Body composition is the split between fat mass and everything else — muscle, bone, organs, water. It is the number most people actually care about when they say they want to lose weight. It is also the number that is hardest to measure well, and the consumer methods most of us use are wrong by margins large enough to hide months of real progress.
This article is about measurement and expectations: what the methods do, how far off they are, how fast composition can plausibly change, and why the bathroom scale keeps telling people the wrong story.
What body composition actually is
The simplest useful model splits the body into two compartments: fat mass and fat-free mass. Fat-free mass includes skeletal muscle, but also bone mineral, organs, connective tissue and — importantly for anyone reading a scale — several liters of water.
Research-grade methods go further, into three- and four-compartment models that separate body water and bone mineral from the rest. Every method below is an estimate of those compartments, not a direct count. They differ in what they assume.
An earlier version of this article told readers to aim for carbohydrate at 35 to 65 percent of intake, protein at 10 to 35 percent and fat at 20 to 35 percent, with the advice to push protein up and fat down to lose weight. Percentages of an unspecified total are close to meaningless: 25 percent protein is a very different amount of protein at 1,600 calories than at 3,200. Protein requirements for preserving muscle are expressed per kilogram of body weight, not as a share of the plate.
The four methods people actually use
Only one of these is a reference method, and it is not the one in your bathroom.
| Method | What it measures | Practical accuracy |
|---|---|---|
| DEXA (DXA) | X-ray attenuation, split into fat, lean soft tissue and bone | The usual field reference. Precise when the protocol is controlled; results shift with hydration, food and machine. |
| Skinfold calipers | Subcutaneous fat at 3 to 7 sites, converted by equation | Depends heavily on the tester and on whether the equation fits the person. |
| Bioelectrical impedance (BIA) | Resistance to a small current, converted by proprietary equation | The weakest of the four, and the most common. See below. |
| Tape measure | Waist, hip and limb circumference | Measures no compartment at all, but is cheap, repeatable and tracks change usefully. |
A 2015 methodology review in the International Journal of Sport Nutrition and Exercise Metabolism set out how much a DEXA result depends on protocol: subjects rested, fasted overnight, in minimal clothing, positioned identically on the bed each time. Without that standardization, the review argued, the machine’s precision is largely wasted.
How far off the consumer methods are
The size of the error is the part nobody puts on the box.
A 2018 retrospective study in PLoS One compared DEXA and bioelectrical impedance across 3,655 paired measurements in a hospital nutrition unit. For people with a BMI above 18.5 and below 40, impedance overestimated fat-free mass by 3.38 to 8.28 kg and underestimated fat mass by 2.51 to 5.67 kg compared with DEXA. Agreement was better for fat mass than for fat-free mass, and worst at the extremes of BMI.
Eight kilograms is not a rounding error. It is more lean mass than most people will gain in several years of training.
Calipers are not obviously better. A 2010 validation study in the Journal of Strength and Conditioning Research tested two commercial calipers, two impedance devices and a three-site skinfold equation against DEXA in 50 college-aged adults, and found the consumer devices predicted percent body fat poorly. A 2014 study of 65 soldiers in Military Medicine reached a similar conclusion: single-frequency impedance was the most reliable field method, which is not the same as being the most accurate.
The practical implication is narrow but important. Any of these methods can track a direction over time if you use the same device, the same way, at the same time of day. None of them can tell you your true body fat percentage.
Why scale weight misleads
Day-to-day and week-to-week weight is mostly not fat.
A 2017 analysis in Physiological Reports tracked two-week weight changes in 46 free-living adults who were, on average, in energy balance. A unit of short-term weight change was composed of about 84% fat-free mass — largely body water. The energy density of those 1 to 3 kg swings averaged 2,380 kcal per kg, nowhere near the roughly 7,700 kcal per kg of stored fat.
Glycogen binds water. Sodium intake moves water. Menstrual cycle phase moves water. A large meal is still in the gut the next morning. Two kilograms can appear and vanish over a weekend without a gram of fat changing hands.
The longer-run picture is also slower than people expect. The 2011 modelling work published in The Lancet found that body weight responds to a sustained change in intake with a half-time of about a year, which means a genuine change in eating takes months to fully show up on a scale.
What a realistic rate of change looks like
Fat loss and muscle gain run on different clocks, and both are slower than marketing implies.
For fat loss with lean mass preserved, the most useful trial is a 2011 randomized comparison in the International Journal of Sport Nutrition and Exercise Metabolism. Twenty-four elite athletes lost weight at either 0.7% or 1.4% of body weight per week, both while strength training. Both groups lost a similar percentage of body weight. The slower group increased lean body mass by 2.1%; the faster group did not change it at all.
For muscle gain, the 2018 meta-analysis in the British Journal of Sports Medicine — 49 trials, 1,863 participants — reported average fat-free mass gains measured in fractions of a kilogram over trials lasting weeks to months, larger in beginners and smaller with age. And in the eight-month STRRIDE AT/RT trial, resistance training raised lean mass while aerobic training lowered fat mass, with neither doing much of the other’s job.
Roughly: half a percent to one percent of body weight per week is a defensible fat-loss rate. Visible change in lean mass is a matter of months. A four-week verdict on either is noise.
The limits of what any of this can tell you
Even DEXA is a model, not a measurement of fat. It infers tissue composition from how two X-ray energies are absorbed, and its estimates move with hydration status and with the manufacturer’s software. Two machines can disagree by several percentage points on the same person on the same day.
Reference ranges are also thinner than they look. Population norms for percent body fat vary by sex, age and ancestry, and most were built from small samples. There is no validated “ideal” body fat percentage for an individual.
And composition is not the only thing that matters, or even the most informative. Where fat sits appears to carry more information than how much there is, which is the argument for the humble tape measure covered in our piece on waist-to-height ratio.
How to track without fooling yourself
A method that is consistently wrong in the same direction is more useful than a method that is occasionally right.
- Weigh at the same time each morning, after the bathroom, before eating, and use a seven-day rolling average rather than any single day.
- Take waist circumference at the same landmark, at the end of a normal exhale, once a week.
- If using an impedance scale, use the same one under the same conditions, and read only the trend line. Ignore the absolute percentage.
- Track training performance. Load, reps and pace are direct evidence that lean mass is being maintained, and they need no equation.
- Judge over 8 to 12 weeks. Anything shorter is water.
What you do between measurements is the part that moves the number. The evidence on that sits in our companion articles on what actually drives fat loss and muscle gain, on calories and energy balance, and on the physical activity dose the guidelines ask for.
Common questions
Are the body fat scales sold for home use accurate?
Not for an absolute number. The 2018 PLoS One comparison found impedance devices misestimating fat-free mass by several kilograms against DEXA. They can be reasonable at showing direction over months if the conditions are kept identical, which in practice means same scale, same time of day, similar hydration.
How often should body composition be measured?
Every 8 to 12 weeks is enough for any method. Measuring more often mostly samples water, and the noise is larger than the signal you are looking for.
Why did I lose two kilograms in three days?
Almost certainly water and gut contents. The 2017 Physiological Reports analysis found short-term weight change was about 84 percent fat-free mass, with an energy density far below that of body fat. Rapid drops reverse just as fast.
Is BMI useless?
It is a poor individual measure of body composition, because it cannot distinguish muscle from fat. It remains useful at population scale and as a rough screen. For an individual, waist circumference or waist-to-height ratio carries more information for the effort involved.
What to take from this
Pick one method, use it identically every time, and judge it over three months rather than three days. The absolute number your device reports is probably wrong by more than the change you are trying to detect; the direction of travel is what it can honestly give you.
Diet quality is one input among several, and a green powder is a convenient way to add plant food to a day rather than a way to change a measurement. That is the case we make for Greens Plus, and we would rather make the smaller true one.
About the Author
