Editorial policy

StrengthMath is written and edited by Jimmy L Wu. The site publishes calculator-driven research and sourced strength math, not hot-take opinion. This page documents the standards every page is held to.

Sourcing tiers

Every numerical claim — formula, band, threshold, conversion factor — carries a source label, on the page that uses it. Tiered:

• Tier 1 — primary references.Original-publication equations (Epley, Brzycki, Lombardi, O'Conner), NIST conversion factors, peer-reviewed validation studies (LeSuer 1997). Reproduced verbatim with the formula or constant cited.
• Tier 2 — standards bodies. NSCA (Essentials of Strength Training and Conditioning, 4th ed.), ACSM 2026 Position Stand on Resistance Training, AAP guidance on youth resistance training. Cited by name where the page leans on their framing.
• Tier 3 — population datasets. ExRx-aligned strength standards (training-population synthesis), OpenPowerlifting (competition-biased; flagged as such whenever referenced). Population data is treated as a reference, not as a ground truth.
• StrengthMath methodology. Where the published literature is sparse — dumbbell-vs-barbell conversion factors, teen-mode soft framing, reliability-band specific percentages — the site fills the gap with practitioner-consensus framing, and labels it as methodology rather than as published truth. This label appears in the calculator UI, the methodology page, and the engine output.

See /methodology for the per-engine sourcing detail.

What we don't do

A short, opinionated list of things that look like content but aren't:

• No invented anecdotes.No “a client of mine once…” padding from training data. If a story is told, it's real and verifiable.
• No fabricated publication numbers.Citations point to actual papers, position stands, and standards. We don't cite “NSCA-2017-XXXX” from training-data fragments.
• No AI-generated photographs as photos. Diagrams and illustrations may be created with AI tools; product photography and gym scenes are real or labeled.
• No competitor scraping.We don't pull calculator outputs, FAQ wording, or comparison tables from competing fitness sites. Numbers come from primary sources; copy is original.

Opinion injection

Sourced math without a point of view reads like AI-template content. Every page surfaces the author's real takes alongside the numbers — preferences across the four 1RM formulas, the ACSM-2026-vs-NSCA judgment for power, why 70–85% incline:flat is a population observation rather than a conversion factor, what dumbbell pressing actually trades off vs the barbell. Opinions are marked as opinions and labeled as the author's, not as consensus. The math stays the math; the framing of how to use the math is opinionated on purpose.

AI policy

AI tools assist with drafting and editing. Every formula, band, and source citation is human-verified before publish — AI is not trusted to invent numbers. The on-site assistant matches user questions against a curated corpus of vetted seeds and deflects when the corpus has no good answer; it does not generate freeform answers about lifting on the user's situation. AI-assisted features are disclosed where they appear.

Review cadence

Pages list a review date in the footer. Time-sensitive claims (any forthcoming federation rule changes, position-stand revisions) are re-checked when the underlying source updates. Engine constants are locked behind tests — a future change to a formula or band triggers a Vitest failure, not a silent drift.

Corrections

If a page has a factual error, email admin@strengthmath.com. Corrections are made promptly, the page's “Updated” date reflects the change, and significant corrections are noted inline so the change history is visible.