Smartwatches Get Called Out

Smartwatches can miss the mark on six popular health numbers, including calories, steps, heart rate, sleep, recovery and fitness scores. (abc.net.au) ABC News in Australia and ScienceAlert on April 21, 2026, both highlighted a new explainer by exercise scientist Hunter Bennett that says many watch readouts are estimates, not direct measurements. Bennett wrote that users often treat those scores as precise even when the devices infer them from wrist sensors and software models. (abc.net.au, sciencealert.com) The wrist sensor at the center of most watches uses light to track changes in blood flow, a method called photoplethysmography. A 2024 review in *Sports Medicine* found consumer wearables showed a mean heart-rate bias of about ±3%, while an American College of Cardiology summary said errors became clinically meaningful during exercise and were larger in people with atrial fibrillation. (link.springer.com, acc.org) Calories are harder because watches do not measure energy burn directly; they estimate it from motion, pulse and profile data. Bennett cited evidence that wearables can under- or overestimate energy expenditure by more than 20%, and a *British Journal of Sports Medicine* meta-analysis found accuracy varied widely by activity type, with large differences across devices and studies. (abc.net.au, bjsm.bmj.com) Step counts look simpler, but the watch is really counting arm motion and translating that into walking. Bennett reported smartwatches can undercount steps by about 10% under normal exercise conditions, with pushing a stroller, carrying weights or walking with limited arm swing making the number less reliable. (abc.net.au, sciencealert.com) Sleep scores are another layer of estimation because the device guesses sleep stages from movement and pulse patterns instead of reading brain waves. In a 2025 validation study against polysomnography, six wrist-worn devices detected more than 90% of sleep epochs but had low wake specificity of 29.39% to 52.15% and only fair-to-moderate agreement overall. (pmc.ncbi.nlm.nih.gov) Recovery and readiness scores stack one estimate on top of another. Bennett said those scores usually combine heart-rate variability and sleep quality, so errors in both inputs can flow into a single number that tells users to train harder or rest longer. (abc.net.au, sciencealert.com) The broader evidence base is still thin relative to the size of the market. The 2024 umbrella review found just 24 systematic reviews covering 249 non-duplicate validation studies and said only about 11% of commercially available wearables had been validated for at least one biometric outcome. (link.springer.com) That gap is feeding a parallel debate over form factor as buyers weigh smart rings against watches. TechGenyz reported on April 21 that rings are gaining attention for discreet wear and longer battery life, while watches still lead on screens, apps, calls, navigation and broader ecosystem features. (techgenyz.com) The practical split is becoming clearer: rings suit people who want passive tracking and fewer interruptions, and watches suit people who want workout tools and on-wrist interaction. The numbers on either device can still help show trends over time, but the April 21 reporting says they are less dependable as exact daily verdicts on health or fitness. (techgenyz.com, abc.net.au, acc.org)