Four of the most popular smartwatches overstated how many calories people burned during a controlled cycling workout, missing the true number by a median of roughly 15% to 25%, according to a study of 58 adults published in PLOS One by researchers at Florida International University. The error ran in one direction, upward, and it got worse as a person’s body fat percentage rose.
Key takeaways
- Median percentage error was about 15% to 25% across devices, and three of the four watches read significantly high.
- Average overestimates ran about 22 calories for Apple, 57 for Samsung, and 69 for Garmin per session.
- Higher body fat predicted larger error on every brand. Skin tone, tested across Fitzpatrick types III to V, showed no effect.
What the study found
Participants wore an Apple Watch Series 8, Fitbit Sense 2, Samsung Galaxy Watch 5, and Garmin Forerunner 955 at the same time while a COSMED K5 metabolic analyzer measured their actual energy use through breath-by-breath gas exchange. That machine is the laboratory reference standard. The workout was a 10-minute recumbent cycling protocol alternating two-minute intervals at moderate and vigorous intensity, with five minutes of rest and recovery on either side.
Average error, in calories, was 21.60 for Apple (very likely between 11.59 and 31.60), 56.76 for Samsung (very likely between 45.11 and 68.41), and 68.61 for Garmin (very likely between 53.28 and 83.94). All three read high.
Fitbit’s average looked clean at 3.14 calories, but the study calls that number misleading. The device produced “seven implausible estimates (> 450% of K5 values), representing approximately 13% of Fitbit trials,” and failed to return any reading in one session. FIU’s own writeup notes it sometimes reported “just one calorie for an entire workout.” Put the outliers back in and Fitbit’s average error jumps to 128.6 calories, the worst of the group.
The body fat finding was the sturdier one. In the statistical models, higher body fat percentage predicted larger percentage error across all devices, with a device-by-body-fat interaction that has about a 2 in 100 chance of being coincidence (p = .02). The device effect itself had less than a 1 in 1,000 chance (p < .001). Fitzpatrick skin type III to V showed nothing (p = .89).
Dr. Kumar’s take
The direction matters more than the size. A watch that guessed randomly would be annoying but harmless. A watch that consistently guesses high tells you that you earned food you did not earn, and it tells that story most loudly to the people using it to lose weight. Lead author Jason Kostrna put it plainly: “You could be off hundreds of calories each week and easily end up in a calorie surplus when you think you’re in a calorie deficit.”
Now the part the headlines skipped. This was 58 Hispanic adults with a mean age of 23, all with Fitzpatrick skin types III to V, on a recumbent bike for ten minutes. That is not running, not lifting, not a real training week, and not a broad patient population. The authors say so themselves, noting that accuracy “may differ across other activities and populations.” Body fat was measured across a range that included an average BMI of 29.73, so the adiposity signal is not coming from a handful of outliers, but a single short cycling bout in young adults is a narrow window.
The practical conclusion is not to adjust the calorie number down by some correction factor. It is to stop using it. An input with a median error of 15% to 25% and a bias that scales with your body composition is not a number you can budget against. The rest of what these watches do, heart rate, minutes in a zone, step counts, and your own trend line over weeks, still works, because those measures are compared against yourself rather than against a metabolic cart.
What it means for you
Keep wearing the watch if you like it. Ignore the calorie figure. Track heart rate, time in each heart rate zone, and whether your numbers are moving in the direction you want over months.
If you are managing weight, anchor to intake and to the scale, not to what the watch says you burned. And do not swing the other way by slashing food to cover a phantom gap. FIU’s researchers caution against aggressive cutting for exactly that reason: recovery and consistent training both require fuel.