Reading health calculators honestly

Body mass index is weight divided by the square of height. That is the entire calculation. It contains no information about muscle, fat, bone density, where weight is carried, or anything else — because none of those were inputs.

It was devised in the nineteenth century as a way of describing populations and was never intended as an individual diagnostic. It survives because it is cheap, needs no equipment, and correlates well enough across large groups to be useful for public-health screening. Those are real virtues at population scale and they do not transfer cleanly to one person.

The squaring is also a known distortion. Because it scales with height squared while bodies scale closer to height cubed, BMI tends to read high for tall people and low for short people of identical build. This is a property of the formula, not a finding about them.

The clearest failure is muscle. Muscle is denser than fat, so a well-trained athlete can land in the "overweight" or even "obese" band while carrying very little fat. The number is arithmetically correct and clinically meaningless in that case.

The opposite failure is less discussed and more consequential: someone can sit squarely in the healthy band while carrying a high proportion of fat and very little muscle. BMI cannot see this at all, and a reassuring number can delay a conversation worth having.

The standard adult categories are also derived largely from European-ancestry populations, and several health bodies apply lower thresholds for people of South Asian, Chinese and other ancestries, where metabolic risk appears at a lower BMI. And the adult categories do not apply to children and adolescents at all, who are assessed against age and sex percentiles instead.

A daily calorie estimate is built in two steps. First, a resting metabolic rate: the energy your body uses doing nothing at all, which is the large majority of daily expenditure. The Mifflin-St Jeor equation is the one most commonly recommended for healthy adults, and it takes weight, height, age and sex as inputs.

Second, that resting figure is multiplied by an activity factor to account for everything you actually do. This is where most of the uncertainty enters. The multipliers are coarse bands with plain-language labels, and people are poor judges of which band they belong to — "moderately active" is claimed far more often than it is true.

Both steps are population averages. Individual resting metabolic rates vary meaningfully from the equation’s prediction even among people with identical inputs, because the equation has no way to see body composition, thyroid function, medication or genetics. Treat the output as a starting hypothesis to be adjusted against what actually happens.

The most useful property of both figures is not their absolute value but their direction over time. One BMI reading is a weak signal; the same measurement taken the same way over a year is a much stronger one. A calorie estimate that turns out to be 200 too high becomes accurate the moment you adjust it against real observed change.

It also helps to pair each number with something it cannot see. Waist circumference, or waist-to-height ratio, captures where weight is carried, which BMI cannot, and is measurable with a tape. For energy needs, two or three weeks of consistent weight data tells you more about your real requirement than any equation can.

And treat a precise-looking figure with appropriate suspicion. A calculator that reports 2,347 calories is reporting the formula’s arithmetic, not a measurement accurate to the calorie. Rounding to the nearest fifty is a more honest representation of what is actually known.