Trackers may wildly overcount

A fitness tracker can count your pulse like a decent speedometer, but calories are more like estimating gas burned from the sound of the engine. In a Stanford study published on May 24, 2017, none of 7 popular wrist devices measured energy expenditure well, and the worst error reached 93%. (med.stanford.edu) “Energy expenditure” is just the calories your body uses while sitting, walking, running, or cycling. Stanford tested that by comparing tracker estimates with indirect calorimetry, a lab method that measures oxygen used and carbon dioxide produced through breathing. (mdpi.com, commondataelements.ninds.nih.gov) The Stanford team put 60 volunteers through sitting, walking, running, and cycling tests while they wore the Apple Watch, Basis Peak, Fitbit Surge, Microsoft Band, Mio Alpha 2, PulseOn, and Samsung Gear S2. The group included 29 men and 31 women with different ages, weights, fitness levels, and skin tones. (mdpi.com) Heart rate was the part these devices handled best. Six of the 7 devices had median heart-rate error below 5% during cycling, and Stanford said heart-rate readings were generally much closer than calorie readings. (mdpi.com, med.stanford.edu) Calories were the messier part because watches do not directly measure burned energy. They infer it from movement sensors, heart-rate sensors, and personal inputs like age, height, weight, and sex, then run that through a proprietary formula. (sciencedirect.com, images.apple.com) That is why two people can do the same 30-minute workout and get different calorie numbers, and why one wrong setup detail can skew the result. In the Stanford paper, device error was higher for walking than cycling and was also higher in men, people with greater body mass index, and people with darker skin tone. (mdpi.com) The 93% number also does not mean every watch is wrong by 93% all the time. In Stanford’s results, the best device still missed energy expenditure by an average of 27%, and no device got below 20% error. (med.stanford.edu, mdpi.com) That gap is big enough to change eating decisions. If your watch says a workout burned 600 calories when the real number is closer to 400, eating back the full 600 can erase the deficit you thought you created. (med.stanford.edu) Researchers still see value in wearables, just not in treating calorie totals like lab measurements. A 2024 umbrella review of systematic reviews found consumer wearables are widely used, but accuracy varies across outcomes and devices, which is why calorie numbers are better used as rough trend lines than precise targets. (link.springer.com) The practical move is to trust your tracker more for consistency than precision. If the same watch shows your long run usually costs more energy than your short walk, that ranking is useful even when the exact calorie total is not. (med.stanford.edu, mdpi.com) Stanford’s senior author, Euan Ashley, put the warning plainly in 2017: people are making life decisions from these numbers. The viral post is new, but the underlying finding is not that your watch is useless; it is that the calorie number on your wrist is an estimate built from guesses, not a direct readout from your body. (med.stanford.edu)