How BMI is calculated
BMI is your weight in kilograms divided by the square of your height in metres (kg/m²). Someone who weighs 70 kg and is 1.75 m tall has a BMI of 70 ÷ (1.75 × 1.75) ≈ 22.9. A second example: a 60 kg adult at 1.60 m gets 60 ÷ (1.60 × 1.60) ≈ 23.4 — also healthy, despite quite different height and weight.
In imperial units the formula needs a conversion factor: BMI = 703 × weight(lb) ÷ height(in)². A 154 lb person at 69 inches gets 703 × 154 ÷ 4,761 ≈ 22.7, which is the same person as the first example, give or take rounding.
BMI categories (adults)
- Under 18.5 — Underweight
- 18.5–24.9 — Healthy weight
- 25–29.9 — Overweight
- 30–34.9 — Obesity, class 1
- 35–39.9 — Obesity, class 2
- 40+ — Obesity, class 3
The full BMI chart sets these out alongside what each band signals.
Where the formula comes from
BMI was devised in the 1830s by the Belgian mathematician Adolphe Quetelet, as a way of describing how body sizes were distributed across a population. He was explicit that it was a statistic for groups and not a measure of any individual — a caveat that did not survive the journey into routine clinical use.
The modern thresholds are much newer. A WHO expert consultation adopted the 18.5 / 25 / 30 cut-offs in 1997, chosen because that is roughly where population-level disease and mortality risk begins curving upward in large cohort studies. Nothing physiological happens at exactly 25.0; a BMI of 24.9 and one of 25.1 describe indistinguishable bodies.
What BMI can and can't tell you
It cannot tell muscle from fat. This is the headline limitation and it is a serious one: a lean, heavily muscled person routinely reads "overweight" or even "obese" while carrying very little fat, because muscle is denser than fat and BMI only sees total mass.
It also cannot see where fat is stored, which matters more for metabolic risk than the total. Visceral fat around the abdominal organs behaves quite differently from fat on the hips and thighs, and BMI treats them identically.
And it does not account for age. Older adults typically lose muscle and bone while gaining fat, so someone can hold a steady, healthy-looking BMI for twenty years while their body composition changes considerably underneath it.
Who should not use this number
- Athletes and muscular people — the number will over-read. Use a body-fat estimate instead.
- Children and teenagers — they need age- and sex-specific percentile charts, not adult bands.
- Pregnant people — BMI is not meaningful during pregnancy.
- People of South Asian, Chinese and some other Asian ancestry — cardiometabolic risk tends to appear at lower BMIs, and many health bodies apply reduced thresholds, often around 23 and 27.5.
- Older adults — a slightly higher BMI may be protective in later life.
- Anyone with a history of disordered eating — please read our medical disclaimer before using tools like this one.
What to pair it with
Two tape-measure numbers cover most of what BMI misses, and neither costs anything. Waist circumference captures fat distribution — the thing that matters most for metabolic risk. Body-fat percentage, estimated with the US Navy method, separates muscle from fat. If BMI reads high but both of those look unremarkable, BMI is probably the number that is wrong. If all three agree, the signal is real.
Using your result
Treat it as a prompt to look further, not a verdict. A single BMI reading says nothing about your blood pressure, blood glucose, lipids, fitness, sleep, or diet quality — all of which predict your health considerably better. If the number concerns you, a conversation with a doctor who can see the rest of the picture is a far better next step than a crash diet.
If you do want to shift it, the durable route is unglamorous: a modest calorie deficit rather than a severe one, enough protein and resistance training to hold onto your muscle, decent sleep, and a timeframe measured in months. Watch the trend over weeks — normal shifts in water, food volume, and glycogen move the scale by a kilogram or two in either direction without any change in body fat at all.