Body composition: what the number can tell you, and what it cannot

A fighter gets a body fat reading of 15 percent and uses it to decide whether he can drop a division. Depending on the method, the true figure for a 70-kilogram athlete could be anywhere from about 13.6 to 16.4 percent. The error is larger than the decision he is making with it.

Measure anyway, and know what you are holding when you do.

The pseudo accuracy problem

Every method reports a clean figure to one decimal place. None of them deserves that confidence.

Take the worked example. A 70-kilogram athlete measured at 15 percent body fat. Applying the error range that impedance methods carry, his actual fat mass sits somewhere between roughly 9.5 and 11.5 kilograms, which is a body fat percentage somewhere between 13.6 and 16.4.

For a 50-kilogram athlete the same error produces a wider spread still, roughly 13 to 17 percent.

Nobody reads the number that way. It arrives as "15", it gets written down as 15, and a division change gets planned on it.

The four methods, and what each one is actually doing

Two things in that table are worth pulling out, because they change what you should do.

Impedance during a cut is worse than no measurement

BIA works by passing a current through the body and inferring composition from the resistance. Water conducts. So the reading is a function of how much water you are carrying and where.

Fluctuations in hydration alone can move a body fat estimate by two to three percentage points. That is the entire range most fighters are trying to detect over a whole camp, produced by nothing more than whether you drank recently.

There is a second complication with consumer devices specifically. A study of impedance devices against DXA found mean absolute error under 2 percent for one unit and no systematic bias, and still concluded that the devices should not be used interchangeably with criterion measures because the error limits at the individual level were unacceptably wide.

A good average across a group tells you nothing about your reading.

The finding almost nobody knows

Standardised B-mode ultrasound has a property none of the others has.

In a study that deliberately manipulated food and fluid intake, the sum of eight subcutaneous adipose tissue thicknesses measured by ultrasound was unaffected by acute food and fluid consumption or by changes in hydration status.

The same interventions altered DXA estimates, particularly the lean mass component.

Read that against a sport where the athlete is dehydrated for a fortnight out of every camp. A method that does not care about hydration is worth a great deal more here than in a sport where the athlete is always in a normal state.

The same work found ultrasound more accurate than skinfolds at both ends of the range, in leaner athletes around 10 to 15 percent and in higher-fat athletes. The cost is real: roughly 20 minutes per test and equipment an order of magnitude more expensive than callipers, which is why you will not find it in most gyms.

What to actually do

1. Choose one method and never mix. A caliper number and a DXA number are not comparable, and switching between them produces a change that is entirely artefact. Most fighters who panic about "gaining fat" have simply changed measurement method.
2. Standardise the conditions. Morning, fasted, before training, after using the bathroom, same room, same operator. This costs nothing and removes most of the day-to-day noise.
3. Measure out of camp, not in it. Pre-camp and post-fight are the two windows where you are in a normal hydration state. Anything measured in the final 3 weeks is measuring your cut, not your composition.
4. Read the trend, not the reading. Three measurements over 8 weeks under identical conditions tell you something real. One measurement tells you almost nothing.

What the number is actually for

This is the decision it is supposed to inform, and precision matters here.

A fighter wants to know how much of his walk-around weight is fat, because that is the part he can lose without losing performance.

Everything below that is muscle, water and organs, and cutting into it is where the damage happens.

That is a legitimate question. The problem is that the answer arrives with an error band wide enough to cover the difference between a comfortable cut and a dangerous one. Which is why the division decision should be made on several inputs rather than one number: your walk-around weight, the gap to the limit as a percentage, how many weeks you have, and what your last 3 camps actually looked like.

The arithmetic is in what weight class should you fight at and the ranges in the calculator.

The honest version of the sales pitch

Body composition testing is sold as precision. What it actually gives you is a repeatable estimate, and its value comes almost entirely from repeating it under identical conditions rather than from any single result.

A fighter with three consistent caliper readings taken by the same person across a camp knows more than a fighter with one DXA scan and no comparison.

The cheap method used properly beats the expensive method used once.

And if the number you are about to pay for is going to be used to decide a division, the measurement is the smallest part of that decision.

The rest of it is the conversation about what your last camp cost you.