Three methods, three answers — which one should you trust?
A DEXA scan costs $50–200 and gives roughly ±2% accuracy. The three free methods on this page give ±3–5% in their training populations and considerably worse outside them. Pick the wrong method for your body type and the gap widens to 8–12%. Here is how to choose, and how to read the result the calculator just gave you.
U.S. Navy circumference method
Inputs: height, neck, waist (and hip circumference for women). Trained against hydrostatic weighing on U.S. Navy personnel in the 1980s (Hodgdon & Beckett, 1984). Best for: average-fitness adults with typical proportions. Worst for: very lean athletes (under-estimates body fat), heavy non-obese individuals with broad frames (over-estimates). Not valid during pregnancy. Accuracy: ±3–4% versus DEXA in the training population.
Jackson–Pollock 3-site skinfold
Inputs: caliper measurements at three sites (men: chest, abdomen, thigh; women: triceps, suprailiac, thigh). Requires a caliper (around $20) and ideally a second person to take the measurements consistently. Best for: anyone with caliper access. The most accurate of the three methods at population extremes (very lean and very heavy). Worst for: anyone trying to measure their own thigh skinfold solo — just don’t. Accuracy: ±3–4% with good measurement technique, ±8–10% without it.
Deurenberg BMI-based formula
Inputs: BMI, age, sex. The formula is Body Fat % = 1.20 × BMI + 0.23 × age − 10.8 × sex − 5.4 (sex = 1 for men, 0 for women). Convenient (no calipers, no tape measure) but inherits BMI’s blindness to body composition: a recreational lifter and a sedentary office worker with identical BMI get identical Deurenberg numbers. Accuracy: ±5% in average populations, ±10%+ in athletes. Use only when measurement is impossible.
How to pick
Average adult, no caliper, want a quick estimate → U.S. Navy. Trained, regularly lifts, has caliper access → Jackson–Pollock. No measurements available at all, just BMI → Deurenberg as a rough placeholder. Want clinical-grade precision → pay for a DEXA scan; nothing on this page replaces it.
Convergence is the real signal
Run all three methods on the same person. If Navy says 18%, Jackson–Pollock says 19%, Deurenberg says 17%, you’re between 17–19% — the agreement is the signal that you fit the average population each method was trained on. If they spread 15–25%, take it as a warning sign that at least one of them is poorly calibrated for your body composition.
Takeaway: Pick the method that matches your inputs and body type, not the one that gives the prettiest number. U.S. Navy for average adults, Jackson–Pollock for trained populations with caliper access, Deurenberg as a fallback only. Re-measure every 4–6 weeks under similar conditions (morning, hydrated, before breakfast) for a comparable trend — a single measurement is a snapshot, not a verdict.
Sources: ACE Fitness body-composition classification · Overview of body-fat measurement methods.