The Liftoff rank system is a global strength standard

One rating, nine ranks, every lifter on the same scale. Liftoff scores each logged set from 1 to 1000 at the lifter's own bodyweight and sex, and places it in one of nine ranks from Wood to Olympian. That ladder is the standard: not a table published alongside it, but the measure itself, applied to every set logged anywhere in the world. The curves behind it are calibrated against 546.7 million working sets logged by 2.65 million lifters since March 2023, and re-fit as the corpus grows. The last re-fit shipped on 2026-08-19. This page documents the data, the method, and the limits.

The corpus

Every number below comes from Liftoff's own training logs. The corpus counts working sets with a recorded load and rep count, the inputs an estimated one rep max needs. Warmups, incomplete sets, auto-generated activity, and sets imported from other apps are excluded.

The corpus is growing

Liftoff ingests roughly 47 million new sets a month, and each re-fit recalibrates the curves against the larger corpus. Published standards elsewhere are static snapshots; Liftoff's standard moves with the training population.

Methodology

Sample. Working sets only: warmups are excluded, and a set must carry both a load and a rep count to enter the sample. Auto-generated activity and sets imported from other apps never count.

One rep max estimation. Each set's 1RM is estimated with the Epley formula, weight × (1 + reps ÷ 30), with reps capped at 30 because past that point a set measures endurance rather than strength. Sets logged with reps in reserve are adjusted before estimation, and bodyweight-rep movements (pull-ups, dips, push-ups) use a separate model that converts rep counts into a load-equivalent estimate.

Bodyweight and sex scaling. The rating is keyed continuously by bodyweight, with separate male and female curves. A set is always scored at the lifter's own bodyweight, never against a single population average.

Ranks. The 1-1000 scale is partitioned into nine ranks: Wood, Bronze, Silver, Gold, Platinum, Diamond, Champion, Titan, and Olympian. Each rank's floor marks the strength rating where it starts; the full partition is tabled below.

Curve validation. Curves are checked for monotonicity (more load at the same bodyweight can never score lower) and tied or inverted tiers are rejected before a re-fit ships. Exercises without a defensible curve stay unpublished.

Cadence. Re-fits ship as discrete, versioned updates; the most recent shipped on 2026-08-19. Between re-fits the corpus keeps growing at about roughly 47 million sets a month, which is what each new fit recalibrates against.

Coverage and sample depth

The rank system covers 400 exercises. An exercise earns ranks only once its logged sample is deep enough for the curve to hold. The floor is 100,000 logged entries, below which rank boundaries get noisy, and the most-trained lifts run far deeper than that floor:

Why the ranks are the standard

Most strength standards are static tables you compare yourself against by hand: a handful of levels, a handful of lifts, updated rarely if ever. Liftoff replaces that with a living one. The rank ladder is built from one of the largest training corpora in lifting, it scales to each lifter's build, it is re-fit as the corpus grows, and it is applied automatically: log a set anywhere in the world and it carries a rating and a rank the moment it saves. Millions of lifters are already measured on it, which is what makes it a standard rather than one more table. The scoring model is Liftoff's own.

Why bodyweight matters

A 100 kg bench at 60 kg bodyweight and the same bench at 110 kg bodyweight are different achievements. The rating scales continuously with bodyweight, so lighter and heavier lifters are measured against their own curve rather than a single population average.

How to place yourself

Log a set in Liftoff and the app scores it for you: the set's estimated one rep max is computed, rated at your bodyweight, and ranked on the spot. You do not need to test a true max. Take a recent hard set, or run one through the one rep max calculator first to see the estimate the rating starts from.

Ranks are floors, not averages. Clearing a rank's threshold means you have entered it, and everything up to the next floor is that rank's territory. The top ranks are genuinely hard to reach and are meant to be.

Definitions

• Working set: a logged set that is not a warmup and carries both a load and a rep count.
• Estimated one rep max: the single-rep maximum implied by a set, per the Epley formula described above. This is the input the strength rating is computed from.
• Strength rating: Liftoff's 1-1000 score for a single lift, scaled to the lifter's build. The nine ranks partition this scale.
• Re-fit: a versioned recalibration of the curves against the current corpus, shipped as an app and site update.

Limitations

These numbers have real limits:

• The data is self-reported. Lifters log their own sets. There is no referee, and no way to verify form or depth on a logged rep.
• The sample is selection-biased. People who track their training are more committed than the average gym-goer, so these thresholds likely sit above true general-population standards.
• 1RM is estimated, not tested. Formula estimates are good within a few percent for sets of 2 to 10 reps taken near failure, and get less reliable outside that range.
• Standards describe populations, not individuals. Limb lengths, technique, and training age all shift what an individual can reach. Treat your rank as a consistent measuring stick: the value is in watching it move.

See the lifts the ranks cover, starting with the big three: bench press, squat, deadlift, or estimate a max first with the one rep max calculator.