What FFMI measures
Fat-Free Mass Index is BMI’s pickier cousin. Where BMI just divides total weight by height squared, FFMI replaces total weight with lean body mass (everything that isn’t fat). The result is a number that scales with how much muscle a body carries, normalized for how tall it is.
A 70 kg lifter at 8% body fat carries 64.4 kg of lean mass. A 70 kg lifter at 25% body fat carries 52.5 kg. Same BMI, different FFMIs (and different physiques). FFMI captures that distinction in one number.
Formula
LBM = weight × (1 − body fat %)
Raw FFMI = LBM / height_m²
Adjusted = Raw FFMI + 6.1 × (1.8 − height_m)
The height adjustment is from Kouri 1995. Without it, tall lifters score artificially low and short lifters score artificially high. The 1.8m anchor and 6.1 coefficient are empirical fits to a sample of 157 male athletes.
Worked example
A 180 cm, 80 kg male at 15% body fat:
LBM = 80 × (1 − 0.15) = 68 kg
Raw FFMI = 68 / 1.80² = 68 / 3.24 = 20.99
Adj FFMI = 20.99 + 6.1 × (1.8 − 1.8) = 20.99 + 0 = 20.99
The same lifter at 175 cm (slightly shorter):
LBM = 80 × (1 − 0.15) = 68 kg
Raw FFMI = 68 / 1.75² = 68 / 3.06 = 22.21
Adj FFMI = 22.21 + 6.1 × (1.8 − 1.75) = 22.21 + 0.305 = 22.51
Same body, lower height, FFMI shifts up. The adjustment normalises this so a 175 cm and a 185 cm lifter at the same lean mass ratio score similarly.
The natural-ceiling reference
Kouri’s 1995 paper compared FFMI distributions in self-reported anabolic steroid users vs non-users. Non-users clustered at FFMI 21-23. Users clustered at 25-27. The 25.0 line (“the natural ceiling”) got popularised as the suggested upper boundary for natural lifters.
What the line means and doesn’t mean:
- It is not a hard ceiling. There are documented natural lifters at FFMI 26-26.5. Outliers exist.
- It is a central tendency based on a 157-subject sample. Larger samples since suggest the natural distribution may have a longer tail than Kouri proposed.
- It assumes accurate body-fat measurement. If your body fat is off by 3 percentage points, your FFMI estimate is off by ~0.5-0.8 points. Not nothing.
- For females, the equivalent suggested ceiling is ~21.5, with similar caveats.
For a coach using FFMI: treat it as a useful framing for the question “how trained is this client” but don’t draw line-in-the-sand conclusions. Body fat measurement error alone wipes out half a point.
Body-fat input quality matters more than the formula
The most common mistake is using a bathroom-scale BIA reading for body fat, then taking FFMI to two decimal places. Bathroom scales are routinely 5-10 percentage points off. That’s a 1.5-3 point swing on the FFMI output, which moves you between “average natural” and “approaching ceiling” without any actual body composition change.
Better body-fat sources, in order of accuracy:
- DEXA scan. Gold standard for body composition. Often available at universities for $50-100.
- Hydrostatic weighing or Bod Pod. Accurate, less common.
- Navy method (circumferences). Within ~3% of DEXA for most populations. Free and replicable at home. See our Navy body-fat calculator.
- Skinfolds with calipers. Accurate if the same trained measurer takes consistent sites.
- BIA scale. Last resort. Read it as ±5%.
When FFMI is useful
For a working coach:
- Tracking lean mass changes during a cut: weight goes down, FFMI ideally stays flat (you lose fat, not muscle).
- Setting realistic goals: a male client at FFMI 20 wanting to reach 24 in 18 months is asking for ~3-4 kg of lean mass over 18 months. Within natural range and ambitious.
- Reality-checking promises: a coaching client at FFMI 23 asking how to get to 27 in a year needs the conversation about what that progression typically requires.
For a self-tracker: it’s a once-a-quarter check-in number, not a daily metric.
How Coach Sheet uses FFMI
The Profile tab takes Weight, Height, Body Fat % once. The Quick Stats tab below auto-calculates Adjusted FFMI alongside BMR and TDEE. Update the body fat measurement quarterly, watch the FFMI trend independently of weight (since cutting cycles drop weight without dropping FFMI).
Sources
- [1]
- [2]
- [3]