BMI was built to describe populations, not people

Body mass index is one division. Take a weight in kilograms, take a height in metres, square the height, divide. That is the whole operation, and it has not changed since it was written down.

Where it came from is the part almost nobody is told, and it explains a great deal about why the number behaves the way it does.

A statistician counting a country

The ratio is usually traced to Adolphe Quetelet, a Belgian astronomer and statistician working in the 1830s. He was not building a screening test. He was doing something closer to demography: measuring large numbers of people and looking for the regularities in the aggregate, the project he called social physics. Weight divided by height squared was one of the descriptive quantities that fell out of that work, because across a population it happened to track reasonably flatly with height, which raw weight does not.

That is the design brief in full. It was chosen to behave well when you average it over thousands of people. Nothing in the derivation concerns any individual, and Quetelet did not propose it as one.

The name arrived much later. The term body mass index comes from a 1972 paper by Ancel Keys and colleagues, which compared several relative-weight indices and concluded that the Quetelet ratio was the best of them for population work. The recommendation was comparative and it was about population work. What happened afterwards — the ratio becoming a routine per-person screen printed on charts in waiting rooms — was an adoption, not a finding.

The arithmetic, with a worked example

Height in centimetres has to become metres first, which is the step people get wrong. For 75 kg at 178 cm:

1. 178 / 100 = 1.78 metres
2. 1.78 * 1.78 = 3.1684 square metres
3. 75 / 3.1684 = 23.671254

The unit of the answer is kilograms per square metre, and it is worth saying that out loud once. A body is not two-dimensional, so the denominator is not a measurement of anything physical. It is a divisor that was picked because it made a population distribution behave, and the square is a convenience rather than a claim about geometry.

What one division cannot know

A scale reports one number for everything you are made of at once. Muscle, bone, organs, fluid and fat all arrive in the same total, and division cannot separate quantities that were never separate in the input. So two people at identical heights and identical weights produce the identical index whatever the composition underneath, and the formula has no mechanism by which it could distinguish them. Our tool says the same thing in its own intro: it ignores muscle, frame and composition.

The cut-points carry a second limitation that is easy to miss. They are conventional boundaries fitted to particular reference populations, and how well they transfer to a population that was not in that reference data is a live and long-running question. Bodies differ systematically in frame and in typical composition across ancestries, and a fixed boundary applied to everyone inherits none of that. That is a statement about the boundaries, not about anybody standing on either side of one.

Reading a ratio as a ratio

The honest description is short. The index is the ratio of a weight to the square of a height, computed exactly, and it is a real number about a real measurement. It is not a measurement of body composition, and it is not a verdict. What it means for one particular person is a clinical question, and answering it requires somebody who has actually examined that person and knows the rest of their history. A calculator has none of that and cannot acquire it.

The useful posture, if you are going to compute it at all, is to treat the number as an input to a conversation rather than the output of one.

Running it

The BMI calculator takes weight in kilograms and height in centimetres, does the conversion and the division on your own device, and prints the index to six decimal places along with the conventional cut-point band the value falls in. Those bands are population conventions and the tool labels them as such rather than as a statement about you.

It is informational and it is not medical advice. It has no view about anybody, and it cannot have one — it has seen two numbers.