How Accurate Is BMI? What It Gets Right and Wrong

BMI is the most widely used health metric in the world and the most widely criticised, and both of those things are true for the same reason: it uses only two numbers. Height and weight, one division, done. That is why every doctor's office, insurer and public health department reaches for it — and why it can be badly wrong about you specifically.

So how accurate is BMI, really? The short version: accurate enough to sort a population, not accurate enough to judge an individual. Here is what the evidence says, where the errors come from, and what to measure instead when BMI is likely to mislead you.

What BMI measures, and what it does not

BMI is weight in kilograms divided by height in metres squared. That is the entire calculation, unchanged since Adolphe Quetelet devised it in the 1830s as a way of describing populations — not, as he was explicit about at the time, individuals.

What it captures is whether your weight is unusual for your height. What it cannot capture, because the inputs simply do not contain the information:

• What the weight is made of. Muscle and fat weigh the same to a scale. A 90 kg athlete and a 90 kg sedentary person of the same height get identical BMIs and wildly different health risks.
• Where the fat is. Visceral fat around the organs is far more dangerous than fat on the hips and thighs. BMI cannot tell them apart at all.
• Bone density and frame size. Both contribute weight and neither is a health risk.
• Age-related composition change. Adults lose muscle and gain fat with age at a stable weight. BMI records no change whatsoever.

How accurate is BMI as a body fat estimate?

When BMI is validated against DEXA scanning, the correlation across a large group is strong — typically in the range of r = 0.7 to 0.8. That is why BMI works for public health: run it across ten thousand people and the group averages come out about right.

For one person, that same correlation implies error bars of roughly 5–10 percentage points of body fat. Studies that compare BMI-based obesity classification with DEXA-based classification consistently find substantial misclassification in both directions:

• False positives: people classed as overweight or obese by BMI whose body fat is in the healthy range. Concentrated among muscular, athletic and physically heavy-framed people.
• False negatives: people with a normal BMI whose body fat is high and whose metabolic markers are poor — the pattern often called "normal weight obesity". This is the more dangerous error, because the normal BMI reassures everyone involved.

Accuracy also varies by group. BMI tends to understate body fat in older adults, in people of South Asian descent (several health systems use lower BMI thresholds for this reason), and in anyone who has lost significant muscle. It tends to overstate body fat in trained athletes and in people with dense builds.

Is there a more accurate BMI calculator?

No, and it is worth being direct about this, because it is a common search. BMI is a single fixed equation. Every calculator on the internet — ours included — performs the same division. A calculator cannot be more accurate than another one; it can only be easier to use, or wrong.

What varies is the quality of what goes into it. Two things genuinely improve the number you get:

1. Measure properly. Weigh yourself in the morning, unclothed, before eating. Measure height without shoes, standing straight against a wall. Guessed height is the single most common source of error in a home BMI — most adults are shorter than they think, and each centimetre of overstated height understates BMI by about 0.3.
2. Interpret it in context. BMI answers one question. Pair it with a measurement that answers a different one.

You can run an accurate BMI in seconds with our quick BMI calculator — it uses the standard WHO formula and nothing is stored or sent anywhere.

Which body measurement is the most accurate?

It depends what you want to know. Each metric answers a different question, and "accuracy" is only meaningful once you have decided which question matters.

If you can only take one measurement, take a waist circumference — it beats BMI on predicting the outcomes people actually care about. If you can take two, take BMI and waist. The pair is a genuinely good screen and costs nothing but a tape measure.

BMI, BMR or waist-to-hip ratio: which should you actually track?

These three get compared constantly, and they are not substitutes for each other.

• BMI is a screening number. Check it once or twice a year, or when your weight has changed noticeably. It is a starting point, not a destination.
• Waist-to-hip ratio is a risk number. It tells you whether the weight you carry is in a place that raises cardiovascular and diabetes risk. If you only track one thing for health, track this. Our comparison of waist-to-hip ratio vs. BMI goes through the evidence.
• BMR is a planning number, not a health measure at all. It estimates resting calorie burn so you can set a sensible intake target. It says nothing about whether you are healthy. Use the calorie calculator when you need it, and do not treat it as a diagnostic.

Does BMI still work at the extremes?

At the high end — BMI above about 35 — BMI becomes more reliable as a body fat indicator, not less. It is very difficult to reach a BMI of 40 on muscle mass, so the false-positive problem largely disappears. Someone with a BMI of 48 is, with near certainty, carrying a body fat percentage well into the range associated with serious risk; the precise figure barely changes the clinical picture, which is why guidelines act on the BMI directly in that range.

At the low end, BMI is less informative than it looks. A BMI of 17 can reflect a naturally slight frame, or it can reflect muscle wasting from illness or undereating — and those need very different responses. Below about 18.5, a body fat percentage and a clinical assessment tell you far more than the BMI does.

The band where BMI is weakest is 25 to 35 — exactly the band most adults fall into. That is where muscular builds get flagged as overweight, and where a normal-looking number can hide a high visceral fat load. It is also the band where adding a waist measurement changes the answer most often, which is why most clinical guidelines now recommend exactly that.

How to use BMI well

The bottom line

BMI is accurate at the job it was designed for and inaccurate at the job people ask of it. As a fast, free, universally available screen that tells you whether your weight is worth a closer look, it works well and nothing has replaced it. As a statement about your individual health, it is a rough guess with a five to ten point margin.

The fix is not a better formula — there isn't one. The fix is a second measurement. Run your BMI, then spend two minutes with a tape measure on your waist-to-hip ratio. Where the two agree, believe them. Where they disagree, believe the tape.