Measure Body Fat

Measuring body fat: the two home methods that work

You cannot measure body fat directly at home. What you can do is take a small number of body measurements and run them through a prediction equation, and there are only two routes worth the effort without a laboratory: circumferences with a tape and skinfolds with a caliper. Both are cheap, both are repeatable, and both carry an error of roughly one percentage point in untrained hands — which is why the second half of this page matters more than the first.

The full landmark-by-landmark protocol — including the mistakes that shift the answer by two points — lives on our fat calculator and how to measure body fat at home pages. What follows here is the part no reference chart gives you: what to do with the number once you have it.

Computed: the age-and-sex corridor

Almost every body fat chart you will find is age-blind. It gives one set of ranges for men and one for women and stops there, which means a 25-year-old and a 70-year-old are being judged against the same line. This page builds a corridor instead: take the Deurenberg equation, evaluate it at the two ends of the World Health Organization's normal BMI range (18.5 and 24.9), and you get the span of body fat percentages consistent with a normal BMI at that age and sex.

Three properties of this table are exact, not approximate, and they follow directly from the equation's shape. First, the corridor is 7.68 points wide at every age for both sexes: 1.20 × (24.9 − 18.5) = 7.68, and neither age nor sex appears in that calculation. Second, it climbs 2.30 points per decade— 0.23 per year, the equation's own age coefficient — at a constant rate with no acceleration. Third, the gap between the women's and men's corridor is exactly 10.80 points at every age, because the sex term is a constant −10.8 and everything else cancels.

That third property is worth dwelling on, because it is the cleanest statement of the sex difference available: in this equation, being female adds 10.8 points of body fat at every age and every BMI, no more and no less. The equation has no room for the sex difference to narrow or widen with age, which is a modelling simplification rather than a fact about bodies.

Computed: where the age-blind chart and the corridor disagree

Set the corridor next to the widely circulated American Council on Exercise category chart — Essential 2–5% / Athlete 6–13% / Fitness 14–17% / Acceptable 18–24% / Obese 25%+ for men, and 10–13 / 14–20 / 21–24 / 25–31 / 32+ for women — and the two frames come apart as you age. Here is the men's comparison.

The women's version is starker. A 45-year-old woman with a completely normal BMI of 24.9 computes to 34.8% body fat, which the age-blind chart calls obese. She is not obese by any BMI criterion — she is at the top of a normal weight range, and two different frameworks disagree about her entirely.

Invert the calculation and you get the exact ages at which the two frameworks cross. These are solved algebraically from the equation, not read off a chart:

Read that table honestly and it indicts both frameworks rather than either one. An age-blind chart that classifies a normal-weight 65-year-old woman as obese is not describing her body; it is describing a population she was never sampled from. But a corridor built on a BMI equation inherits every weakness of BMI, and the Deurenberg equation is known to behave differently outside the population it was fitted on. The useful conclusion is narrow and defensible: compare yourself to an age-specific band, and treat every fixed cut-off with suspicion.

Computed: what happens if you never change

Because the corridor rises 2.3 points per decade and your body does not have to follow it, the same body fat percentage means different things at different ages. Here are two men holding their numbers constant from thirty to seventy.

The man at 21.0% starts marginally above the corridor, spends three decades comfortably inside it, and falls out of the bottom by seventy without losing a gram. The man at 15.0% — an impressive number at thirty — is out of the corridor by forty. Neither of them changed; the reference moved.

The geometry behind all of this is simple enough to state in one line: the corridor is 7.68 points wide and drifts 2.30 points per decade, so holding your body fat perfectly constant walks you from the ceiling to the floor in 3.34 decades — about 33 years. A man at the top of his corridor at 25 falls out of the bottom at 58.4 if he never changes anything.

Note the direction this cuts. Falling below the corridor as you age is not automatically a problem — it means you are leaner than a normal-BMI person of your age typically is, which is usually the result of carrying more muscle than average. The corridor is a description of a reference population, not a target to stay inside.

Computed: the same question asked backwards

Instead of asking whether your number fits your age, ask which ages your number fits. Solving the corridor for age gives a window: below the lower age the corridor is still beneath you, above the upper age it has climbed past you.

Every window is 33.4 years wide — the same 3.34 decades as before, because the corridor's width and drift are both constants. The windows differ only in where they sit. This is the most useful way to read a single body fat measurement: not “am I good or bad” but “this is the age range my number is typical for.”

Computed: how much fat that actually is

Percentage points are abstract. Converting them to kilograms is arithmetic: at body weight W, one percentage point of body fat is W ÷ 100 kilograms, provided lean mass holds still. Which means the cost of moving through the corridor scales with how much you weigh.

Two worked cases make this concrete. An 80 kg man at 21.0% who wants to reach the floor of the 45-year-old corridor (16.35%) needs 3.72 kg of fat lost with lean mass intact. A 65 kg woman at 30.0% reaching for the 45-year-old floor (27.15%) needs 1.85 kg. Both are ordinary amounts of fat — a few months of work, not a transformation — which is a useful piece of perspective next to the two-to-three-point measurement noise described below.

The lean-mass caveat is doing real work in that arithmetic, though. If you lose 4 kg and 1 kg of it is lean, the percentage moves less than the table predicts, and at a body fat percentage of around 20 the difference is worth roughly half a point. Our body recomposition calculator models the fat and lean split explicitly if you want that level of detail.

Computed: is your measurement precise enough to place you?

Knowing whether you are inside or outside a band is only meaningful if your measurement error is smaller than the distance to the edge. Here is that comparison, using the two worked profiles — a 178 cm, 80 kg man with a 92 cm waist and 39 cm neck, and a 165 cm, 65 kg woman with a 78 cm waist, 98 cm hip and 32 cm neck — with three readings averaged per site.

The good news: the 7.68-point corridor is between 4.5 and 5.6 noise-widths wide, so deciding whether you are inside it or outside it is a decision a tape measure can actually make. The bad news is what happens with narrower bands. The ACE “fitness” band is four points wide — 14–17% for men, 21–24% for women — which resolves to only 2.5 to 2.9 separate positions. You can tell whether you are below it, in it, or above it, and not much more.

And any single fixed cut-off is worse than a band, because there is no width to absorb error at all. Standing one point above the 25% line with a ±0.58 point measurement is not a finding; it is a coin toss. This is the practical argument for thinking in corridors rather than thresholds: a band gives your measurement error somewhere to live.

How to run this at home

• Measure the same way every time, and write down the raw numbers. The corridor question is answered by your measurements being consistent, not by being right.
• Three readings per site, averaged. It cuts the noise by roughly 40% and costs two minutes. Going to five buys another 22% and is rarely worth it.
• Same time of day, same state. Morning, after the bathroom, before food and training. This removes the largest single source of within-person variation that has nothing to do with fat.
• Re-measure monthly at the most often. The detectable-change numbers above put the smallest trustworthy difference at 1.4 to 1.7 points, and a month of genuine fat loss is often under one point.
• Track the waist in centimetres alongside the percentage. It carries no prediction error at all, and it moves in the direction you care about.
• Update the corridor when you have a birthday, not every session. It climbs 2.3 points per decade, so it takes years to matter and a year to notice.
• Cross-check against an independent method occasionally. A photo-based estimate or a reference chart will not settle the truth, but a method that disagrees with everything else is a method you are using wrong.

Where this is weak

• The corridor is this page's construction, not a standard. It is one published equation evaluated at two published BMI cut-offs. No organisation endorses it, and it should not be used as a clinical threshold.
• It inherits every weakness of BMI.BMI cannot distinguish muscle from fat, so a muscular person's BMI-derived corridor sits too high for them. The corridor is built on a normal-BMI population and describes that population, not you.
• The age coefficient is a single constant. The equation applies +0.23 points per year uniformly from 18 to 90. Real body composition change with age is not uniform, and this equation has no way to represent that.
• The sex difference is fixed at 10.80 points.That is a property of the equation's structure, not a measured fact that holds across all ages and body types.
• None of the error figures include the equations' own error. The ± bands cover measurement error only. How far these equations sit from a laboratory reference method for your body is a separate and usually larger problem that this page does not compute.
• The slip size you select is a guess about yourself. Every noise figure is conditional on it. If you are worse than you think, every confidence statement above is optimistic.
• Nothing here is a health assessment. A body fat percentage is one number with a wide error band, and a corridor is a description of a reference population. Neither is a diagnosis.

Frequently asked questions

Related tools

fat calculator with an error budget · body fat calculator · Navy body fat calculator · how to measure body fat at home · body fat percentage chart · body fat with a cross-method error bar · BMI and smart scales

Not medical advice.The corridor on this page is this site's own construction from a published equation and two published BMI cut-offs; it is not an official standard. All error figures cover measurement error only and are conditional on the slip size you selected. See our disclaimer.