Independent Test Finds Popular Smartwatches Often Overestimate Workout Calories

Four popular fitness watches overestimated how many calories users burned during a controlled cycling workout, according to an independent evaluation by researchers at Florida International University. The findings, published in PLOS One, also showed that accuracy varied among devices and participants.

The study involved 58 adults who rode a recumbent bike while alternating between moderate and vigorous effort. Each participant wore an Apple Watch Series 8, Fitbit Sense 2, Samsung Galaxy Watch 5 or Garmin Forerunner 955. The researchers compared the watches’ calorie estimates with measurements from a COSMED K5 metabolic analyzer, clinical laboratory equipment used to assess actual energy expenditure.

Typical errors were approximately 15% to 25%, although the average error was considerably higher for the poorest-performing watches. Apple’s device was the most accurate in the evaluation, while Garmin and Samsung produced the largest overestimates. The researchers could not draw a clear conclusion about Fitbit because the watch sometimes returned implausibly low values—including one calorie for an entire workout—or failed to produce a reading. Much of its data was therefore excluded from the analysis.

The error also appeared to be related to body-fat percentage. Participants with higher body fat generally experienced larger discrepancies between the watches’ estimates and the metabolic analyzer’s measurements. The researchers did not establish why this pattern occurred, but lead author Jason Kostrna said the algorithms used to estimate calorie expenditure may contribute. The study did not establish that any particular algorithm was responsible.

That limitation matters for people who use wearable calorie estimates to monitor energy balance. An inflated estimate could make a workout appear to offset more food intake than it actually did, although the study did not directly examine weight change or long-term calorie balance.

The researchers said the watches may still be useful for tracking broader exercise trends. They described calorie-burn figures as rough guides rather than precise measurements and pointed to measures such as heart rate and time spent in different heart-rate zones as data users can act on during training. They also cautioned against responding to inaccurate estimates by cutting food intake too aggressively, noting that the body needs fuel for recovery and consistent exercise.

Kostrna is continuing the research by examining how well the watches estimate energy expenditure during traditional strength training, where brief, intense efforts may be difficult for wearable algorithms to interpret. The current findings apply to the tested devices and controlled cycling protocol; they do not establish how accurately the watches perform across all workouts or users.