What the circumference method actually estimates
The Navy method does not measure fat. It predicts body density from a small number of tape measurements, and then converts that density into a fat percentage using a two-compartment model. Understanding that two-step structure explains both why the method works and where it fails.
Step one is a regression. Hodgdon and Beckett, working at the Naval Health Research Center in 1984, measured hundreds of Navy personnel by hydrostatic weighing and looked for the combination of tape measurements that best predicted the result. For men, abdominal girth minus neck girth, scaled logarithmically against height, did most of the work: the abdomen is where men store the fat that varies, and the neck is a site that barely changes with adiposity, so subtracting it removes a large part of the frame-size effect. For women, hip circumference had to be added, because women store a much larger share of fat gluteofemorally.
Step two is the Siri equation, %BF = 495/D − 450. It assumes the body is exactly two compartments: fat at a density of 0.900 g/cm³ and fat-free mass at 1.100 g/cm³. Plug those in and you can see where the constants come from — at D = 1.100 the formula returns 0% fat, and at D = 0.900 it returns 100%.
So a Navy body fat result is a prediction of a prediction. That is not a reason to distrust it, but it is the reason it should be read as a tracking number rather than a measurement.
Reading the equation term by term
Take the male form: D = 1.0324 − 0.19077·log10(waist − neck) + 0.15456·log10(height).
The circumference term carries a minus sign, so a larger waist-minus-neck value lowers predicted density, and lower density means more fat. That is the whole mechanism. Because the term is logarithmic, a fixed number of centimetres matters more to a small person than a large one: going from 40 to 45 cm changes log10 by 0.051, while going from 60 to 65 cm changes it by only 0.035.
The height term carries a plus sign, so at the same waist-minus-neck value, a taller person is predicted to have higher density and less fat. This is the frame correction. Without it, every tall person would read as lean and every short person as fat, because girth scales with size.
The neck is doing more work than it looks like. It is subtracted rather than divided, which means it acts as a per-person offset for frame and for the non-fat contribution to abdominal girth. It is also the measurement people most often get wrong — a tape sitting too low picks up the trapezius and shoulders, inflates the neck value, shrinks the circumference term, and returns an implausibly low body fat.
The female equation swaps sites, not structure. Adding hip to waist before subtracting neck captures both major storage depots, and the steeper coefficient (0.35004 against 0.19077) reflects the wider range that sum takes.
Worked example: a man 178 cm tall, 84 kg, neck 39 cm, waist 90 cm
- Circumference value. 90 − 39 = 51 cm.
- Take the logarithms. log10(51) = 1.70757 and log10(178) = 2.25042.
- Circumference term. 0.19077 × 1.70757 = 0.32575.
- Height term. 0.15456 × 2.25042 = 0.34782.
- Predicted density. D = 1.0324 − 0.32575 + 0.34782 = 1.05447 g/cm³.
- Siri conversion. 495 ÷ 1.05447 = 469.43; 469.43 − 450 = 19.43% body fat.
- Fat and lean mass. 84 × 0.1943 = 16.32 kg of fat, leaving 84 − 16.32 = 67.68 kg of lean mass.
Now test the sensitivity. Add one centimetre to the waist, so the circumference value becomes 52 cm: log10(52) = 1.71600, the circumference term becomes 0.32736, D falls to 1.05286, and body fat rises to 495 ÷ 1.05286 − 450 = 20.15%. One centimetre of tape moved the answer by 20.15 − 19.43 = 0.72 percentage points, which is why the measurement protocol matters more than the arithmetic.
That 67.68 kg of lean mass is also the input the BMR calculator needs if you want to use the Katch-McArdle equation rather than Mifflin-St Jeor, and it is the number the protein intake calculator uses when you dose protein on lean mass instead of total weight.
How to interpret your body fat percentage
Read the category, not the decimal. The circumference method's error against hydrostatic weighing is large enough that a single reading should be treated as a band rather than a point, so 19.4% and 21.0% are the same answer.
What the number is genuinely good for is tracking. Measure under identical conditions — same time of day, same tape tension, same sites, before eating — and the errors largely cancel, so the change between two readings is far more reliable than either reading alone. A fall of two percentage points over eight weeks is real information; a fall of 0.3 between two mornings is noise.
Cross-check the result against your waist-to-height ratio, which this calculator prints from the same measurements. The two should tell a consistent story. If your body fat comes out at 12% while your waist exceeds half your height, one of the measurements is wrong — almost always the neck. If you want a third opinion from a completely different model, run the same height and weight through the BMI calculator: a low body fat alongside a BMI in the overweight band is the expected signature of a muscular build, whereas both reading high is consistent.
If you are measuring for a military or first-responder assessment, note that each service sets its own maximum allowable body fat in its own instruction, that those limits have been revised more than once, and that the official assessment uses a specified tape, a specified measurer and a specified protocol. Use this calculator to know roughly where you stand, not to predict a pass.
Published body fat categories by sex
| Category | Men | Women |
|---|---|---|
| Essential fat | 2–5% | 10–13% |
| Athletes | 6–13% | 14–20% |
| Fitness | 14–17% | 21–24% |
| Average | 18–24% | 25–31% |
| Obese | 25% and above | 32% and above |
Women carry roughly 10 percentage points more essential fat than men, which is why the two columns are offset rather than identical.
Measurement mistakes that change the answer
- Measuring the neck too low. The tape belongs just below the larynx, sloping slightly downward at the front. Drop it onto the shoulders and you add two or three centimetres, which can subtract several points of body fat from the result.
- Pulling the tape tight. It should sit flush against the skin without compressing it. Tension is the largest source of difference between two people measuring the same body.
- Measuring the waist in the wrong place. The male equation was fitted at the navel and the female equation at the narrowest point of the waist. Swapping the sites biases the result in a direction you cannot correct for.
- Holding your breath or sucking in. Measure at the end of a normal, relaxed exhalation. Being consistent about this matters more than which point of the breath you pick.
- Mixing units. Entering a neck in inches and a waist in centimetres produces a plausible-looking number that is entirely wrong. This calculator converts everything to centimetres, but only if you set each unit correctly.
- Comparing readings taken by different people. Tracking works because the errors are consistent. A new measurer resets the baseline.
Why very lean and very heavy people read badly
The equation was fitted on a military population, so it is at its best across the middle of the range and degrades at both ends. In very lean people the abdominal circumference stops responding to further fat loss, so the estimate flattens out and can read higher than reality. In people carrying a great deal of fat, the two-compartment Siri model breaks down because hydration of the fat-free mass shifts, so the estimate tends to run low. If your result sits below the essential-fat line for your sex, treat it as a measurement error until you have re-measured the neck.
How this compares with other body composition methods
DEXA is the practical reference standard. It separates fat, lean soft tissue and bone mineral, and reports them regionally, so it can tell you that you gained a kilogram of lean mass in the legs. It costs a scan appointment and a small radiation dose.
Hydrostatic weighing is what the Navy equations were validated against. It measures whole-body density directly by underwater weighing and then applies the same Siri conversion this calculator applies — so it shares the two-compartment assumption, and differs only in that the density is measured rather than predicted.
Skinfold calipers in trained hands are more accurate than tape circumferences, because subcutaneous fat is being sampled directly. In untrained hands they are worse, because site location and pinch technique take practice.
Bioelectrical impedance scales are the most convenient and the most affected by hydration. A morning reading and an evening reading on the same scale can differ by several percentage points with no change in body composition at all.
The circumference method's advantage over all of them is that it costs a tape measure and repeats identically anywhere. For tracking a trend, that consistency is worth more than a one-off accurate reading — which is exactly why a service with hundreds of thousands of people to assess chose it. If you want to convert the lean mass it gives you into a training and nutrition plan, feed it into the TDEE calculator and the macro split calculator.