12 formulas · citations
The Methodology
Open methodology · Audit anything

Show your work.

Every formula in this engine is published, cited, and verifiable. No hidden multipliers. No "trust me, I'm a coach." Pull our citations, plug your numbers, and re-derive every output yourself.

12 cited formulas 7 peer-reviewed sources Open changelog v1.0 audit Q3 '26
The pact

Three commitments
that compound trust.

01 · Citations

Every output, cited.

Every number this engine returns lists its source — Mifflin–St Jeor (J Am Diet Assoc, 1990), ISSN Position Stand on Protein (2017), Schoenfeld 2016 hypertrophy volume meta. No black boxes.

02 · Versioning

No silent math upgrades.

If we change a formula coefficient, you'll see a changelog entry. Past calculations remain reproducible. Your Blueprint from v1.0 will still be derivable when v2.0 ships.

03 · Auditability

Bring your own second opinion.

Open source core arriving Q3 2026 with a public bug bounty for math errors. Until then, every formula is documented below — pull a calculator, run our inputs, get the same outputs. If you can't, file an issue and we ship a fix in the next deploy.

The stack

Every formula
in the engine.

12 calculators, 7 peer-reviewed sources. Click any to see the citation + plug-in form.

ToolFormulaSource · YearAudit
BMR · Mifflin10·W + 6.25·H − 5·A + sJ Am Diet Assoc · 1990Run →
BMR · Katch-McArdle370 + 21.6·LBMKatch & McArdle · 2002Run →
TDEEBMR × activity factor (1.2–1.95)ACSM Position Stand · 2018Run →
Protein target1.6–2.2 g/kg body weightISSN · 2017Run →
Body fat % · Navy495 / (1.0324 − 0.19077·log10(W−N) + 0.15456·log10(H)) − 450U.S. Navy Method · 1981Run →
BMI · Asian-PacificW / H² (cutoffs: 23/27.5)WHO Expert Consultation · 2004Run →
1RM · EpleyW × (1 + R/30)Epley · 1985Run →
1RM · BrzyckiW × (36 / (37 − R))Brzycki · 1993Run →
Hypertrophy volume10–20 sets/muscle/weekSchoenfeld et al. · 2016Library →
HIIT zones%HRmax = 220 − ageACSM HIIT Stand · 2022Today →
Cycle-sync (women)Phase-adjusted RPESims et al. · 2023Read →
Indian micronutrientsICMR-NIN RDA tables · 2020ICMR · 2020Diet hub →
Reproduce it

Audit your own BMR.

Sample worked example. Plug your numbers, get the same answer. If you don't, mail us — we ship a fix.

Worked example

BMR for a 32-year-old male, 75 kg, 178 cm

Mifflin–St Jeor (male):
BMR = 10·W + 6.25·H − 5·A + 5
BMR = 10·75 + 6.25·178 − 5·32 + 5
BMR = 750 + 1112.5 − 160 + 5
BMR = 1707.5 kcal/day

Open /bmr.html in another tab. Plug 75 / 178 / 32 / male. Engine returns 1,708 kcal/day (rounded). If yours doesn't match, screenshot & mail.

Public roadmap

What comes next.

Version
What
When
Status
v1.0
Open methodology page (this page) · 12 calculators audited
Q2 2026
Shipped
v1.1
100-profile validation report · signed PDF + raw CSV
Q3 2026
Drafting
v1.2
Public bug bounty · open-source calculator core
Q4 2026
Planned
v2.0
Live /api/validate · signed JSON for any output
2027
Planned

Changelog auto-publishes to /changelog. No silent updates.

Why audit-able matters

Your body deserves
verifiable math.

Most fitness apps quote a TDEE without showing the formula. So a 75 kg man might see 2,100 from one app and 2,400 from another — both "scientific". Cumulative error over a 12-week cut: 25,000 kcal · 3.5 kg of mis-attributed fat loss / gain.

When the math is open, you can pin down where the difference comes from (activity factor assumed? Mifflin or Harris-Benedict?) and pick what fits your body, not a black box's average.

Honest math also means honest expectations. If you know your BMR is 1,650 and the engine says you can lose 0.5 kg/week at a 500 kcal deficit, you can verify the assumption (1 kg fat ≈ 7,700 kcal) and stop chasing magic.

FitnessByMaddy's whole stance is that information transparency creates better outcomes than fancy metaphors. This page is the proof.

Ready to audit your body?

Start with the Blueprint — it derives all 12 outputs from one form. Every output cites its formula.

Open the Blueprint →