Skinfold Body Fat Calculator (Jackson–Pollock)

Background.

This calculator turns caliper skinfold measurements into an estimated body fat percentage using the Jackson–Pollock generalized equations — the 3-site and 7-site forms that have been the working standard in exercise physiology since 1978. Pick your sex and equation, enter your age, your weight and the relevant skinfolds in millimetres, and you get a predicted body density plus the same density converted to a fat percentage two different ways. Take the result as a screening estimate, not a measurement: the typical individual error against an underwater weighing or DXA reference is around three and a half percentage points, and how you pinch the fold matters more than which equation you pick.

The two-step structure is worth understanding because most calculators hide it. Skinfold equations do not predict body fat directly. They predict body density — how many grams a cubic centimetre of you weighs — because that is what hydrostatic underwater weighing, the reference method Jackson and Pollock validated against, actually measures. A second equation then converts density to a fat percentage by assuming fat has a density of about 0.900 g/cm³ and everything else about 1.100 g/cm³, and solving for the mixture. This calculator shows you the density it predicted, so you can see the step rather than trust it.

There are two competing conversions and this page runs both. Siri's, published in 1961, gives percent fat as 495 divided by density, minus 450. Brozek and colleagues published a revised set of assumptions in 1963 giving 457 divided by density, minus 414.2. They are close but not identical, and — contrary to a widespread claim — neither one always reads higher. They cross at a density of 1.06145 g/cm³: below that, Siri reads higher; above it, Brozek does. In the worked example on this page a body density of 1.064323 gives Siri 15.08 % and Brozek 15.18 %, with Brozek the larger of the two, because that density sits above the crossover. A separate 2010 study in the Journal of Nutrition, Health and Aging compared both against DXA in older adults and concluded that Brozek was the more accurate alternative in that group. That finding is the reason both numbers appear here rather than one being chosen for you.

The Jackson–Pollock equations themselves came from two papers: a 1978 study in the British Journal of Nutrition on 308 men aged 18 to 61, cross-validated on a further 95, and a 1980 study in Medicine and Science in Sports and Exercise on women aged 18 to 55. Both fitted body density as a quadratic in the sum of skinfolds plus a linear term in age. The multiple correlations exceeded 0.90 with standard errors of about 0.0073 g/ml in density, which translates to roughly three to four percentage points of body fat. Age is genuinely in the model, not decoration: at the same skinfolds, a 60-year-old is predicted to carry more fat than a 20-year-old, because subcutaneous fat makes up a smaller share of total fat as people age. That is also why this calculator refuses ages outside 18 to 61 — the equations were never fitted there and are documented as invalid in the elderly.

The 3-site and 7-site versions ask different things of you. For men the 3-site sites are chest, abdomen and thigh; for women they are triceps, suprailiac and thigh — a genuine difference, not an oversight, and one that reflects where each sex stores subcutaneous fat. The 7-site version adds midaxillary, subscapular and, for women, chest and abdomen, and uses the same seven sites for both sexes with sex-specific coefficients. More sites reduce the effect of a single bad measurement, so 7-site is the better choice if you can take all seven reliably; 3-site is the better choice if you cannot, because a poorly taken subscapular fold does more damage than leaving it out. When you choose the 3-site equation, this calculator ignores the other four inputs completely — the 'sites consumed' tile tells you exactly which ones went in.

One more limitation is built into the mathematics and worth knowing. Each of these equations is a downward-then-upward parabola in the sum of skinfolds, so past a certain total it turns around and starts predicting less body fat for thicker folds. Those turning points sit at 258 mm for the men's 3-site, 216 mm for the women's 3-site, 395 mm for the men's 7-site and 419 mm for the women's 7-site. Beyond them the equation is not conservative, it is wrong, so the calculator refuses rather than printing a number. It also rejects any single fold outside 3 to 80 mm, 80 mm being the jaw range of a Harpenden caliper. Within those limits the predicted density always lands between 0.900 and 1.100 g/cm³, which is what keeps the percentage physically possible.

Finally, technique. Grasp the fold firmly with thumb and forefinger about a centimetre away from where the caliper jaws will sit, lift it clear of the underlying muscle, apply the caliper perpendicular to the fold at mid-depth, and read after two seconds while still holding the fold. Take every site on the right side of the body, rotate through all of them, then repeat the whole rotation twice more and use the median. Two measurements of the same site should agree within a millimetre. If they do not, your technique is the largest source of error in your result and no choice of equation will fix it.

What is skinfold body fat calculator?

A skinfold body fat test estimates body composition by measuring the thickness of a pinched double layer of skin and the subcutaneous fat beneath it at standardised anatomical sites, then feeding the sum of those measurements into a regression equation that predicts whole-body density. The Jackson–Pollock generalized equations are the most widely used set. Andrew Jackson and Michael Pollock published the men's equations in the British Journal of Nutrition in 1978, having measured skinfolds, circumferences and hydrostatically determined body density in 308 men aged 18 to 61 and cross-validated the resulting models on a separate sample of 95; Jackson, Pollock and Ann Ward published the companion women's equations in Medicine and Science in Sports and Exercise in 1980. The word 'generalized' is the point: earlier skinfold equations were population-specific and broke down outside the group they were fitted on, whereas these were designed to hold across a wide range of ages and fatness by including age as a predictor and using a quadratic rather than linear term in the skinfold sum. Body density is then converted to percent fat by a two-compartment model, either Siri's 1961 formulation or the revised constants published by Brozek, Grande, Anderson and Keys in 1963. Both models assume the body is exactly two tissues of fixed density, which is the method's deepest limitation: bone mineral density, hydration and the density of fat-free tissue all vary between people and are all assumed constant. Skinfolds also see only subcutaneous fat, so they are blind to visceral fat, and the proportion of total fat that sits under the skin shifts with age and differs between populations. Used consistently by a practised tester, however, the method is reproducible to about a percentage point, which makes it a genuinely useful tool for tracking change.

How to use this calculator.

1. 01Choose your sex and equation. If you can reliably take all seven folds, use 7-site; if not, 3-site with three good measurements beats 7-site with four bad ones.
2. 02Enter your age. It is a real predictor in the equation, not a label — the calculator only accepts 18 to 61, the range these equations were fitted on.
3. 03Take every measurement on the RIGHT side of the body, standing relaxed, with dry skin and no lotion.
4. 04Pinch the fold with thumb and forefinger about 1 cm from where the caliper jaws will go, lift it clear of the muscle underneath, and keep holding while you measure.
5. 05Place the caliper jaws perpendicular to the fold, at mid-depth between its crest and its base, release the caliper's spring fully, and read after two seconds.
6. 06Rotate through all your sites once, then repeat the whole rotation twice more. Use the median of the three readings for each site — not the first, and not the smallest.
7. 07Enter each fold in millimetres. The calculator ignores the sites your chosen equation does not use, and the 'sites consumed' tile confirms which ones went in.
8. 08Read both the Siri and Brozek percentages. If they differ by more than a point you are near the crossover density; the gap is a property of the models, not of you.
9. 09Repeat monthly with the same caliper, same tester and same sites. Absolute skinfold numbers are only worth what your technique is worth, but the change over time is reliable.

The formula.

Step one predicts body density: Db = a − b × S + c × S² − d × age, where S is the sum in millimetres of only the sites the chosen equation uses. The four published coefficient sets are: men 3-site (chest, abdomen, thigh) 1.10938, 0.0008267, 0.0000016, 0.0002574; women 3-site (triceps, suprailiac, thigh) 1.0994921, 0.0009929, 0.0000023, 0.0001392; men 7-site 1.112, 0.00043499, 0.00000055, 0.00028826; women 7-site 1.0970, 0.00046971, 0.00000056, 0.00012828. Step two converts density to percent fat with a two-compartment model: Siri gives 495 ÷ Db − 450, which is the same as (4.950 ÷ Db − 4.500) × 100, and Brozek gives 457 ÷ Db − 414.2, the same as (4.570 ÷ Db − 4.142) × 100. A common published error multiplies the first form by 100 a second time and produces a number around 1500 %; this calculator does not. Working the shipped example: a 30-year-old man with chest 12 mm, abdomen 22 mm and thigh 16 mm has S = 50 mm, so 0.0008267 × 50 = 0.041335, 0.0000016 × 2500 = 0.004 and 0.0002574 × 30 = 0.007722, giving Db = 1.10938 − 0.041335 + 0.004 − 0.007722 = 1.064323 g/cm³ exactly. Siri then gives 495 ÷ 1.064323 − 450 = 15.0843775809 %, and Brozek gives 457 ÷ 1.064323 − 414.2 = 15.1809304130 %. At 80 kg body weight that is 12.0675 kg of fat and 67.9325 kg of fat-free mass. Note the ordering: Brozek reads HIGHER here, because 1.064323 is above the two models' crossover density of 38 ÷ 35.8 = 1.0614525140; below that density the ordering flips. ROUNDING STAGE: body density is computed as an exact decimal, the two divisions run at full precision, and rounding to ten decimal places happens once, at the end. DOMAIN: each equation is a parabola in S and turns around at S = b ÷ 2c — 258.34 mm (men 3-site), 215.85 mm (women 3-site), 395.45 mm (men 7-site), 419.38 mm (women 7-site) — beyond which it predicts falling body fat for thicker folds. The calculator refuses those sums, refuses any single fold outside 3–80 mm, and refuses ages outside 18–61.

A worked example.

Frequently asked questions.

References& sources.

1. [1]Jackson AS, Pollock ML. "Generalized equations for predicting body density of men." Br J Nutr. 1978;40(3):497–504. doi:10.1079/BJN19780152. Origin of the men's 3-site and 7-site equations; 308-man development sample plus a 95-man cross-validation sample, ages 18–61, multiple correlations above 0.90 with standard errors of about ±0.0073 g/ml. Abstract open access; full text paywalled at Cambridge Core — the coefficients themselves are not readable in the abstract (see the next citation for the document in which they were verified).
2. [2]University of Michigan, MVS 240 Exercise Physiology — Lab A6-3, "Alternative Skinfold Measurement Formulas to Calculate Percent Body Fat" (Fahey, Insel & Roth, Fit and Well, 6th ed., McGraw-Hill, 2005; conversion table adapted from Heyward & Wagner, Applied Body Composition Assessment, 2nd ed., Human Kinetics, 2004; equations sourced there to Jackson & Pollock 1985 and ACSM's Guidelines for Exercise Testing and Prescription, 6th ed.). This is the document in which every coefficient shipped on this page was read verbatim. Note: it prints the men's 3-site age coefficient as 0.000257 where other reproductions give 0.0002574; this page uses 0.0002574, a difference of 0.007 percentage points of body fat at age 40.
3. [3]Jackson AS, Pollock ML, Ward A. "Generalized equations for predicting body density of women." Med Sci Sports Exerc. 1980;12(3):175–181. PMID 7402053. Origin of the women's 3-site (triceps, suprailiac, thigh) and 7-site equations; derivation sample ages 18–55. Abstract open access; publisher full text paywalled.
4. [4]Guerra RS, Amaral TF, Marques E, Mota J, Restivo MT. "Accuracy of Siri and Brozek equations in the percent body fat estimation in older adults." J Nutr Health Aging. 2010;14(9):744–748. PMID 21085903, PMC12880224. Independent second authority consulted for this build: prints both conversions verbatim — Siri "%BF = [4.950 / BD – 4.500] × 100" and Brozek "%BF = [4.570 / BD – 4.142] × 100" — and concludes that Brozek is the more accurate alternative in older adults. Open access.
5. [5]Siri WE. "Body composition from fluid spaces and density: analysis of methods." In: Brozek J, Henschel A, eds. Techniques for Measuring Body Composition. Washington DC: National Academy of Sciences, 1961:223–244. Reprinted in Nutrition 1993;9(5):480–491, PMID 8286893. Bibliographic origin of the 4.950/4.500 conversion. PRINT REFERENCE — the 1961 National Academy of Sciences volume is not available online; the linked PubMed record is for the 1993 reprint.
6. [6]Brozek J, Grande F, Anderson JT, Keys A. "Densitometric analysis of body composition: revision of some quantitative assumptions." Ann N Y Acad Sci. 1963;110:113–140. Bibliographic origin of the 4.570/4.142 conversion. PRINT / PAYWALLED reference — no open full text; cited for provenance of the constants only.
7. [7]Jackson AS, Pollock ML. "Practical assessment of body composition." Phys Sportsmed. 1985;13(5):76–90. doi:10.1080/00913847.1985.11708790. The practitioner paper that put the 3-site short forms into general use and the source the ACSM materials cite for them. Paywalled at Taylor & Francis.