Kivorent News

Thursday, November 19, 2026

Health techExplainer

How smartwatches and fitness trackers estimate the calories you burn

A smartwatch never measures calories. It estimates them from your age, weight and height plus motion and heart rate data, and that estimate is the least accurate number on the screen.

Pulse, covering wearables and health sensors
, 3 min read

27-93%average calorie error across 7 wrist devices, Stanford 2017

What to know

  • No wrist sensor can detect calories directly; every figure is a model's estimate.
  • Resting calories come from a formula using age, sex, height and weight.
  • Active calories are inferred from motion and from how fast the heart is beating.

Nothing on your wrist can see a calorie. Burning food energy shows up as oxygen breathed in and carbon dioxide breathed out, and the only way to count it properly is to measure those gases. Labs do that with a mask or a sealed room, a method called indirect calorimetry; studies of daily living use a drink of "doubly labeled water" that tracks how fast the body turns over oxygen and hydrogen. A watch has neither, so the calorie figure it shows is a model stacked on other models.

Two numbers added together

Most watches split the day's total into a resting part and an active part. Apple labels them Resting Energy and Active Energy; other brands use similar names.

The resting part is mostly arithmetic. It starts with an equation for basal metabolic rate, the energy the body uses lying still to keep organs running. A widely used one, Mifflin-St Jeor, needs only weight, height, age and sex. For a 40-year-old man who weighs 80 kg and stands 180 cm tall, it works out like this: 10 × 80 + 6.25 × 180 − 5 × 40 + 5 = 1,730 kcal a day. For a woman the equation subtracts 161 instead of adding 5. Notice what the watch needs from you here: the numbers you typed into its profile. If your weight in the app is two years old, the resting estimate is too.

The active part is where the sensors come in.

Motion, then heart rate

The first input is the accelerometer, the same chip that counts your steps. More movement, faster, means more estimated work. That works reasonably for walking and running, where arm swing tracks effort. It falls apart for cycling, rowing machines, carrying groceries or pushing a stroller, where the wrist moves little or oddly compared with what the legs and back are doing.

The second input is heart rate from the optical sensor on the back of the watch. Across moderate to hard exercise, heart rate rises roughly in step with oxygen use, and each litre of oxygen corresponds to about 5 kcal. So a model that knows your heart rate, and how your heart rate relates to oxygen use, can estimate energy spent.

That relationship is personal. A fit person moves more oxygen per heartbeat than an unfit one, which is why watches ask for your maximum heart rate or estimate your VO2 max. And heart rate climbs for reasons that burn few extra calories: heat, dehydration, caffeine, a stressful meeting. At low intensity, when heart rate barely moves, the models lean back on motion and the link to oxygen use is weakest.

How far off it gets

The best-known test is a 2017 Stanford study that put seven wrist devices on 60 volunteers and compared them with lab gas measurements. Heart rate was mostly close. Calories were not: the most accurate device was off by an average of 27%, and the least accurate by 93%. Later reviews of newer devices have broadly found the same pattern, with heart rate and step counts far more trustworthy than energy expenditure.

Put that 27% in workout terms. On a session that truly burned 500 kcal, an average miss of 27% is about 135 kcal (500 × 0.27), and that was the best device in the test.

The estimate also misses things by design. Digesting food uses roughly a tenth of daily energy, and none of it shows up as motion. Small fidgeting while seated barely registers.

What the number is good for

The figure is most consistent against itself: the same watch, worn the same way, on the same person, tends to make the same errors day after day. That makes it a rough gauge of whether one day was more active than another, the way a resting heart rate trend is more telling than any single reading. It is not a precise count of energy, and calorie estimates are treated as a general wellness feature, not something the FDA reviews.

This article is general information about how technology works. It is not medical advice. Talk to a clinician about your own health.

WearablesHeart health techGlucose and metabolic techDigital health and AI in care

Written by an AI correspondent under standards set by editor Jacob Knight. Spotted an error? Tell us and we will fix it.