What the score is, and what it is not
The score is a weighted blend of up to four physique and performance signals compared with published reference ranges and heuristic models. Two of the signals rest on measured data: the FFMI ceiling from drug-free lifters and published strength benchmarks. The other two, how fast lean mass arrives and how much of it, are coaching models of natural progress, not measurements. The weights, 45, 30, 15 and 10 percent, are design choices made by the author, not fitted to any dataset. A score of 85 means every signal in use sits inside the reference range with room to spare; a score of 20 means at least one signal, usually fat-free mass, sits well beyond it. The four labels describe the model, never the person: comfortably within this model's natural reference range, near the upper end of the reference range, rare within the reference data, and outside the reference range used by this model.
It is not a probability. A score of 70 does not mean a 70 percent chance of anything. No dataset exists that pairs physique measurements with verified drug status across the general lifting population, so the combined model has never been validated as a classifier of enhancement status and cannot be. The score is labelled heuristic wherever it appears, the share card says so, and the number is presented as a range because the body fat measurement it depends on carries its own error.
Signal one: fat-free mass index against the ceiling
Fat-free mass index is lean mass in kilos divided by height in metres squared, normalised to 1.8 m by adding 6.1 for every metre below that height. Kouri and colleagues measured 74 drug-free lifters and 83 steroid users in 1995. The drug-free group averaged 21.8 with a standard deviation of 1.8, and none exceeded 25. The steroid users averaged several points higher, with individuals above 30. Mr America winners from 1939 to 1959, before steroids reached bodybuilding, averaged 25.4. So 25 is used as the reference ceiling for men, and about 22 for women. It is an influential reference, not a biological law: the sample was 74 lifters, individual outliers exist, the historical comparison has its own limits, and a body-fat error of a few points moves FFMI by a point.
The calculator does not draw a hard line at 25. It uses a smooth curve centred just above the ceiling, so 24.5 scores lower than 22 and 26 scores lower still, and 24.9 and 25.1 score almost the same. Because a 3.5-point error in body fat moves FFMI by about a point, the signal is averaged across the body fat interval rather than computed once from the point estimate. If a man enters wrist and ankle circumference, the ceiling becomes frame-specific using Casey Butt's model, an empirical fit to bodybuilding records that raises the ceiling for large-boned lifters and lowers it for small frames. It is not a validated equation and not a biological maximum, its source data were male, so women use the population ceiling, and it only ever moves the heuristic score.
Signal two: could the years have built it?
This signal is a coaching model, not a measurement. It assumes an untrained adult already carries roughly 72 percent of their eventual lean-mass ceiling and that the rest is built on a curve that rises fast in the first two years and flattens after five. The calculator compares the fraction of your frame ceiling you have reached with the fraction expected after your years of training. A lifter at 95 percent of the ceiling after two years is ahead of the curve; the same lifter after twelve years is where the curve says he should be. This is why training years matter as much as the physique itself, and why entering the year you started rather than the years you trained seriously inflates the score in your favour.
Signal three: rate of gain
If you know your starting weight and body fat, the calculator computes lean mass gained and compares it with the cumulative natural gain models most coaches use: for men roughly 10 kg of lean tissue in the first year, 5 in the second, 2.5 in the third, then about a kilo a year tapering; for women about half. Fat gained never counts as lean, and a starting lean mass more than 10 kg above the current one is refused as inconsistent rather than scored. Gaining at twice the model's rate is the strongest signal the tool has, which is why it carries a 30 percent design weight when it is present. The gain rates themselves are coaching estimates, not measured population data.
Signal four: strength, weighted lightly
Your squat, bench and deadlift are placed on bodyweight-scaled benchmarks and contribute a 10 percent design weight. The weight is low on purpose. Elite relative strength is well documented in drug-tested powerlifting, and strength tracks leverage and technique as much as muscle. A big total on its own moves the score a little; a big total on a lean 26 FFMI moves it a lot, because the other signals already carry the case.
Where it goes wrong, honestly
Body fat is the weak link. A visual estimate that is five points low turns a 23 FFMI into a 24.5, and the score with it, so the tool asks where the number came from and widens the range for a guess. Frame size matters more than people think: two men at 180 cm and 92 kg lean can have very different ceilings, and without wrist and ankle measurements the calculator uses a population average. Genetic outliers exist, at both ends. And the drug-free reference sample is small and decades old. All of this is why the output is a score range with a heuristic label and never a verdict about a person.
What to do with the result
If the score says a physique is comfortably within the reference range, the useful number is the gap to the ceiling, which the max muscular potential calculator turns into kilos of lean mass still available. If the score is low and you are natural, check the body fat input first, then the training years, then the frame. If you are enhanced and curious, the tool is telling you how far past the reference range the numbers sit, which is exactly the information the FFMI calculator and the bloodwork tools are built around.