Body composition calculator
Body fat scale accuracy calculator
Your body fat reading, plus the error the method actually carries. Find out whether that number can tell you what you think it can. Free, instant, no sign-up.
Your reading's error crosses a category boundary, so the category itself is not information — the honest answer is that you could be in either.
One reading is noise. The coach works from the trend, which is the part that means something.
Why every body fat number has an error bar
No consumer method measures body fat. Every one of them measures something else — how much current passes through you, how thick a fold of skin is, how much water you displace — and runs it through an equation fitted on a sample of other people. The gap between that estimate and the truth is the standard error, and for most methods it is large relative to the differences people actually care about. A smart scale carries roughly ±5 percentage points. The gap between "fitness" and "average" on the usual descriptor tables is four. That arithmetic is the entire point of this page.
What the error figures are
DEXA is quoted at about ±1.5 points, hydrostatic weighing ±2.5, the Bod Pod ±2.7, skinfold calipers and the Navy tape method about ±3.5, handheld BIA around ±4.5, and a foot-to-foot smart scale about ±5. These are the method's error under decent conditions — a skinfold taken by a practised tester with calibrated calipers, a scale used properly hydrated. Operator error and bad timing sit on top of them. DEXA deserves a caveat of its own: its precision is excellent, meaning a repeat scan gives nearly the same answer, but precision and accuracy are different properties, and two machines from different manufacturers can disagree about the same person by more than the figure above.
Why your scale jumps overnight
Because two separate readings each carry the full error, a change has to clear roughly twice it before it means anything at all. A smart scale that reports 18.2 % one morning and 20.6 % the next has not detected two kilograms of fat appearing overnight, which is physiologically impossible — it has detected how much water was in your legs. Foot-to-foot devices pass a current up one leg and down the other, so they sample the lower body and are unusually sensitive to hydration, recent food, alcohol, salt and whether you trained yesterday. Weighing in first thing, fasted, before drinking, is not superstition; it is the only way to hold the systematic part of the error constant.
So what is a body fat measurement good for?
Trend, not verdict. Measured the same way, at the same time of day, in the same state, most of the error is systematic — it is the same wrongness every time, so it largely cancels out of the difference between readings. A scale that reads three points high still shows you a real downward trend over eight weeks. What it cannot do is tell you whether you are at 17 % or 22 % today, and no amount of re-weighing will fix that, because re-weighing measures the same thing the same wrong way. If you need an accurate single number — for a competition class, a research protocol, a medical decision — you need a laboratory method, and even then you should ask which one and on whose machine.