FFMI Calculator

Calculator

What FFMI measures, and why BMI cannot

Body mass index divides weight by height squared, and its great weakness is that it does not care what the weight is made of. That is fine as a population screen and useless for anyone who trains. Fat-free mass index applies the identical arithmetic to lean mass alone: strip the fat out first, then scale for height. What comes back is an index of muscularity rather than of heaviness.

The scales run differently, and that trips people up. A BMI of 25 is a boundary; an FFMI of 25 is near the top of what has been recorded in drug-free athletes. Ordinary untrained men cluster around 18 to 19, ordinary untrained women around 15. A Swiss survey of 5,635 adults found median fat-free mass index of 18.9 kg/m² in men aged 18 to 34 and 15.4 kg/m² in women, and, unlike BMI, those medians barely moved with age, while the companion fat mass index climbed by more than half. That single finding is a good argument for tracking both indices instead of a scale reading.

The useful property is that FFMI is roughly height-independent once normalised, which BMI is not. It lets you compare a 1.65 m and a 1.95 m lifter fairly, compare yourself against published reference percentiles, and, most practically, compare yourself against yourself over years without the number drifting for reasons that have nothing to do with training.

The formula, and why it needs a height correction

Raw FFMI is lean mass in kilograms divided by height in metres squared. Lean mass comes from your body fat percentage: weight minus weight times body fat over a hundred. So for an 82 kg person at 15 per cent body fat, lean mass is 69.7 kg, and at 1.78 m the raw index is 69.7 ÷ 1.78², which is 22.0.

The problem with stopping there is the same problem BMI has. Lean mass does not scale with the square of height: taller people carry less lean tissue than height squared predicts, so the raw index quietly penalises them and flatters short people. The standard fix, introduced with the index itself, is to add a linear correction that expresses everyone's index as what it would be at 1.80 m: normalised FFMI = FFMI + 6.1 × (1.80 − height in metres). Below 1.80 m the term is positive, above it the term is negative, and at exactly 1.80 m it vanishes, which is why some calculators appear to skip the step entirely for people of that height.

One coefficient note, because it is the sort of thing that quietly differs between tools. The original 1995 paper used 6.3; the figure in general use, and the one this page computes, is 6.1. Across any realistic adult height the two differ by less than 0.05 index points (at 1.65 m the correction is 0.92 rather than 0.95), so it changes nothing about your band. It is worth knowing only because it explains small disagreements between calculators that are both quoting the same source.

The interpretation bands and where they came from

The men's bands on this page follow the trained-and-untrained literature, anchored at the top by Kouri and colleagues' 1995 study of 157 male athletes (83 users of anabolic-androgenic steroids and 74 non-users), in which the normalised values of the drug-free group reached a well-defined limit of 25.0. Below that, roughly 18 to 20 is average for an untrained adult man, 20 to 22 reads as above average, and 22 to 25 is territory that ordinarily takes years of consistent training to reach.

The women's bands need a health warning, and this page prints one next to them. There is no equivalent study of drug-free female athletes establishing a comparable ceiling. The boundaries you will see quoted everywhere (around 14 to 16 average, 16 to 18 above average, 18 to 21 muscular) are the men's scale shifted down by about four points, which is a convention that happens to sit roughly where population data put women, not a regression anybody published. Read them as a rough orientation and put more weight on your own trend than on which label you land in.

Population percentiles are a different and more solid thing, and they answer a different question. Schutz and colleagues' 2002 reference data describe ordinary adults rather than trained ones, which makes them the right comparison if you are asking whether you have enough muscle for health, and the wrong comparison if you are asking whether you are muscular by gym standards. Both questions are legitimate; they just need different rulers.

The natural limit of 25, handled honestly

The claim that a normalised FFMI above 25 means drugs is the most repeated thing about this index and the most overstated. Here is what the source actually says. Kouri's team computed the index for 74 athletes who reported never using steroids and found their normalised values extended up to a well-defined limit of 25.0, while many steroid users exceeded it and some passed 30. The authors described their own findings as preliminary and proposed the index as a screening measure, not a test.

Two details that usually get dropped change the reading considerably. The same paper estimated the normalised index for twenty Mr America winners from the pre-steroid era of 1939 to 1959 and found a mean of 25.4, above the supposed ceiling, in men competing decades before anabolic steroids were available. And a 2024 analysis of a large sample of collegiate American football players reported 21 per cent of them above a raw FFMI of 25, with values ranging to 27.7.

So the honest statement is this: a normalised FFMI above 25 is uncommon without pharmacological help, and genuinely rare much above 26, but it is a population observation with real outliers rather than a physical law. It is also computed from a body fat estimate that can easily be two or three points out, which moves the index by close to a point on its own. Nobody should be accused, or accuse themselves, on the strength of a calculator result.

How much of your FFMI is really measurement error

FFMI is not measured. Your height and weight are; everything else is derived from a body fat percentage that is itself an estimate, and the index magnifies whatever error that estimate carries. Skinfold equations validated against a four-compartment model sit around three percentage points of error in good hands; the tape-based Navy method is nearer four; consumer bioimpedance scales move with hydration and can be further out again.

Work the arithmetic through and the effect is easy to state. For an 82 kg person at 1.78 m, moving the body fat figure from 15 per cent to 11.5 per cent takes normalised FFMI from 22.1 to 23.0; moving it to 18.5 per cent takes it down to 21.2. That is a swing of about ±0.9 index points from the body fat estimate alone, enough to cross a band boundary in either direction without a single gram of muscle changing hands. The result panel above does this calculation on your own numbers so you can see the size of it rather than take it on trust.

The practical response is not to despair of the number but to use it properly. Pick one body fat method and stay with it, measure under the same conditions each time, and read FFMI as a slow trend line rather than a label. Systematic error largely cancels when you compare a reading against your own earlier readings taken the same way; it does not cancel at all when you compare your caliper number against somebody else's DEXA number on the internet.

• Changing measurement method resets your baseline; treat it as starting a new series, not continuing the old one.
• A body fat figure you guessed produces an FFMI you guessed, dressed up to one decimal place.
• Three or four readings across six months tell you more than any single reading ever will.

How FFMI changes with training over years

The index moves slowly, and knowing how slowly is a useful defence against disappointment. An untrained man starting around 18 might reach 20 to 21 in his first serious year, because beginners gain lean mass fastest, perhaps eight to twelve kilograms in twelve months under good conditions. The second year typically delivers half of that, the third less again, and by the fourth or fifth year an honest annual gain is one to two kilograms of lean tissue, which is a quarter to a half of an index point.

That is why chasing a target number is the wrong use of FFMI. A jump of a full point inside three months is almost always the body fat estimate moving, or glycogen and water, or both. The genuine signal is a line that drifts upward across years while your measurement method stays constant.

It is also why FFMI is the better progress metric during a cut than scale weight. In a well-run deficit, fat mass falls and lean mass holds, so FFMI stays flat while the scale drops and body fat percentage falls. If FFMI is falling too, the deficit is taking lean tissue with it: the signal to ease the deficit, lift the protein, and keep training heavy rather than to push harder.

Who FFMI misreads

Very tall and very short people are the first group. The height normalisation is a linear correction fitted around a population of ordinary adult heights; it is not validated at the far tails, so someone at 1.50 m or 2.05 m should treat the normalised figure with more scepticism than the raw one and rely mostly on their own trend.

People with a lot of fat mass are the second. At high body fat, every method of estimating body fat percentage becomes less reliable, so the lean mass that FFMI is built on becomes less reliable too. And the index, being indifferent to fat, can read perfectly average for someone whose health risks are anything but. FFMI is a muscularity index, not a health screen; the waist-based measures answer that question and should be read alongside it, never replaced by it.

Anyone carrying unusual fluid is the third group: pregnancy, oedema, kidney or heart failure, or simply a very high-carbohydrate few days before measuring. Extra water counts as lean mass in every method here and inflates the index accordingly. And finally the index says nothing about where your muscle is, how strong it is, or whether it is trained: two people at an identical FFMI of 22 can have completely different physiques and completely different logbooks.

What to do with your number

If your index is below average for your sex and you want to change that, the levers are unglamorous and well established: resistance training two to four times a week with progressive load, protein in the region of 1.6 to 2.2 grams per kilogram of bodyweight, a small energy surplus if you are already lean, and enough sleep to recover from any of it. Nothing about FFMI changes that advice; it just measures whether it is working.

If your index is already high and you are trying to lose fat, the job is to keep the index flat while the scale falls. Watch lean mass rather than weight, keep the deficit moderate and the protein high, and expect FFMI to hold rather than climb. Gaining lean mass in a deficit is realistic mostly for beginners and people returning after a layoff.

And if the number surprised you in either direction, check the input before believing the output. Re-measure your body fat by the same method a week later, look at what the ±3.5 point panel above does to your index, and remember that the honest version of your result is a range about two points wide, not a figure with a decimal place.

How it's calculated

Worked example: a man of 178 cm and 82 kg at 15% body fat

1. Fat mass: 82 × 15 ÷ 100 = 12.3 kg. Lean mass: 82 − 12.3 = 69.7 kg.
2. Height squared: 1.78 × 1.78 = 3.1684 m².
3. Raw FFMI: 69.7 ÷ 3.1684 = 22.0 kg/m².
4. Height normalisation: he is 2 cm under the 1.80 m reference, so add 6.1 × (1.80 − 1.78) = 6.1 × 0.02 = 0.12. Normalised FFMI = 22.1.
5. Read the band: 22.1 sits just inside the 22–25 'muscular' band for men. That is years of training, and a long way short of the 25 ceiling.
6. Now propagate the error. If his real body fat were 11.5% instead, lean mass is 82 × 0.885 = 72.6 kg, raw FFMI 22.9 and normalised 23.0.
7. And if it were 18.5%, lean mass is 82 × 0.815 = 66.8 kg, raw FFMI 21.1 and normalised 21.2, back in the 'above average' band.
8. State it honestly: a caliper reading of 15% means a normalised FFMI of about 22.1 ± 0.9, or somewhere between 21.2 and 23.0. The band label changed twice in that range while his body did not change at all.

Where this number is used in the real world

• Training progress tracking, where FFMI is the one figure that separates muscle gain from scale-weight noise over years.
• Body composition assessment in sport, as a height-fair way to compare athletes of very different statures in the same squad.
• Screening for possible anabolic steroid use in athletic, medical or forensic settings, the purpose the index was originally proposed for.
• Nutritional assessment and sarcopenia screening in clinical practice, where a low fat-free mass index for age flags muscle deficit that BMI hides.
• Chronic disease research, where fat-free mass index predicts outcomes in conditions such as COPD and chronic kidney disease better than body weight.
• Correcting a misleading BMI, for muscular people whose weight-based reading classes them as overweight without any excess fat.
• Setting realistic goals, by comparing your current index against what published data say is actually attainable without drugs.

Frequently asked questions

Keep going

A single number rarely tells the whole story. Alongside the FFMI result, the body fat calculator, the skinfold calculator, the lean body mass calculator, the BMI calculator and the BMR calculator each add a different angle on the same measurements. For the reasoning behind the numbers, read BMI for athletes, Body fat methods and Track without DEXA.

Sources

1. Kouri EM, Pope HG, Katz DL, Oliva P. Fat-free mass index in users and nonusers of anabolic-androgenic steroids. Clin J Sport Med 1995;5:223–8. doi.org/10.1097/00042752-199510000-00003
2. Schutz Y, Kyle UUG, Pichard C. Fat-free mass index and fat mass index percentiles in Caucasians aged 18–98 y. Int J Obes 2002;26:953–60. doi.org/10.1038/sj.ijo.0802037
3. Fields JB, Jones MT, Kuhlman NM, Magee MK, Feit A, Jagim AR. Fat-free mass index in a large sample of collegiate American football athletes. Int J Exerc Sci 2024;17:129–39. doi.org/10.70252/FGRL8917
4. Peterson MJ, Czerwinski SA, Siervogel RM. Development and validation of skinfold-thickness prediction equations with a 4-compartment model. Am J Clin Nutr 2003;77:1186–91. doi.org/10.1093/ajcn/77.5.1186
5. Boer P. Estimated lean body mass as an index for normalization of body fluid volumes in humans. Am J Physiol 1984;247:F632–6. doi.org/10.1152/ajprenal.1984.247.4.F632