The four numbers, and how they connect
Body composition is really one measurement and three things derived from it. Body fat percentage is the measurement — everything else on this page is arithmetic performed on it. Lean body mass is bodyweight minus fat mass, and it is the figure that actually matters for setting a protein target and for the one BMR equation that knows anything about composition. FFMI divides lean mass by height squared, so it answers a different question: not how much muscle you have, but how much you carry for your frame. And recomposition runs the whole thing backwards — hold lean mass constant, pick a target body fat, and solve for the weight that gets you there.
If you only measure one thing, measure body fat with a tape, the same way, every four weeks. The other three fall out of it, and none of them can be better than it is.
Measure the trend, not the reading
Every method here is more reliable at tracking change than at pinning down an absolute figure. Measured the same way, at the same time of day, before eating, with the same tape, a circumference estimate will show you a real trend over a month. Compared against a smart scale, a DEXA scan or a friend's calculator, it will disagree — and so will they with each other.
The practical consequence is that a body-fat number is most useful as a series. One reading of 21% tells you very little, because the method's own error is around three percentage points. Six readings taken the same way over three months, trending from 24 to 21, tell you the thing you actually wanted to know, and the error largely cancels because it is consistent rather than random.
The error bars, in one place
The US Navy tape method carries roughly ±3 percentage points against hydrostatic weighing, and it is more accurate at average body compositions than at the extremes. The three lean-mass regressions — Boer, James and Hume — disagree with each other by around 8% on the same person, which at 80 kg is about 5 kg of claimed lean tissue. Bioimpedance scales are highly sensitive to hydration and can move several points between morning and evening on an unchanged body. Even DEXA, the usual reference standard, has a meaningful test-retest range and different machines do not agree with each other.
None of that makes the numbers useless. It makes them estimates, which is what they are called on every page here. A calculator that prints 17.4% without saying that the real figure is somewhere between 14 and 20 is not more accurate — it is just quieter about the same uncertainty.
Why there is no BMI calculator on this site
Two reasons, and the first is commercial rather than principled: Google answers “BMI calculator” with its own inline widget, so a page targeting it earns essentially nothing. The second reason is that BMI is a ratio of weight to height squared with no information about composition in it at all. It cannot tell a lifter from someone carrying the same weight as fat, which is the exact distinction every tool on this page exists to make.
What this site will not tell you
Whether your body fat percentage is healthy. That is a medical judgement about a specific person, and a web page applying a threshold table to one number is not qualified to make it. These pages report what the method returns and how far out it can be. What that means for you is a conversation with someone who can see you.
You will also not find goal weights presented as targets, before-and-after framing, or any category label attached to a number. The FFMI page prints orientation bands because the figure is meaningless without some sense of scale, and it says in the same breath that they come from small samples and are not standards.