How these standards are computed

Every percentile on this site is computed from the OpenPowerlifting database, which is dedicated to the public domain. We publish the method because a percentile without a stated population and sample is not information.

Dataset

Build 2026-07-25-606a59af. 3,991,443 meet results read, of which 1,059,855 passed the filters below. These are judged competition results — measured one-rep maxes performed under competition commands, not estimates and not self-reports.

Filters

• Raw equipment only — no supportive suits or shirts.
• Age 20–45, so youth and masters divisions don't distort the curve.
• Best-ever per lifter, not per meet entry — a lifter who competes twenty times counts once, at their best.
• Bodyweight window of ±5 lb around each page's centre.
• Minimum sample of 1,000 to publish a table at all.

What we don't do

• We don't interpolate. If a cell falls below the minimum sample, the page for it does not exist rather than being filled in from neighbouring values.
• We don't publish the extreme tail. A 99th percentile is computable but invites "elite" framing that is meaningless without knowing the federation, meet, and year.
• We don't generate the age × bodyweight cross-product. That is where the sample runs out, and thin cells produce confident-looking numbers with nothing behind them.
• We don't publish a general-population average. No dataset of measured maxes among non-competitors exists to support one.

Known limitations

No stable lifter id exists in this export, so deduplication is by name. Distinct lifters sharing a name collapse into one; a lifter recorded under name variants counts more than once.

More fundamentally, this population self-selected into competition. It answers "what do strong, trained lifters at this bodyweight do" — not "what is normal". Any reading of these tables as a general-population benchmark is a misreading, and we'd rather say so than let the framing do quiet work.

Source: OpenPowerlifting database, build 2026-07-25-606a59af (public domain). Percentiles computed by Motus. Method.