How does a watch calculate calories?
Most devices do not measure calorie expenditure directly. It is estimated using age, sex, height, weight, accelerometer data, optical heart rate sensors, and proprietary algorithms specific to the manufacturer. For this reason, two different watches may give different calorie values for the same workout.
2026 direct comparison study
In a 2026 PLOS ONE study, the Apple Watch Series 8, Garmin Forerunner 955, Samsung Galaxy Watch 5, and Fitbit Sense 2 were compared with portable COSMED K5 indirect calorimetry. The median absolute percentage error of the devices for energy expenditure was approximately in the 15–25% range, and the error could vary with individual characteristics such as body fat percentage.
What do systematic reviews say?
A meta-analysis including 60 validation studies showed that energy expenditure accuracy varied significantly depending on the type of activity. Another systematic review including 158 publications reported that step counts and certain heart rate measurements could be acceptable, but no brand consistently demonstrated adequate accuracy in energy expenditure.
Heart rate measurement
Optical PPG sensors generally work better at rest and during rhythmic aerobic activity. Resistance exercise, intervals, wrist movements, cold, low peripheral perfusion, tattoos, and sensor fit can increase the margin of error. For clinical rhythm diagnosis, a watch ECG or heart rate measurement alone is not sufficient.
Step count
Step count is a more reliable behavioral metric than calorie estimation for most users. However, in situations such as walking very slowly, using a cane/walker, pushing a stroller, or keeping the wrist stationary, steps may be undercounted.
Sleep measurement
Wearable devices do not measure true sleep stages directly like an EEG; they estimate them from movement and heart rate patterns. Total sleep time and bedtime-wake-up trends can be useful, but interpreting single-night details such as “23% deep sleep” at the level of clinical polysomnography is not accurate.
HRV and recovery scores
HRV can provide valuable information about the autonomic nervous system; however, time of measurement, respiration, alcohol, illness, exercise load, and device algorithms affect the results. Rather than a single HRV value, the trend relative to an individual’s own long-term baseline is more meaningful.
The most common mistake in weight management
The watch saying “you burned 700 kcal” is not a reliable basis for eating an extra 700 kcal. When energy expenditure estimation errors and nutrition label errors are combined, a significant discrepancy can occur. In weight management, long-term body weight, waist circumference, activity, and dietary trends should be evaluated together.
When is a clinical device needed?
If a clinical decision is required for VO₂max evaluation, metabolic rate, exercise prescription, or cardiopulmonary performance, validated methods such as indirect calorimetry/CPET are more appropriate than consumer watches.
Related reading
2026 Smartwatch Calorie Validation Study · VO₂max and Cardiorespiratory Fitness
References
- Kostrna J, et al. Body fat, skin tone, and the accuracy of smartwatch caloric expenditure estimates. PLOS ONE. 2026;21(7):e0353261.
- O’Driscoll R, et al. How well do activity monitors estimate energy expenditure? Br J Sports Med. DOI: 10.1136/bjsports-2018-099643.
- Fuller D, et al. Reliability and Validity of Commercially Available Wearable Devices… JMIR Mhealth Uhealth. 2020.
Last medical review: September 25, 2026
