You finish a workout and your watch congratulates you: 512 calories burned. It is a pleasingly precise number, and it feels like a measurement.
It is not. It is an estimate, and according to the independent research, it is often a poor one. The same studies that vindicate your watch's heart rate sensor find that its calorie figure carries an average error of nearly 28%.
That gap matters more for weight loss than any other feature on the device, because calorie burn is the number people act on. Here is what the evidence actually shows, what your tracker is genuinely good at, and how to use one without letting it steer your progress off course.
What the research actually found
The clearest picture comes from a meta-analysis that pooled 56 validation studies of consumer wearables. For heart rate, the devices were impressive: a mean absolute percentage error of just 4.43%. Step counting came in at 8.17%, perfectly serviceable. Energy expenditure, the calorie burn figure, recorded an average error of 27.96%.
The lead researcher's advice was blunt: "Think of it as a helpful guide, not a diagnostic tool."
This is not a problem with one cheap brand. The most-studied wearable on earth, the Apple Watch, has been the subject of a living systematic review covering 82 studies and 430,052 participants. Its heart rate sensing was excellent, undershooting by an average of just 0.27 beats per minute. Its energy expenditure error was rated "inconsistent and frequently large".
So the pattern is consistent across the industry: the sensors are good, and the calorie arithmetic built on top of them is not.
Why the calorie number goes wrong
Your watch does not measure calories. Nothing on your wrist can. It measures heart rate and movement, then feeds those into a model along with your age, weight, sex and height, and the model guesses how much energy someone with your profile probably spent.
The guess breaks down because people are not averages. Two people of the same age and weight can burn meaningfully different amounts doing the same session, depending on fitness, muscle mass, technique and how efficiently they move. Activity type matters too: wrist-based estimates struggle most with anything the wrist cannot see properly, such as cycling, strength work or carrying shopping.
Worse, the error is not a fixed bias you could learn to correct for. The Apple Watch review's finding that errors are "inconsistent" is the important word: the same watch might overestimate your run on Monday and underestimate your gym session on Wednesday. There is no reliable conversion factor between the number on your wrist and the truth.
What a 28% error does to a weight-loss plan
Take a concrete example. Your watch credits a gym session with 500 calories. At the documented average error, the honest range runs from roughly 360 to 640. You do not know which end you are on, and the watch presents the midpoint with total confidence.
Now imagine the common habit of "eating back" exercise calories. If the estimate runs high and you faithfully eat back the full figure, you can erase a planned daily deficit without ever breaking a rule. A modest weight-loss plan typically relies on a deficit of a few hundred calories a day; an overestimate the size of the documented average error on a single workout can swallow most of it.
This is one reason people who feel they are doing everything right still see the scale refuse to move, and then conclude their body is simply broken. Sometimes the arithmetic was never real in the first place. There are genuine biological reasons progress varies between people too, which we cover in why losing weight is more difficult for some people, but a flattering calorie counter deserves to be ruled out first.
The practical rules are simple. Never treat the calorie burn figure as currency you can spend. If you account for exercise at all, count a fraction of what the watch displays. And judge your plan by results over weeks, not by the daily numbers on your wrist.
Where trackers genuinely earn their place
None of this means the device is useless. It means you should lean on the parts the research supports.
Heart rate and effort. At around 4% error, heart rate is trustworthy enough to train with. Effort zones, resting heart rate trends and workout intensity are on solid ground.
Steps and consistency. An 8% step error does not matter when the job is turning 4,000 steps a day into 8,000. For most people, the tracker's real value is behavioural: it makes inactivity visible and streaks satisfying.
The trend line. Because you wear the same device, running the same algorithm, on the same wrist, week-on-week comparisons carry information even when the absolute numbers are off. If your watch says this month's activity is up a quarter on last month's, that direction is probably real.
The price of entry has collapsed. Google's screenless Fitbit Air launched in the US at $99.99 with an AI coach built in, and in the UK, insurer schemes change the sums entirely: Vitality offers the Apple Watch Series 11 from £9 upfront on 40-month interest-free credit, with monthly payments falling as low as £0 depending on your activity.
One cost to watch is the subscription. Oura charges £5.99 a month or £69.99 a year and gates most of its analytics behind it, while WHOOP sells its hardware only as part of memberships running £169 to £349 a year in the UK. Garmin and Polar take the opposite approach and include their full platforms free with the device. For weight loss specifically, the free basics, steps, heart rate and trends, are the part that works anyway.
The numbers the watch cannot give you
Even a perfectly accurate tracker would only ever describe one side of the equation: energy out, across your whole body. Weight-loss progress is better judged by cheaper, older instruments. A tape measure around the waist, monthly photographs in the same light, and how clothes fit will tell you more than any wrist-worn estimate, and the scale is best read as a weekly trend rather than a daily verdict. BMI has its own blind spots, which we unpack in our guide to BMI.
The whole-body point matters for one more reason. Calorie balance determines how much fat you carry overall, but it has no say in where you carry it. Diet and activity cannot target a specific area, which is why stubborn pockets on the stomach, flanks or thighs often outlast an otherwise successful plan. We compare the two approaches honestly in fat freezing versus diet and exercise: lifestyle change moves the scale, while a localised treatment like fat freezing changes the shape of a treated area without meaningfully moving the scale at all. They answer different questions, and no tracker measures either one directly.
Use the watch, do not obey it
A fitness tracker is a good servant and a poor accountant. Trust its heart rate, use its step count to build the habit, watch the trend lines, and treat the calorie figure as a rough sketch that flatters you more often than not.
If the number that actually bothers you is a specific area rather than the one on the scale, that is a different conversation, and a tracker was never going to settle it. A free consultation at a clinic can give you an honest assessment of whether a localised treatment suits you, with no obligation either way.
Fitness trackers for weight loss
What your watch gets right
- Heart rate is genuinely reliable, with an average error of about 4% across 56 validation studies
- Step counts are accurate enough to build a daily activity habit around, at roughly 8% average error
- Worn consistently, the trend line is meaningful even when the absolute numbers are not
Where it misleads
- Calorie burn estimates carry an average error of 27.96%, and the error is inconsistent, so you cannot simply correct for it
- Recovery and readiness scores are proprietary composites with no independent validation behind any brand's version
- On some brands the useful analytics sit behind a subscription that costs more than the device over time
Frequently Asked Questions
How accurate are fitness tracker calorie counts?
Not very. A meta-analysis pooling 56 validation studies found an average error of 27.96% for energy expenditure, against 4.43% for heart rate and 8.17% for steps. A separate living review of Apple Watch studies described calorie error as inconsistent and frequently large. Treat the number as a rough guide, not a measurement.
Should I eat back the calories my watch says I burned?
Not the full figure. If the estimate runs high, and on average it can be off by more than a quarter, eating back everything the watch credits you with can quietly erase your deficit. If you want to account for exercise at all, count only part of the displayed figure and judge the result by what your weight does over several weeks.
Which fitness tracker has the most accurate calorie counter?
No brand has solved this. The large error shows up across manufacturers, and the most-studied device on the market, the Apple Watch, still shows calorie errors described by reviewers as inconsistent and frequently large. Choose a tracker for comfort, battery life and cost, and treat every brand's calorie figure with the same caution.
Is anything on a fitness tracker actually reliable?
Yes. Heart rate is the standout, accurate to within roughly 4% on average, and step counting sits around 8% error, which is easily good enough for building an activity habit. The problems concentrate in the derived estimates: calories burned, sleep stages, and composite scores like readiness or recovery.
Do I need a paid subscription to use a fitness tracker for weight loss?
No. Oura and WHOOP gate their core analytics behind memberships, but Garmin and Polar include their full platforms free with the device, and the basics that matter for weight loss, steps, heart rate and activity trends, are free on most watches.
