Body Fat Calculator
Estimate your body-fat percentage from simple tape measurements using the US Navy circumference method — no calipers or scales required. Choose metric or imperial units; the hip measurement appears when the female formula needs it.
Estimate body fat
Example: male, 175 cm, neck 37, waist 85 → 17.7%.
How the Navy method works
The circumference method was developed by Hodgdon and Beckett at the Naval Health Research Center in 1984, and it is the equation this page implements. It works in two steps. The first predicts body density from height together with the neck and waist girths — plus the hip girth in the equation fitted for women. The second converts that predicted density into a percentage of body mass. The expression this page implements for men — 495 ÷ (1.0324 − 0.19077·log₁₀(waist − neck) + 0.15456·log₁₀(height)) − 450 — is those two steps composed into one line, which is why the arithmetic looks stranger than the idea behind it. The measurements in the example above give 17.7%, computed at build time by the same tested engine the calculator runs in your browser.
What the equation senses is the shape described by a few tape readings, not fat itself. Two people with identical height and identical girths receive identical estimates, whatever else differs between them. That is the trade the method makes: a tape measure in place of a laboratory, at the cost of everything the tape cannot see.
Two equations, one method
The male and female equations use different inputs and entirely different coefficients, so they are not variations on one formula. The male equation works from waist − neck; the female equation works from waist + hip − neck, which is why the hip field appears only when it is needed. A worked female example — 165 cm tall, neck 32 cm, waist 74 cm, hip 96 cm — gives 27.4%. Reading that figure against a male figure compares two different equations rather than two bodies, so a figure from one is not on the same scale as a figure from the other.
Entering the measurements
The unit radio swaps in a different set of fields rather than converting what is already typed, so an entry made in one mode does not carry across to the other. In imperial mode the girths are read in inches and converted before the equation sees them, at 2.54 cm to the inch, and height is entered as feet plus inches with an empty inches box counted as zero rather than treated as an error. The hip field stays hidden until the female equation is selected, because the male equation has no term that would use it.
The estimate recalculates on every keystroke. A field left blank leaves the result box waiting rather than showing an error, and a value the engine will not accept is explained underneath the field it came from rather than silently dropped. The result box prints the percentage to one decimal place and names the method that produced it. Nothing is transmitted anywhere — the arithmetic runs in the page, and the worked figures in this text are produced by the same functions at build time, so the copy and the code cannot drift apart.
What the error band means in practice
Against reference methods such as underwater weighing, the standard error of the circumference method is commonly reported at roughly 3–4 percentage points. That figure, not the decimal place, is the right scale at which to read a result. It is the reason the tool is more useful for comparing readings taken the same way than for settling any one of them: a single estimate is consistent with a true value several points either side of it. The single decimal place is a display choice, not a claim that the tenth of a percent means anything.
Where the tape sits changes the answer
That error band belongs to the method. Measurement error sits on top of it, and the size of it is easy to show. Take the worked example again — 175 cm, neck 37 cm, waist 85 cm, 17.7%. Enter the waist as 87 cm instead of 85 — a two-centimetre difference in what the tape reads — and the estimate becomes 19.2%. Put the waist back and enter the neck one centimetre larger instead, and it falls to 16.9%. Nothing about the body changed in either case; only the number written down did. Tape position and tension alone can move a reading by a point or more.
The technique that keeps readings comparable is unglamorous: the same tape, held the same way, at the same landmarks, under the same conditions every time. Repeatability matters more than picking the anatomically perfect landmark, because a landmark you can find again cancels out of the trend.
Track trends, not single readings
A single reading carries the method's full error band on top of whatever the tape contributed. Measuring the same way each time — same tape, same landmarks, same tension — removes the part of the variation that has nothing to do with the body, which is why the sequence of readings, rather than any one of them, is the part that carries information. For context beyond one number, the BMI Calculator gives the screening ratio that population guidelines are written around, and what BMI does and does not measure explains why a circumference estimate and a BMI figure can describe the same person differently without either being wrong.
Why the calculator refuses some measurements
Three checks stand between the tape and a number. Circumferences have to be physically plausible: neck 10–100 cm, waist and hip 30–300 cm, with height 50–280 cm. The waist has to exceed the neck for the male equation — that difference goes inside a logarithm, so a waist at or below the neck has no defined value at all rather than a merely wrong one — and waist plus hip has to exceed the neck for the female equation. Finally, any measurement set implying a body fat outside 1–75% is refused, because that is the point at which the regression has been pushed past anything it can speak to. In each case the engine returns the reason instead of a plausible-looking answer, which is the whole point of validating input rather than rounding it into range.
Frequently asked questions
How do I take the measurements?
The engine cannot see how the tape was held, so it treats every reading as exact. Use a flexible, non-stretch tape, keep it snug without compressing the skin, and — most importantly — use the same landmarks and the same conditions every time. A landmark you can find again reliably matters more than the anatomically perfect one, because the tape reading is the only thing the equation responds to.
How accurate is the Navy method?
It is a regression fitted to a population, so it produces an estimate rather than a measurement. The standard error it carries against reference methods such as underwater weighing is set out in full on this page, under "What the error band means in practice"; the practical reading of that figure is that the method is dependable for comparing readings taken the same way and not dependable for settling any single one. The decimal place shown is a display convention, not a claim about tenths of a percent.
Why does this page not classify the result into a category?
Because the engine returns a percentage, not a category. The method’s own error is wide enough that one reading can fall on either side of any boundary without anything about the body having changed, so the direction of a trend across consistently taken measurements carries more information than the label a single number would receive.
Does switching to imperial convert what I have already typed?
No. The unit radio swaps in a different set of fields rather than converting the entries in the fields you were using, so a measurement taken in one mode has to be re-entered in the other. Inches are converted to centimetres before the equation sees them, so both modes reach the same tested function.
Does the method measure body fat directly?
No. It predicts body density from height and the girth measurements, then converts that predicted density into a percentage. It responds only to the tape readings you enter, so two people with identical measurements receive identical estimates regardless of anything the tape cannot see.
Why does the female formula need a hip measurement?
The two equations are different: the male equation works from waist − neck; the female equation uses waist + hip − neck, so the hip field appears only when that equation is selected.
Why were my measurements rejected?
The engine refuses input it cannot answer honestly: circumferences outside the accepted ranges (neck 10–100 cm, waist and hip 30–300 cm), a waist at or below the neck for the male equation, and any measurement set implying a body fat outside 1–75%. Check for a unit mix-up — inches typed into a field labeled cm — or a dropped digit.
Not medical advice: this calculator provides general educational estimates from a published population formula. It cannot account for your individual health context and must not be used to diagnose or treat any condition — consult a qualified healthcare professional. Measurements are processed locally in your browser and never transmitted. See the methodology page.