The formula takes two measurements: your weight and your height. It divides one by the square of the other. It then places you in one of four categories. Elite athletes who spend thirty or more hours per week training, professional cyclists, competitive rowers, sprinters, routinely land in the overweight or obese categories. The formula has been doing this for decades. The reason is not that the formula is broken. It is that the formula was designed in 1832 to describe the statistical shape of citizens, and nobody has replaced it.
BMI, body mass index, calculates weight in kilograms divided by height in meters squared. The result is a single number that millions of people receive annually from their doctors as a primary screening result for metabolic health. What is less commonly communicated is that the formula was designed for population statistics, its creator explicitly said it should not be applied to individuals, and when the physiologist who coined the term body mass index formally evaluated its accuracy in 1972, he concluded it accounted for less than half of the actual variation in body fatness.
The gap between what BMI claims to measure and what it actually measures has become increasingly difficult to ignore, especially for anyone who carries a higher-than-average amount of muscle.
The Astronomer Who Invented BMI
Adolphe Quetelet was born in 1796 in Ghent and spent his career working across mathematics, astronomy, and statistics for the Belgian state. In 1832, he published research analyzing body measurements collected from European populations with a specific goal: characterizing the statistical "average man," the l'homme moyen, as a mathematical representative of normal human dimensions. The weight-to-height-squared ratio emerged from this work because it held relatively constant across the population he studied. As people grow taller, their weight tends to scale with height squared rather than height alone, which made the ratio a convenient index for comparing population-level statistics across cities or nations.
Quetelet was explicit about what the index was for. He was building a statistical model of society, the tool of a demographer, not a clinician. A population average tells you something meaningful about a group. It tells you considerably less about any specific person, whose weight and body composition are products of genetics, training history, diet, age, occupation, and dozens of other variables that a two-measurement ratio cannot capture. Quetelet's index was designed to smooth over individual variation; that is precisely what makes it useful as a statistical tool and problematic as a diagnostic one.
The Quetelet Index remained in the statistical literature for approximately 140 years without becoming a clinical standard. In the 1950s, the Metropolitan Life Insurance Company developed separate weight tables used by actuaries to price life policies, but these were built from mortality data and used different methodology. The path from 1832 to the modern examination room runs through a different decision, made in 1972.
How Ancel Keys Repurposed a 140-Year-Old Formula
In 1972, American physiologist Ancel Keys published a paper evaluating how well different height-weight indices correlated with directly measured body fat across a sample of 7,426 men drawn from Finland, Italy, Japan, South Africa, and the United States. Keys tested several formulas and found that the Quetelet Index correlated better with actual body fat than the alternatives he compared. He formally renamed it body mass index.
Keys was candid about the findings. He wrote that BMI accounted for "not more than half of the total variance of body fatness." He acknowledged that direct body density measurement, using hydrostatic weighing, which requires submerging a person in water and measuring displacement, was considerably more accurate. He chose BMI over better alternatives because of practical advantage: all you need is a scale and a measuring tape. For large-scale population research in an era before widespread clinical imaging, that practicality made BMI the workable choice.
The demographic composition of Keys' sample also matters for understanding what the derived ranges reflect. All 7,426 participants were men. Participants of Asian descent made up 13.9% of the sample; participants of African descent made up 1.56%. The sample was predominantly European and American in composition. The BMI ranges derived from this data were subsequently applied globally and to both sexes, which has produced documented inaccuracies. Research has established that Asian populations tend to show significantly higher body fat percentages at lower BMI values than the original sample predicted, meaning the standard healthy range systematically underestimates metabolic risk for a substantial portion of the world's population.
The Athlete Problem in Numbers
The mechanism that misclassifies athletes is straightforward physics. Muscle tissue is denser than fat tissue: a given volume of muscle weighs more than the same volume of fat. BMI measures weight divided by height squared with no way to differentiate what the weight is made of. A person who adds five kilograms of muscle through training will see their BMI rise by the same amount as someone who adds five kilograms of fat. The formula treats both outcomes identically.
This produces classifications that are clinically meaningless for people with substantial muscle mass. Consider a male rugby player at 1.85 meters tall and weighing 105 kilograms. His BMI is 30.7, placing him in the Class I obesity category, which begins at 30. Body composition analyses of similarly built athletes using DEXA scanning, which distinguishes fat mass from lean mass and bone mass, regularly show body fat percentages in the 10 to 15 percent range. By that direct measurement, the same player would be classified as lean or athletic. By BMI, he is obese.
The problem runs in the opposite direction as well. A sedentary individual who carries low muscle mass can maintain a BMI in the healthy range, 18.5 to 24.9, while accumulating visceral abdominal fat in amounts associated with insulin resistance and elevated cardiovascular risk. The clinical literature calls this normal-weight obesity. It is specifically the kind of metabolic risk profile that BMI cannot detect because BMI does not measure where fat is distributed. Research on BMI's limitations has found that abdominal fat distribution is a stronger driver of insulin resistance and cardiovascular disease than total body fat percentage alone, and BMI provides no information about fat distribution whatsoever.
A population-based cohort study published in PLOS ONE examining U.S. adults found that muscle mass was an independent predictor of mortality outcomes: participants with the highest muscle mass in their cohort showed significantly better survival rates even after controlling for their BMI category. Two people with identical BMI scores can have dramatically different health trajectories depending on what their weight is made of, which is information that BMI, by design, does not capture.
What More Accurate Measurement Looks Like
Researchers studying BMI's limitations have converged on waist-based measurements as the most practical improvement that does not require specialized clinical equipment. Waist-to-height ratio, your waist circumference divided by your total height measured in the same units, has demonstrated significantly stronger predictive power for cardiovascular disease risk in comparative studies. A literature review published in PMC in 2023 found that waist-to-height ratio showed "significantly greater discriminatory power" for cardiovascular disease outcomes than BMI across diverse populations. The generally recommended screening threshold is 0.5: a waist circumference less than half your height is associated with lower metabolic risk across both European and Asian populations.
Body fat percentage measured by DEXA scanning gives the most informative single number for understanding body composition. DEXA distinguishes fat mass, lean mass, and bone mass with precision, correctly identifying the muscular athlete with a high BMI as lean and flagging the normal-weight obese individual whose standard measurements appear healthy. Hydrostatic weighing and air displacement plethysmography provide similar accuracy through different physical principles. These measurements remain less accessible than a scale and tape measure, which explains why BMI persists in primary care settings despite its documented limitations.
For most people without access to clinical body composition analysis, combining waist circumference with BMI captures more of what BMI alone misses. The National Institutes of Health identifies a waist circumference above 40 inches in men and 35 inches in women as an independent cardiovascular risk factor regardless of BMI category. This paired approach catches a meaningful portion of the people that BMI alone misclassifies: the athlete who is lean despite a high BMI number, and the sedentary individual who is at metabolic risk despite a normal one.
Conclusion
Adolphe Quetelet built his index to answer a question about populations, not people. Ancel Keys repurposed it in 1972 because it was the best convenient proxy available at the time, while being clear about its accuracy limits. Neither man intended for it to become the primary clinical metric used to assess individual metabolic health across the entire global population.
Understanding where BMI comes from makes it easier to interpret what a BMI score actually tells you. It is a rough population-level screening tool that works reasonably well for identifying broad trends in large groups and poorly at characterizing any individual, especially one with more muscle than average. ToolHQ's BMI calculator gives you the standard number quickly, but it means most when read alongside other indicators: your waist circumference, your activity level, and, if accessible, a direct body composition measurement. The formula is 194 years old. It was never designed to be the final word.
Frequently Asked Questions
Who invented BMI and why?
Adolphe Quetelet, a Belgian mathematician and astronomer, created the formula in 1832 to statistically describe average population dimensions. It was a demographic tool, not a health assessment for individuals.
Why does BMI classify muscular athletes as overweight or obese?
Because BMI measures only weight relative to height. Muscle is denser than fat, so muscular people weigh more for their frame without excess fat, and BMI cannot distinguish between the two.
What is a healthy BMI range?
The standard ranges are: underweight below 18.5, healthy 18.5 to 24.9, overweight 25 to 29.9, and obese at 30 and above. These thresholds were derived primarily from a 1972 study of European and American men.
What should I use instead of BMI?
Waist-to-height ratio (waist circumference divided by height, with a healthy threshold below 0.5) and waist circumference alone have both shown stronger predictive power for cardiovascular risk than BMI.
Is BMI accurate for women?
Ancel Keys' original 1972 BMI study used only male subjects. Women tend to carry proportionally more body fat at the same BMI than men, making the standard ranges less precise as a health indicator for women.