The WIMSTA Formula Registry

A verified, machine-readable registry of body-composition formulas — with sources, samples and error margins. 15 formulas · v2026.08.11 · updated 2026-08-11.

Use this data

The registry is open under CC BY 4.0 — the coefficients are facts; our value is the verification and the structure. Use them anywhere, just credit WIMSTA (wimsta.com) with a link.

curl https://wimsta.com/api/registry.json | jq '.formulas[] | select(.id=="cun-bae")'

Every formula, with its coefficients

Verification is honest: ✅ primary = checked against the original paper, 🟡 secondary = checked against secondary sources only. The same status the method page and the JSON report.

Formula registry v2026.08.11
Formula Category Inputs Equation & coefficients Source Error Status
Deurenberg 1991
id: deurenberg
body-fat bmi, age, sex
male=1, female=0
BF% = 1.20·BMI + 0.23·age − 10.8·sex − 5.4 (BMI in kg/m², age in years) {"bmi":1.2,"age":0.23,"sex":-10.8,"intercept":-5.4} Br J Nutr 65(2):105–114
Deurenberg, Weststrate & Seidell
SEE 4.1 pp · R² 0.79
CUN-BAE 2012
id: cun-bae
body-fat bmi, age, sex
female=1, male=0 (⚠️ inverse of Deurenberg)
BF% = −44.988 + 0.503·age + 10.689·female + 3.172·BMI − 0.026·BMI² + 0.181·BMI·female − 0.02·BMI·age − 0.005·BMI²·female + 0.00021·BMI²·age {"intercept":-44.988,"age":0.503,"female":10.689,"bmi":3.172,"bmi2":-0.026,"bmi_female":0.181,"bmi_age":-0.02,"bmi2_female":-0.005,"bmi2_age":0.00021} Diabetes Care 35(2):383–388
Gómez-Ambrosi et al.
SE ≈ 4.7% · r 0.89
US Navy (Hodgdon–Beckett) 1984
id: us-navy
body-fat waist, neck, height, sex, hip (women only)
separate equations for men and women (women also require hip)
men: BF% = 495 / (1.0324 − 0.19077·log₁₀(waist−neck) + 0.15456·log₁₀(height)) − 450; women: BF% = 495 / (1.29579 − 0.35004·log₁₀(waist+hip−neck) + 0.221·log₁₀(height)) − 450 (cm) {"men":{"c0":1.0324,"cLogWaistNeck":-0.19077,"cLogHeight":0.15456},"women":{"c0":1.29579,"cLogWaistHipNeck":-0.35004,"cLogHeight":0.221},"transform":"495 / denominator − 450"} U.S. Navy NHRC Reports 84-11 / 84-29
Hodgdon & Beckett
±3–4 pp · r ≈ 0.90 🟡
RFM (Relative Fat Mass) 2018
id: rfm
body-fat height, waist, sex
intercept 64 male / 76 female (the sex term is the +12 offset)
men: RFM = 64 − 20·(height/waist); women: RFM = 76 − 20·(height/waist) (cm) {"men":{"intercept":64,"heightOverWaist":-20},"women":{"intercept":76,"heightOverWaist":-20}} Sci Rep 8:10980
Woolcott & Bergman
more accurate than BMI vs DXA
Covert Bailey 1999
id: covert-bailey
body-fat waist, hip, forearm, wrist, age, sex, thigh (women), calf (women)
separate equations; women use thigh+calf, men use waist+hip+forearm
men (inches): waist + 0.5·hip − forearmCoef·forearm − wrist (forearmCoef 3 if age≤30 else 2.7); women (inches): hip + thighCoef·thigh − 2·calf − wrist (thighCoef 0.8 if age≤30 else 1.0) {"men":{"waist":1,"hip":0.5,"forearm":"coef","wrist":-1,"forearmCoef_le30":-3,"forearmCoef_gt30":-2.7},"women":{"hip":1,"thigh":"coef","calf":-2,"wrist":-1,"thighCoef_le30":0.8,"thighCoef_gt30":1}} “Ultimate Fit or Fat” (book)
Covert Bailey
not formally published 🟡
Gallagher healthy ranges 2000
id: gallagher-ranges
body-fat-ranges age, sex
ranges differ by sex
Age- and sex-specific HEALTHY body-fat % ranges (a reference band, not an estimator) Am J Clin Nutr 72(3):694–701
Gallagher et al.
age- and sex-specific ranges 🟡
WHO BMI cut-points 2000
id: who-bmi
bmi bmi BMI cut-points (kg/m²): underweight <18.5 · normal 18.5–25 · overweight 25–30 · obese I 30–35 · II 35–40 · III ≥40 {"cutPoints":[18.5,25,30,35,40]} WHO Technical Report Series 894
World Health Organization
bands 18.5 / 25 / 30 / 35 / 40
Trefethen “new BMI” 2013
id: trefethen-bmi
bmi height, weight new BMI = 1.3·weight / height^2.5 (weight kg, height m) {"factor":1.3,"heightExponent":2.5} The Economist (letter) / Oxford page
Nick Trefethen (Oxford)
1.3 · w / h^2.5
Ashwell boundary 2016
id: ashwell-whtr
waist-to-height waist, height waist ÷ height; screening boundary at 0.5 ("keep waist under half your height") {"boundary":0.5,"lowFlag":0.4} BMJ Open 6:e010159
Ashwell & Gibson
boundary 0.5 (“keep waist under half your height”)
Tanaka 2001
id: tanaka
max-heart-rate age HRmax = 208 − 0.7·age (bpm) {"intercept":208,"age":-0.7} J Am Coll Cardiol 37(1):153–156
Tanaka, Monahan & Seals
208 − 0.7·age · SD ±10 bpm 🟡
Gulati 2010
id: gulati
max-heart-rate age, sex
women only (derived in a female cohort)
HRmax = 206 − 0.88·age (bpm), women {"intercept":206,"age":-0.88} Circulation 122(2):130–137
Gulati et al.
206 − 0.88·age 🟡
Fox (220 − age) 1971
id: fox
max-heart-rate age HRmax = 220 − age (bpm) {"intercept":220,"age":-1} no strict primary derivation
Fox, Naughton & Haskell
220 − age · SD ±10–12 bpm 🟡
Harris–Benedict 1919
id: harris-benedict-1919
bmr weight, height, age, sex
separate equations for men and women
men: BMR = 66.473 + 13.7516·weight + 5.0033·height − 6.755·age; women: BMR = 655.0955 + 9.5634·weight + 1.8496·height − 4.6756·age (kcal/day; kg, cm, years) {"men":{"intercept":66.473,"weight":13.7516,"height":5.0033,"age":-6.755},"women":{"intercept":655.0955,"weight":9.5634,"height":1.8496,"age":-4.6756}} Carnegie Inst. of Washington, Pub. 279
Harris & Benedict
overestimates modern BMR ~5% 🟡
Harris–Benedict revised 1984
id: harris-benedict-1984
bmr weight, height, age, sex
separate equations for men and women
men: BMR = 88.362 + 13.397·weight + 4.799·height − 5.677·age; women: BMR = 447.593 + 9.247·weight + 3.098·height − 4.33·age (kcal/day; kg, cm, years) {"men":{"intercept":88.362,"weight":13.397,"height":4.799,"age":-5.677},"women":{"intercept":447.593,"weight":9.247,"height":3.098,"age":-4.33}} Am J Clin Nutr 40(1):168–182
Roza & Shizgal
re-derivation of Harris–Benedict
Katch–McArdle (Cunningham 1991) 1991
id: katch-mcardle
bmr weight, bodyFat BMR = 370 + 21.6·LBM, where LBM = weight·(1 − bodyFat/100) (kcal/day; kg) {"intercept":370,"leanBodyMass":21.6} Am J Clin Nutr 54(6):963–969
Cunningham (popularised by Katch & McArdle)
strong with a measured body-fat %

Why this exists

Popular calculators copy each other, and they copy each other’s mistakes. The YMCA body-fat formula is everywhere despite having no locatable primary source and the worst fit against NHANES data in our testing (mean absolute error 8.4 percentage points). “Ideal weight” coefficients get quoted as health targets when they were built for drug dosing. The same equation shows up under two names — “Katch–McArdle” is, term for term, Cunningham 1991 — which lets sites fake a wider “range” by counting one estimate twice.

We publish the verified version instead: every coefficient checked against the source we cite, every verification status shown as it is, and the machine-readable file kept in lockstep with the human pages by generating both from one registry.

Changelog

  • 2026-08-11 — first public release of the machine-readable registry (JSON + this page), generated from the same source as the method page.
  • August 2026 — YMCA formula removed; CUN-BAE added and verified against the primary source; Deurenberg status corrected to verified.