How BMI Is Calculated, and What It Cannot Tell You
BMI is your weight in kilograms divided by your height in metres squared. The BMI Calculator does it and shows the category, and the history explains most of its limitations.
Where it came from
It was devised in the 1830s by Adolphe Quetelet, a Belgian mathematician and statistician who was not studying health at all. He was interested in describing the characteristics of populations, and he wanted a simple index that stayed roughly constant across people of different heights.
It became a clinical screening tool in the 1970s because it was cheap, required only a scale and a tape, and correlated reasonably well with body fat across large groups. That is the key qualifier: it was built to describe populations and is routinely used to judge individuals, which is a different job.
The categories, and the squaring problem
The conventional bands are under 18.5 underweight, 18.5 to 24.9 normal, 25 to 29.9 overweight, and 30 and above obese.
The arithmetic has a known quirk. Dividing by height squared rather than cubed means the index does not scale neutrally with height - it tends to read slightly high for tall people and slightly low for short ones, because bodies do not grow in exactly two dimensions. Various alternative indices have been proposed to correct it; none has displaced BMI, because BMI is simple and entrenched.
What it cannot see
Muscle. BMI counts total weight and does not care what the weight is. A muscular athlete routinely reads as overweight or obese on it, which is the most frequently cited criticism and a fair one.
Where fat is carried, which matters considerably for cardiovascular risk. Abdominal fat behaves differently from fat on the hips and thighs, and BMI is blind to the distinction - which is why the Waist-to-Hip Ratio Calculator adds real information that BMI cannot.
Ethnicity. The relationship between BMI and health risk differs between populations, and several health bodies use lower thresholds for people of South Asian and Chinese descent for exactly that reason.
Age and sex, beyond the separate charts used for children.
What it is still good for
Screening at scale, and tracking one person over time. Both of those are legitimate.
If your BMI has moved several points in a year, something has changed, and that is worth knowing regardless of the number's limitations. If it sits far outside the normal range, it is a reasonable prompt for a conversation with a doctor - not a diagnosis, a prompt.
What it is not is a verdict on a person, and a figure from a 1830s population index is a poor reason to feel anything in particular. Anything you are actually weighing up about your health belongs with a GP, who can see the things BMI cannot.
Common questions
Does BMI work for children?
Only against age and sex percentile charts, not the adult bands.
Is there a better single measure?
Waist circumference, or waist divided by height, both predict cardiovascular risk better. The Body Fat Percentage Calculator estimates composition directly.
Why is the imperial formula different?
The same ratio with a conversion constant of 703 folded in.
Curious what the arithmetic gives?