Why there is no single ideal weight
Where the four classic formulas came from, why they disagree by several kilograms, and what to use instead when the answer actually matters.
Last reviewed · 1,377 words
In short
- The four common formulas disagree by around 3 kg at the same height. None of them is more correct than the others.
- All of them were built for drug dosing and insurance tables, not for personal health targets.
- They use height and sex and nothing else — no frame size, no muscle mass, no age.
- A healthy weight is a range, not a point. For most heights that range spans 15 kg or more.
- Waist-to-height ratio and body fat percentage answer the question these formulas are usually asked to answer.
There is no ideal weight. There is a range of weights compatible with good health at a given height, and it is wide.
The formulas on this page produce single numbers because they were designed for purposes that required single numbers — calculating drug doses, setting ventilator volumes, filling in insurance tables. None was designed to tell an individual what to weigh.
The four formulas
All of them work the same way: a base weight at five feet, plus a fixed amount per inch above it.
| Formula | Year | Men | Women |
|---|---|---|---|
| Hamwi | 1964 | 48.0 kg + 2.7 kg/inch | 45.5 kg + 2.2 kg/inch |
| Devine | 1974 | 50.0 kg + 2.3 kg/inch | 45.5 kg + 2.3 kg/inch |
| Robinson | 1983 | 52.0 kg + 1.9 kg/inch | 49.0 kg + 1.7 kg/inch |
| Miller | 1983 | 56.2 kg + 1.41 kg/inch | 53.1 kg + 1.36 kg/inch |
At 175 cm for a man:
| Formula | Ideal weight |
|---|---|
| Hamwi | 72.0 kg |
| Devine | 70.5 kg |
| Robinson | 68.9 kg |
| Miller | 68.7 kg |
At 162 cm for a woman:
| Formula | Ideal weight |
|---|---|
| Miller | 58.2 kg |
| Robinson | 55.4 kg |
| Devine | 54.2 kg |
| Hamwi | 53.8 kg |
Three and a half kilograms separate the extremes in each case, and the ordering even reverses between the sexes — Hamwi gives the highest figure for men and the lowest for women. That is not a subtle disagreement about biology; it is four different regressions fitted to four different datasets for four different purposes.
Where they came from
Devine, 1974, was published to standardise dosing of gentamicin. Devine described it as a starting point requiring clinical judgement. It remains the most-used formula in medicine, and it is still doing the job it was written for.
Hamwi, 1964, was a rule of thumb for diabetes dietitians estimating calorie needs. It was never validated against outcomes.
Robinson and Miller, both 1983, were attempts to fit the Metropolitan Life insurance tables more closely. Those tables came from policyholders — a self-selected group that was wealthier, more urban and more likely to be white than the general population — and the height measurements were often self-reported with shoes on.
None of these origins involves asking what weight is best for a person's health. They are useful history rather than personal guidance.
The 5-foot problem
Every formula uses inches above five feet as its variable, which creates two failures at the edges.
Below five feet, the formulas either return the base weight or go negative. Someone 145 cm tall gets whatever the base is, regardless of how much shorter than five feet they are.
Well above it, the linear extrapolation drifts. Bodies scale in three dimensions and these formulas add a fixed weight per inch, so they progressively understate tall people — the same structural problem BMI has in the opposite direction.
They are most reliable in the middle of the range they were fitted to, roughly 155 cm to 185 cm, and least reliable exactly where a person is most likely to wonder whether the number applies to them.
What they cannot see
Frame size. A person with wide shoulders and thick wrists carries more bone and connective tissue than a slight person of the same height. The classic wrist-circumference method estimates frame, and no formula here uses it.
Muscle mass. A trained 175 cm man can weigh 80 kg at 12% body fat and be in excellent health. Every formula above says he is 10 kg overweight.
Age. Body composition shifts across decades. Slightly higher weights are associated with better outcomes in older adults, and no formula has an age input.
Ethnicity. All four were derived from Western populations. South Asian bodies carry more visceral fat at the same weight and height, which is why Indian BMI thresholds are lower — and which none of these formulas reflect.
Use a range instead
The more honest answer to "what should I weigh" is a range, derived from BMI at your height.
At 175 cm, the WHO range of 18.5 to 24.9 spans 56.7 kg to 76.3 kg — nearly 20 kg. Under the ICMR range of 18.5 to 22.9 it is 56.7 kg to 70.1 kg.
At 162 cm, WHO gives 48.6 kg to 65.3 kg; ICMR gives 48.6 kg to 60.1 kg.
Where you sit inside that range depends on how much muscle you carry, and the range is wide because health is genuinely compatible with a wide span of weights. A single number invites people to treat a two-kilogram difference as meaningful when it is not.
What to measure instead
Waist-to-height ratio. Keep your waist under half your height. It needs no tables, works across ages and ethnicities, and predicts cardiometabolic risk better than weight does.
Body fat percentage. This is what "ideal weight" is really trying to approximate. A tape measurement using the Navy method lands within about 3% of a DEXA scan.
Waist circumference. Risk rises above 90 cm for Indian men and 80 cm for Indian women — lower than the international thresholds.
How you actually function. Sleep quality, energy through the day, blood pressure, fasting glucose and lipids all describe health more directly than any weight target, and all of them respond to changes long before a scale reaches a particular number.
Where the formulas genuinely belong
They are not useless — they are simply misapplied.
Drug dosing. Many medications are dosed on ideal or adjusted body weight rather than actual weight, because they distribute into lean tissue rather than fat. Devine is standard here.
Ventilator settings. Tidal volume is set from predicted body weight, since lung size scales with height and not with fat mass. Using actual weight in an obese patient causes real harm, which is why this calculation is done carefully in intensive care.
Population-level nutrition planning, where an average is what is wanted.
In each case a single defensible number is required and precision about an individual is not the point. That is the opposite of what a person asking about their own weight needs.
Frame size, the input everybody wants
The usual objection to these formulas is that they ignore build, and the objection is correct. The classic workaround is wrist circumference relative to height.
Measure the wrist at its narrowest point, just below the bone on the hand side. Divide height in centimetres by wrist circumference in centimetres.
| Ratio, men | Ratio, women | Frame |
|---|---|---|
| Above 10.4 | Above 11.0 | Small |
| 9.6 – 10.4 | 10.1 – 11.0 | Medium |
| Below 9.6 | Below 10.1 | Large |
The convention is to subtract about 10% from the formula result for a small frame and add about 10% for a large one. At 175 cm that is roughly 7 kg either way, which is a larger adjustment than the disagreement between all four formulas combined.
Wrist circumference is a reasonable proxy because it is almost entirely bone and tendon — it changes very little with weight, unlike almost every other measurement on the body.
It is still an approximation, and it is not part of any validated formula. Treat it as a reason to widen the range you consider acceptable rather than as a new single number to aim at.
What to do with the result
If you came here to find a target, the useful output is the healthy range, not the four point estimates.
Compare your current weight to that range and to your waist measurement. If you are inside the range and your waist is under half your height, there is nothing to act on. If you are above it, a 5% reduction improves blood pressure, fasting glucose and lipids measurably, well before any threshold is crossed.
And if you train with weights, expect every formula on this page to be wrong about you in the same direction. Measure body fat instead and use lean mass as the reference.
What this calculator assumes
- Four historical formulas, reported side by side rather than averaged, because averaging them would imply an agreement that does not exist.
- Height in centimetres, converted at 2.54 cm per inch, with five feet as the base.
- Results are least reliable below 155 cm and above 185 cm.
- The BMI-based healthy range shown alongside is the more useful output for most purposes.
- None of this is a clinical target. For a weight goal that matters medically, the person to ask has your blood work in front of them.