Is BMI accurate?

BMI is accurate about populations and unreliable about individuals. It compares your weight with your height and knows nothing else, so it cannot separate muscle from fat or see where you carry weight. Studies of national health data find it misclassifies the cardiometabolic health of roughly one adult in three. Treat it as a first question, not an answer.

Last reviewed 25 September 2026. Not medical advice.

What BMI actually measures

Your weight in kilograms divided by your height in metres squared. That is the entire formula. It was devised in the 1830s by Adolphe Quetelet, a statistician interested in describing populations, and it was never designed to assess an individual. Its usefulness and its limits both come from the same place: it is extremely cheap to calculate and it throws away everything except two numbers.

At population scale that trade works. The largest analysis of the question, pooling 239 prospective studies and 10.6 million participants, found a clear relationship between BMI and mortality, with the lowest risk between about 20 and 25. If you want to describe the health of a country, BMI is a reasonable instrument.

The problem is that you are not a country.

Illustrative, not measurements of real people. The tissue proportions come from Clarys, Martin and Drinkwater's cadaver dissection study (Human Biology, 1984), which found muscle at 50 percent of fat-free weight with a standard deviation of 4.4, and bone at 20.6 percent with a standard deviation of 2.6. Each of these three sits at a different point inside that normal spread.

How often it gets an individual wrong

This has been measured. Tomiyama and colleagues took 40,420 adults from the US national health survey and compared their BMI category against actual cardiometabolic markers: blood pressure, triglycerides, cholesterol, glucose, insulin resistance and C-reactive protein.

Nearly halfof people classed as overweight were cardiometabolically healthy

29%of people classed as obese were cardiometabolically healthy

Over 30%of people in the normal weight band were cardiometabolically unhealthy

Scaled to the US population, that is around 75 million adults whose health BMI describes incorrectly. Note that the error runs in both directions. BMI does not simply worry healthy people; it also reassures people who have something worth looking at.

The four things it cannot see

Muscle. Muscle is denser than fat, so a lean, well trained person routinely lands in the overweight band. This is the failure everyone knows about, and it is real.
Where the weight sits. Fat stored around the organs behaves differently from fat under the skin, and carries more metabolic risk. Two people at the same BMI can differ enormously here, and BMI reads them identically.
Height, at the extremes. Dividing by height squared is a convenient approximation, not a law of nature. As Rothman set out, it flatters short people and penalises tall ones. Trefethen proposed an exponent of 2.5 in 2013 to correct this; no health body has adopted cut points for it, so it remains a footnote.
Age and ancestry. Height falls and composition changes over a lifetime, and several health services accept a slightly higher BMI after about 65. Separately, the WHO advises lower cut points of 23 and 27.5 for people of South and East Asian descent, because the same BMI carries more body fat in those populations.

What to measure as well

The single best addition is a tape measure. In a systematic review and meta-analysis, waist to height ratio outperformed both waist circumference and BMI as a screening tool for cardiometabolic risk factors. It costs one extra measurement and the rule fits in a sentence.

The rule is one line long: your waist should measure less than half your height.

A waist measurement also buys you a far better body fat estimate than BMI can give. Which brings us to the honest part.

The methods disagree with each other too

It would be neat to say "use body fat instead" and stop. But body fat estimated from a tape measure is itself a family of competing formulas, and they do not agree.

One man, 1.70 m, 70 kg, 80 cm waist, 38 cm neck. Three published methods, a spread of eight points.

Those three are the US Navy circumference method (Hodgdon and Beckett, 1984), the Deurenberg equation from BMI (Deurenberg and colleagues, 1991), and Relative Fat Mass (Woolcott and Bergman, Scientific Reports, 2018). Nearly eight percentage points separate the highest and lowest estimate for the same man on the same day. That is not a reason to distrust all of them; it is a reason to pick one method, measure yourself the same way each time, and watch your own trend rather than comparing your number with somebody else's from a different formula.

This is also why every calculator on this site tells you which method produced your number and cites the paper it came from. A body fat percentage with no method attached is not information.

So what should you do with your BMI?

Use it as the cheap first pass it is. If it sits inside the healthy band and your waist is under half your height, those two agreeing is genuinely reassuring. If they disagree, the waist is usually the more informative of the two and that disagreement is worth raising with a doctor.

What you should not do is treat a single BMI reading as a description of your health, in either direction.

Calculate your BMI Waist to height ratio Body fat percentage

Common questions

Is BMI accurate?

It is accurate at what it measures, which is weight against height, and that is all it measures. Across large populations it tracks health outcomes well, which is why health services use it. For one individual it is a screening number with known blind spots: it cannot separate muscle from fat, it does not know where you carry weight, and it misreads the very tall, the very short, the heavily trained and the elderly.

How often does BMI get an individual wrong?

More often than most people expect. Analysing 40,420 adults in the US national health survey, Tomiyama and colleagues found that nearly half of people classed as overweight and 29 percent of those classed as obese were cardiometabolically healthy, while more than 30 percent of people in the normal weight band were not. On those numbers, BMI alone misclassifies roughly 75 million American adults.

Should I ignore my BMI then?

No. Ignoring it is as wrong as treating it as a verdict. It is a cheap, fast screening number that tells you whether a closer look is worthwhile. The mistake is stopping there. Add a waist measurement and you learn something BMI structurally cannot tell you.

Is BMI accurate for athletes and muscular people?

This is its most famous failure. Muscle is denser than fat, so a lean, heavily trained person routinely lands in the overweight or obese band on BMI alone. The arithmetic is correct and the conclusion is wrong. For this group a waist based measure or a body fat estimate is far more informative.

What is more accurate than BMI?

For predicting cardiometabolic risk, waist to height ratio has outperformed BMI in meta-analysis. It needs one extra measurement, a tape around your middle, and the rule is simply to keep your waist under half your height. For body composition specifically, a waist based body fat estimate tells you considerably more than BMI can.

Sources

Every figure above comes from one of these. Check them.

Every method, with its source attached

The app runs BMI, waist ratios and three body fat methods side by side, and tells you when they disagree about you.

Coming soon

Free, no account, and it works offline. 100,000+ downloads on Google Play.

This is not medical advice. Body Numbers is a calculator. It does not diagnose, treat or monitor any condition, and no number it produces should replace a conversation with a doctor or a dietitian. If a result worries you, or if you are pregnant, unwell, or under eighteen, speak to a health professional.