/

Training

Why your workout calorie count is probably wrong

Patricia Bedoya

·

·

6

min read

A person glancing at their watch after a workout, towel over one shoulder, kitchen behind them in soft daylight (AI-generated image)

Cover image generated with AI — © Chubby Studio S.L.

KEY TAKEAWAY

Wrist wearables measure heart rate to within a few percent but estimate calorie burn far less reliably: in one seven device comparison, no device came within 20 percent of the true value. Treat the calorie figure as a rough, consistently biased score for comparing your own sessions, not as an accurate measure of energy burned.

Short answer: the heart rate on your wrist is good, the calorie number next to it is not. In the best known head to head test of seven wrist wearables, not one got its energy estimate within 20 percent of the truth, while the same devices read heart rate to within a few percent. Nothing since has closed that gap, because the two numbers are produced in completely different ways.

That does not make the calorie figure useless. It makes it useful for something narrower than most people assume.

How does your watch work out calories at all?

Your watch does not measure energy. Measuring energy expenditure properly means capturing the air you breathe, in a lab, with a mask on your face, or dosing you with labelled water and testing you for a fortnight. A watch has an optical pulse sensor, an accelerometer, a gyroscope and some GPS, and from those it runs a model.

The model is fed by the things you typed in when you set the watch up. Apple's own documentation is explicit that it uses your height, weight, sex, age and wheelchair status to calculate how many calories you burn, and that the watch keeps learning your fitness level the more you use the Workout app. Every calorie figure you have ever seen is an estimate built on top of another estimate, and the first one is a form you filled in once, probably years ago.

How wrong is it, really?

In 2017 a Stanford group put seven wrist devices on 60 people and compared them against clinical grade equipment across walking, running, cycling and sitting. Heart rate held up well: six of the seven kept median error below 5 percent while cycling, and the best device managed 2.0 percent overall. Energy expenditure did not hold up at all. Median error across tasks ran from 27.4 percent for the best device to 92.6 percent for the worst, and no device in that study got within 20 percent of the true value.

A separate 2018 validation reported the same split from the other direction: heart rate errors of roughly 1 to 8 percent, and energy expenditure errors reaching 61.8 percent on one device depending on the activity. The authors put it plainly, that these devices give a usable heart rate for everyday activity and general exercise, but that energy expenditure from them may be significantly over or underestimated.

Those studies are several hardware generations old and today's watches are better. The structure of the problem has not changed, though. A 2020 meta analysis of 40 wrist and arm worn monitors found accuracy still depended heavily on the activity being done, with very large disagreement between studies of the same device. Adding heart rate to motion sensing improved things, which tells you the direction of travel, not that the destination has been reached.

See what a workout actually cost you

FitWoody analyses every session against your own cardiac load instead of a generic calorie formula, so you can see how hard it really was for you. 7-day free trial of Plus.

Download the app

★★★★★

4.5 · 732 ratings on the App Store

Why is the pulse fine but the calories are not?

Because your pulse is measured and your calorie burn is inferred. Light goes into your wrist, some of it comes back, and the shape of that signal is a real physical property of the blood underneath. There is a correct answer and the sensor is trying to read it.

Calories have no equivalent. The watch takes your pulse and your movement, adds the profile you typed in, and asks a statistical model what someone like you would probably burn doing this. Two people with an almost identical heart rate trace can burn genuinely different amounts, which is the same reason the same workout costs two people different amounts. The model knows the average of people like you. It does not know you.

This is also why the error is not evenly spread. The Stanford group found error was higher for men, at higher body mass index, for darker skin tones, and for walking rather than cycling. An error that tracks who you are is worse than a random one, because it does not average out over a month. It just quietly follows you.

What makes your own number worse than average?

Some of it is genuinely fixable. An out of date weight in the Health app feeds straight into the model. Never calibrating means the watch has less idea what your walking and running actually look like: Apple suggests about 20 minutes outdoors at your normal pace with good GPS reception, which improves distance, pace and calorie accuracy. Picking the wrong workout type applies the wrong model entirely.

Then there is the part you cannot fix. Wrist worn optical sensors have a harder time when the wrist is doing little and the rest of you is working, so lifting, rowing and anything grip heavy is tougher for them than steady cycling. Calibration cannot fix a structural problem, but it does fix the part that is yours to fix.

So what is the calorie number good for?

It is a poor absolute measure and a decent relative one. If your watch reads 30 percent high, it reads roughly 30 percent high on Tuesday and on Saturday too, so the comparison between your Tuesday and your Saturday still holds. That is the useful case, and it is the whole useful case. Compare today with your own last month, and not with the person next to you, whose watch is wrong in a different direction.

It follows that a calorie total is a weak basis for eating decisions and a reasonable basis for effort decisions. If you want a number that says how hard a session actually was for you, training load and your cardiac response carry far more signal than an energy estimate. And if you want a daily movement target, a goal that adapts to how recovered you are beats a fixed calorie target that treats every day as identical.

One distinction is worth clearing up here, because it accounts for a lot of confused comparisons. Active calories are what the watch thinks you burned through movement. Total calories add your basal metabolic rate, the energy you would have spent lying still all day. The move ring counts active calories only, which is why the ring number looks small next to a treadmill readout that quietly includes the resting portion. Neither is more correct. They answer different questions, and comparing one against the other is the fastest way to conclude your watch is broken when it is not.

What to watch instead

The numbers on your wrist that come closest to being measured rather than modelled are the heart rate derived ones. Resting heart rate, heart rate variability and how quickly your heart rate falls after a hard effort all rest on the sensor that actually works well. So does your cardio fitness estimate, which is modelled too, but from a much better behaved input.

None of this means you should ignore the move ring. Closing it is a good habit and the habit is most of the point. Just hold the number loosely. It is a rough score for showing up, not an accounting of your body's fuel.

Frequently asked questions

Are Apple Watch calories accurate?

Not in absolute terms. Validation studies of wrist wearables have repeatedly found energy expenditure errors well above 20 percent, even when the same devices read heart rate to within a few percent. The calorie figure is best used to compare your own sessions with each other, not as a true measure of energy burned.

Should I plan my eating around the calories my watch reports?

The measurement is not precise enough to build a food plan on, and small daily errors compound quickly. FitWoody does not count food calories and does not give nutrition advice. If you want to change your eating, a registered dietitian is the right place to start.

Does calibrating an Apple Watch improve calorie accuracy?

It helps. Apple recommends about 20 minutes of outdoor walking or running at your normal pace with good GPS reception, which improves distance, pace and calorie accuracy. Keeping your height, weight, age and sex up to date in the Health app matters just as much, because those values feed the estimate directly.

Sources

Keep reading