Most Accurate Body Fat Method: U.S. Navy vs Skinfold vs BMI-Based Estimates

Which is the most accurate body fat method? Take one 30-year-old man: 180 cm tall, 78 kg, a 38 cm neck, an 85 cm waist, and a combined chest-abdomen-thigh skinfold thickness of 45 millimetres. Run those same measurements through three different, commonly used body fat estimation methods, and you get three different answers: 16.1% from the U.S. Navy tape method, 13.6% from Jackson & Pollock skinfold calipers, and 19.6% from the BMI-based Deurenberg formula. Three different numbers for one unchanged body, spanning three different category labels on the standard American Council on Exercise chart: Athletes, Fitness, and Average.

None of the three methods made a mistake. Each one is estimating the same underlying quantity, body fat percentage, from a completely different physical signal, using a different regression equation built from a different reference population. Understanding what each method actually measures, and where each one's particular blind spot sits, is what turns "which number do I believe" into an answerable question instead of a shrug, and it is the only honest way to answer "what is the most accurate body fat method" without pretending one formula is universally right and the other two are simply wrong.

What each method is actually measuring

The U.S. Navy method reads neck and waist circumference for men, adding hip circumference for women, along with height, and runs them through a formula built from regression analysis against measured body density in a large military sample. It assumes that a thicker waist relative to neck and height generally indicates more body fat, an assumption that holds reasonably well across most people and breaks down for anyone with an unusually muscular neck or midsection. This is the exact method behind any navy body fat calculator or body fat calculator U.S. Navy label you will find online, since the U.S. military still uses this same tape-measure formula for its own service members today.

Skinfold calipers work differently. Rather than reading an external circumference, a tester pinches a fold of skin and the fat directly underneath it at several specific anatomical sites, then reads the fold's thickness in millimetres off a caliper. The Jackson & Pollock 3-site protocol used in this comparison measures chest, abdomen, and thigh for men (different sites are used for women), sums the three readings, and feeds that sum, plus age, into a formula that estimates whole-body density:

Body density = 1.10938 − 0.0008267 × sum − 0.0000016 × sum² − 0.0002574 × age for men, where "sum" is the total of the three skinfold measurements in millimetres. That density is then converted to a body fat percentage using the Siri equation, the same 495 ÷ density − 450 conversion the Navy method's own formula is built around internally.

The BMI-based method, sometimes called the Deurenberg formula after the researchers who published it in 1991, skips physical measurement of the body entirely. It takes only BMI, age, and sex:

Body fat % = 1.20 × BMI + 0.23 × age − 10.8 × genderTerm − 5.4 for adults, where genderTerm is 1 for men and 0 for women. No tape, no calipers, just a height, a weight, a birth year, and a checkbox for sex.

The worked comparison, step by step

Here is the same 30-year-old man run through all three formulas, with every input held constant except the one each method actually needs.

Navy method. Waist minus neck is 85 − 38 = 47 cm. Taking log10 of 47 and log10 of 180 (his height in centimetres) and applying the male coefficients gives a denominator of about 1.0619888. Dividing 495 by that and subtracting 450 gives 16.106606, which rounds to 16.1%. On the American Council on Exercise chart, that lands in the Fitness band (14 to 17% for men).

Skinfold method. His chest, abdomen, and thigh skinfolds sum to 45 mm. Plugging that sum and his age of 30 into the Jackson & Pollock body density formula: 1.10938 − 0.0008267 × 45 − 0.0000016 × 45² − 0.0002574 × 30 works out to a body density of 1.0676965. Running that through the Siri equation, 495 ÷ 1.0676965 − 450, gives 13.614894, which rounds to 13.6%. On the same ACE chart, the Athletes band for men runs from 6% up to just under 14%, so 13.6% sits inside Athletes, one full category leaner than the Navy method's reading of the same person.

BMI-based method. His BMI is 78 ÷ 1.8² = 78 ÷ 3.24 = 24.07. Plugging that into the Deurenberg adult male formula: 1.20 × 24.07 + 0.23 × 30 − 10.8 × 1 − 5.4 = 28.89 + 6.9 − 10.8 − 5.4 = 19.59%, which rounds to 19.6%. That lands in the Average band (18 to 24% for men), one category higher than the Navy method and two categories higher than the skinfold reading.

A 6-percentage-point spread, and three different labels, for one person who did not change at all between measurements. That spread is the entire reason a comparison like this is worth doing before trusting any single number.

A second comparison: the same three methods for a woman

The male example above is not a one-off pattern specific to men. Take a 28-year-old woman: 165 cm tall, 62 kg, a 32 cm neck, a 75 cm waist, a 95 cm hip, and a combined triceps-suprailiac-thigh skinfold sum of 55 mm, the three-site protocol typically used for women rather than the chest-abdomen-thigh sites used for men.

Navy method. Waist plus hip minus neck is 75 + 95 − 32 = 138 cm. Running that and her height of 165 cm through the female coefficients gives a denominator of about 1.0368107, and 495 divided by that minus 450 comes out to 27.425639, rounding to 27.4%. That lands in the Average band (25 to 31% for women).

Skinfold method. Her body density, from the women's Jackson & Pollock formula using a skinfold sum of 55 and an age of 28, works out to 1.0479425. The Siri equation converts that to 495 ÷ 1.0479425 − 450 = 22.354161, rounding to 22.4%. That lands in the Fitness band (21 to 24% for women), a full category leaner than her Navy result.

BMI-based method. Her BMI is 62 ÷ 1.65² = 62 ÷ 2.7225 = 22.77. The Deurenberg adult female formula (genderTerm 0) gives 1.20 × 22.77 + 0.23 × 28 − 5.4 = 27.32 + 6.44 − 5.4 = 28.37%, rounding to 28.4%, which also lands in the Average band, close to her Navy result rather than her skinfold result.

This second example shows something the first one did not: the disagreement does not always point the same direction. For the man above, skinfold read leanest and the BMI-based method read highest, with Navy in between. For this woman, skinfold is again the outlier, but this time the Navy method and the BMI-based method land close together instead of at opposite ends. There is no fixed rule for which method will read highest for a given person; it depends on how that individual's actual fat distribution compares to whatever the specific formula assumes as typical.

Why the spread happens: what each method cannot see

The skinfold method comes out leanest here because it is the only one of the three that samples subcutaneous fat directly, at specific sites, rather than inferring fat from an external circumference or from a weight-to-height ratio. If this man happens to carry relatively little fat specifically at his chest, abdomen, and thigh sites, compared to how his neck-to-waist ratio or his overall BMI would suggest, skinfold will report a leaner number than the other two.

The Navy method sits in the middle because it only sees two circumferences and height. It has no way to tell whether a wider waist relative to neck reflects fat or muscle, or whether this particular person's fat distribution matches the regression sample the formula was built from.

The BMI-based method reads the highest here for a structural reason: it has the least information of the three. It never touches the body at all, so it cannot see this specific man's actual fat distribution, only his overall weight relative to height, adjusted by a general age trend. For someone whose actual body composition sits leaner than what their BMI alone would predict, which describes plenty of moderately active adults, the BMI-based method will tend to overstate body fat, exactly the pattern in this example.

So which is the most accurate body fat method for you?

For a single individual, in order of how directly each one measures the actual thing you are trying to estimate: skinfold, done correctly by a consistent tester, generally edges out the Navy method, which in turn is more informative than the BMI-based method for anyone whose build is not close to average. None of the three should be trusted as a precise, lab-grade number the way a DEXA scan or hydrostatic weighing would be, since all three are estimates built from an external proxy rather than a direct measurement of fat, lean tissue, and bone.

That ranking comes with real caveats, though, and the caveats matter more than the ranking itself in practice:

Skinfold accuracy depends entirely on tester skill. Finding the exact anatomical site, pinching a fold consistently rather than a mix of skin and muscle, and reading the caliper at the right moment all take practice. A skilled, consistent tester using the same technique every time produces genuinely useful numbers. An inconsistent tester, or worse, someone trying to pinch their own hard-to-reach sites, can produce numbers less reliable than the Navy method's simpler, harder-to-get-wrong tape measurement.

The Navy method needs no special skill, which is its real advantage. A tape measure around the neck and waist is far harder to measure badly than a skinfold pinch is, which is exactly why the U.S. military adopted it for large-scale, low-equipment fitness testing in the first place. It trades some individual precision for consistency across a huge number of people measured by people with no special training.

The BMI-based method is most useful when nothing else is available. With no tape measure and no calipers, a height, a weight, an age, and a sex are enough to get a rough estimate. It is the right tool when it is the only tool, not the first choice when a tape measure is sitting right there.

Where these three formulas actually came from

None of the three methods compared here is a recent invention, and knowing where each one came from explains a lot about what it was actually built to do. The skinfold approach traces back to A.S. Jackson and M.L. Pollock, who published their generalized skinfold equations for estimating body density in adult men in 1978, later extended to women with A. Ward in 1980. Both versions were derived by taking skinfold measurements and ages from a large sample of adults alongside a direct, lab-measured body density reading (from hydrostatic weighing), then fitting a regression equation to predict that lab-measured density from the much simpler skinfold and age inputs. The conversion from body density to a body fat percentage in all three methods compared in this article, not just skinfold, uses the Siri equation, published by Rudolf Siri in 1961, a two-component model that treats the body as fat mass and fat-free mass with two different, well-established densities.

The Navy method has a more specifically military origin. It was developed by J.A. Hodgdon and M.B. Beckett at the Naval Health Research Center and published in 1984, built for exactly the same practical constraint the U.S. military still uses it for today: assessing body composition for a very large number of service members quickly, without needing calipers, a hydrostatic tank, or any specially trained tester beyond someone who can read a tape measure correctly.

The BMI-based method is the youngest of the three. Paul Deurenberg and colleagues published the prediction formulas behind it in 1991 in the British Journal of Nutrition, explicitly framing BMI as an imperfect but convenient proxy for body fatness once age and sex are factored in, since BMI alone was already widely collected in population health surveys where a tape measure or calipers were not.

That lineage matters for judging which is the most accurate body fat method for a given situation: the skinfold and Navy formulas were both fitted directly against lab-measured body density, which is why they tend to track an individual's actual composition more closely than the BMI-based formula, which was fitted as a population-level proxy from the start and was never trying to be the most accurate body fat method for any one person, only a reasonable stand-in when nothing more direct was available.

Common measurement mistakes that quietly change the answer

A meaningful share of the disagreement between these three numbers, in practice, comes from measurement error rather than a genuine limitation of any formula. Each method has its own common failure point.

Skinfold calipers. The single biggest source of error is inconsistent site location and pinch technique: pinching too much underlying muscle along with the skin fold, not lifting the fold cleanly away from the muscle first, or reading the caliper before it has settled. Self-testing sites that are hard to reach on your own body, such as the subscapular or suprailiac sites, is particularly unreliable for exactly this reason.

The Navy method. The waist measurement is the most commonly mismeasured input, usually by measuring at the narrowest point of the torso instead of at the navel, which produces an artificially lean-looking result. The neck measurement, taken too low or with the tape pulled too tight, is the second most common error.

The BMI-based method. Because this method only ever sees BMI, age, and sex, any error here is really a height or weight entry error, not a technique error. A stale weight, or converting pounds and inches incorrectly, moves this method's result more directly than the same-size error would move either of the other two, since BMI feeds into the formula with a larger coefficient than either the Navy method's circumference gap or the skinfold method's summed thickness.

How much does skinfold measurement error actually matter?

Because the skinfold formula responds directly to the summed measurement, a look at how much the result shifts as that sum changes shows exactly how much a measurement error can move the final number, for the same 30-year-old:

A 15 mm difference in the summed skinfold reading, easily within the range of tester-to-tester variation on the same person, shifts the result by roughly 4 to 5 percentage points, enough to move someone from Athletes to Fitness on the ACE chart. This is the practical argument for consistency mattering more than any single reading: the same slightly-imprecise tester, measuring the same way every time, will show a real trend over weeks even if their absolute number carries a few points of bias.

A practical way to use all three together

Rather than treating this as a contest with one winner, the more useful habit is running more than one method and paying attention to when they agree and when they diverge. When the Navy method, skinfold, and BMI-based estimate land close together, that agreement is itself informative: it suggests this particular body's proportions are close to what all three formulas assume as typical, and any one of the three is probably a reasonable number to track going forward. When they diverge sharply, the way they did in the worked example above, that divergence usually points to something specific, most often an above-average or below-average amount of muscle relative to the general population each formula was built from, and it is worth treating the direct measurements (skinfold or Navy) as more trustworthy than the BMI-based estimate in that case, since the direct measurements are at least responding to this specific body rather than to a population average.

This site's body fat calculator runs the Navy method and the BMI-based method side by side automatically once you enter your age and weight alongside your circumference measurements, which is the fastest way to see your own version of the comparison above without doing the arithmetic by hand, and about as direct an answer to how to find your body fat percentage without lab equipment as any of these three methods can give on their own. If your build is unusually muscular or unusually lean, cross-checking against the BMI calculator is also worth doing, since a large gap between your BMI category and your body fat category is exactly the signal that one of these formulas is reading your build less accurately than the others.

The takeaway

No single field method here deserves blind trust, and that is not a flaw unique to any one of them, it is the nature of estimating body composition from an external proxy instead of measuring it directly. Skinfold, done well, generally comes closest among these three. The Navy method is the most forgiving of tester skill. The BMI-based method is the fallback when no tape or calipers are available at all. Run more than one, notice whether they agree, and treat a real disagreement as information about your own build rather than a reason to distrust the whole idea of estimating body fat outside a lab.