade. The lean-mass equations trade a measurement most people can't supply for accuracy at the extremes most people don't occupy. If you have the number and you sit at the far end of the composition curve, reach for them. If you are somewhere in the broad middle with a bathroom scale and a tape measure, the extra input is a liability, not an upgrade, and the 1990 equation built for exactly your situation is the right default.
Where accuracy actually comes from, once the equation is chosen sensibly, is being honest about your activity bucket and rerunning the math as your weight moves, since the baseline shifts under you every few pounds. If you want the Mifflin-St Jeor number with every activity level laid out beside it, so you can see that 1,200-calorie spread instead of committing to one guess, TDEEfy runs it entirely in your browser and keeps your inputs ready for the next recalculation.