Are Fitness Trackers Accurate? What 71 Studies Show
Last Updated: October 26, 2026
Fitness trackers are accurate for some things and not for others, and no single wearable is the most accurate on every metric. I went through 71 peer-reviewed studies that tested seven brands (Oura, Apple Watch, Garmin, Fitbit / Google, Samsung, WHOOP and Polar) against lab equipment on ten measures, and the brand that ranked first changed from one metric to the next. Oura ranked first on all four overnight measures: sleep stages, sleep timing, HRV and resting heart rate. Apple Watch ranked first on exercise heart rate and blood oxygen, Garmin on steps and VO2 max, and Fitbit / Google on calories and breathing rate. Calories are the weakest number on every wrist: no brand came within 9% on a max treadmill test. So the right device depends on which number you actually check, and some numbers deserve less trust than the screen suggests. Here is what the studies show.
Want to sort these results by metric yourself? Open the wearable accuracy hub, then come back for the story behind the numbers.
I did this because I kept running into the same problem: every accuracy claim I found was either a brand's own marketing or a review that tested one watch on one person. So I read more than 500 sources, kept the 71 studies that tested a device against a lab reference, and took every figure from the paper's own results tables. No brand paid for it, and no brand saw it before it went out.
Which wearable ranked first, metric by metric
One ranking per metric, decided by the numbers. Sleep gets two rankings because telling sleep stages apart and knowing when you were asleep are two different jobs with different winners.
Metric | Ranked 1st | 2nd | 3rd | Confidence |
Sleep stages | Oura Ring | Apple Watch | Fitbit / Google | High for 1st, medium after |
Sleep timing | Oura Ring | Fitbit / Google | Apple Watch | High for 1st, medium after |
HRV | Oura Ring | WHOOP | Garmin | High |
Resting heart rate | Oura Ring | Polar | WHOOP | Medium |
Exercise heart rate | Apple Watch | Garmin | Fitbit / Google | Medium |
Blood oxygen (SpO2) | Apple Watch | Samsung | Garmin | High for 1st, medium after |
Steps | Garmin | Apple Watch | Samsung | Medium |
Calories | Fitbit / Google | Samsung | Apple Watch | Medium |
VO2 max | Garmin | Fitbit / Google | Apple Watch | High for 1st, medium after |
Breathing rate | Fitbit / Google | WHOOP | No qualifying third | Low |
Models tested: sleep is Oura Ring 3, Apple Watch Series 8 and Fitbit Sense 2; HRV and resting heart rate are Oura Ring 4, WHOOP 4.0, Garmin Fenix 6 and Polar Grit X Pro; Garmin's first place on VO2 max is the Forerunner 245 and Fenix 6 with a chest strap. Confidence is high when a head-to-head study of current models decides the order, medium when the order depends on which single-brand study is used, and low when the figures come from different references or older models. Skin temperature is the eleventh metric studied and is not ranked.
How accurate are fitness trackers? Five findings
1. No wearable is the most accurate on every metric
Four brands share the ten first places, and the spread underneath is wider. Apple Watch finished in the top three on the most metrics, seven of ten: first on two, second on two and third on three. Oura ranked first four times and finished nowhere else. Samsung, WHOOP and Polar never ranked first, though each finished second at least once.
2. Overnight HRV is close to lab grade
The cleanest result comes from one independent home study (Dial 2025, 13 adults, funded by the Air Force Research Laboratory) that put four brands against a chest-strap ECG at the same time. Oura Ring 4 was 5.96% off over 139 nights. WHOOP 4.0 was 8.17% off and Garmin Fenix 6 was 10.52% off in the same study. When a ring on your finger lands within 6% of an ECG at night, the sensor is no longer the weak link.
3. Watches miss the time you lie awake
Most of the night is sleep, so any device that guesses "asleep" looks accurate. The harder test is catching the minutes you lie awake. In a lab sleep study, Apple Watch Series 8 caught 52.4% of awake time (Robbins 2024, Oura-funded †). Garmin Forerunner 245 Music caught 27% and counted 43.8 extra minutes of sleep (Miller 2022, 53 healthy young adults, lab receives WHOOP support †). Oura Ring 3 caught 73.0% (Svensson 2024, Oura-funded authors †). If your watch says you slept longer than you feel you did, this is usually why. The full sleep breakdown is in our sleep tracker accuracy comparison.
4. Calorie burn is the least accurate number
No brand came within 9% on the max treadmill test, and weight training errors ran from 41.7% to 114.9%. Even Fitbit / Google, which ranked first on calories, was 13.97% off on that treadmill test. In weight training, Apple Watch was 112.2% off and Garmin 114.9% off, which means the number on screen can be roughly double the real burn. Oura has one older lab test, listed as context and not ranked: Oura Ring Gen 2 was 21.1% off in METs across eight lab activities (Kristiansson 2023, independent, older model ‡), and Oura has changed its calorie algorithm since. Our wearable calorie burn breakdown goes device by device.
5. Blood oxygen fails when it matters
Apple Watch passed the low-oxygen lab test at 2.2 Arms, where 4.0 or less passes (Walzel 2023). But every watch drifts when oxygen is truly low, and fails to give a reading 10 to 45% of the time. A watch that reads well at 97% saturation tells you little about how it behaves at 85%.
How accurate is your wearable? Brand by brand
Every brand has something it does well. Here is each one, strengths first, then the gaps.
Oura Ring. Ranked first on four metrics, all overnight: sleep stages, sleep timing, HRV and resting heart rate. It is not in the top three on exercise heart rate (15% off), steps (older tests read 50 to 69% over, all before its May 2025 step algorithm), blood oxygen and VO2 max, which have never been tested, or calories.
Apple Watch. Ranked first for exercise heart rate and blood oxygen, and in the top three on seven of ten metrics, more than any other brand. It is not in the top three on HRV, where Apple reports a different statistic (SDNN) and read 28.9% off, on resting heart rate, or on breathing rate, where it gave no readings in two hospital studies.
Garmin. Ranked first for steps and for VO2 max with a chest strap, second for exercise heart rate, third for HRV and blood oxygen. It misses most of the time you lie awake, is not in the top three on resting heart rate, runs 77.6% to 114.9% off on calories in cycling and weight training, and its breathing rate feature failed a hospital test.
Fitbit / Google. Ranked first for calories and breathing rate, second for sleep timing and VO2 max, third for sleep stages and exercise heart rate. Its HRV, resting heart rate and blood oxygen tests are weak, and it counts 15% too many steps in free living across pooled studies.
Samsung. Second for blood oxygen and calories, third for steps. It has no test of the shipped feature for sleep in healthy adults, HRV, resting heart rate, VO2 max or breathing rate.
WHOOP. Second for HRV and breathing rate, third for resting heart rate. Outside overnight metrics it has no top-three finish, because it has no peer-reviewed test for blood oxygen, steps, calories or VO2 max.
Polar. Second for resting heart rate (Grit X Pro, 2.71% off). Everything else is either far behind, such as steps at 72% off, or untested.
How the rankings were decided
Numbers decide. Within each metric, brands are ranked on every sub-measure and the ranks are combined with equal weight. Head-to-head studies, where several brands are tested on the same people with the same protocol, outrank single-brand studies when the two disagree.
Brand-funded studies count. They are marked †, and a funded figure is replaced only when independent tests of the same brand disagree with it. I checked every placing, and no top-three place rests on a funded figure that independent work contradicts. Older models are used only when nothing newer has been tested, and are marked ‡.
Only the feature a buyer actually gets is ranked, and only from peer-reviewed tests. Research algorithms run on a watch's raw data, company blog posts, unpublished internal tests and preprints are listed in a "Not counted" section with the reason, so nothing is hidden. That is where you will find figures like WHOOP's internal VO2 max test and Oura's 2020 breathing rate blog. Healthy adults are the main population; clinical samples are used as a check on the order, not as the ranking. The full rules and the not-counted list are in the free State of Wearable Accuracy 2026 report.
What to do with a number that's a little off
A wearable that is 10% off on a single reading can still be very useful, as long as you compare it with itself. Your device makes roughly the same error night after night, so a trend in your own HRV or resting heart rate over weeks is far more trustworthy than any one morning's number.
The same logic applies to calories and VO2 max, the two weakest metrics. Treat them as a direction, not a target. If your VO2 max estimate climbs over a training block, the climb is more meaningful than the starting figure.
The real value shows up when you pair those trends with what you ate and how you slept. A late dinner, a second coffee or a heavy training day will show up in your overnight numbers, and that is something no lab study can tell you about yourself. For the deeper ranking of every sub-metric, see the most accurate wearable, metric by metric. If you are shopping for a new device, our best wearables page sorts them by the metric you care about.
Common questions
Are fitness trackers accurate?
For some metrics, yes. The best overnight readings land within 2 to 6% of a chest-strap ECG (Oura Ring 4 was 1.94% off on resting heart rate and 5.96% off on HRV), but no brand came within 9% on calories during a max treadmill test. How far to trust yours depends on the metric and the brand.
Is there one wearable that is accurate for everything?
No. Four brands share the ten first places, and no brand finished in the top three on every metric. Apple Watch finished in the top three most often, seven of ten, but not on HRV, resting heart rate or breathing rate.
How accurate are fitness trackers for calories burned?
Less accurate than any other metric. No brand came within 9% on the max treadmill test, and weight training errors ran from 41.7% to 114.9%, so the number on screen can be roughly double the real burn.
How accurate are smartwatches for steps?
Close in everyday life for the best brands. Across a 59-study analysis of free living, Garmin read 3% high, Apple Watch 7% high and Fitbit 15% high. Our step count accuracy ranking has every model.
Why does my watch say I slept longer than I did?
Watches miss much of the time you lie awake. Apple Watch Series 8 caught 52.4% of awake time in a lab study and Garmin Forerunner 245 Music caught 27%, counting 43.8 extra minutes of sleep.
How accurate is VO2 max on a smartwatch?
Good with a chest strap, weak from the wrist alone. Garmin with a chest strap was 6.7% and 7.05% off a lab test, Apple Watch was 13.2% to 15.8% off and read low, and Garmin's wrist-only Vivoactive 4S had a median error of 56% in 207 mostly older women. Our VO2 max accuracy ranking has the full comparison.
Is wearable skin temperature accurate?
Nobody has checked yet. No brand has an independent test of its temperature sensor against a thermometer. The published figures measure ovulation detection, not temperature, and most are brand-funded. More on that in is your wearable's skin temperature accurate.
Did any brand fund these studies?
Some, and they count. Brand-funded studies are marked † and are replaced only when independent tests of the same brand disagree. No top-three place rests on a funded figure that independent work contradicts.
Can I use two wearables at once and get the best of both?
Yes. Many people wear a ring for sleep and a watch for training. The hard part is combining the data without double-counting. Kygo handles that by taking each metric from the device that measures it best.
Where can I see every result?
In the free State of Wearable Accuracy 2026 report: 78 pages covering seven brands and 71 studies, at kygo.app/state-of-wearable-accuracy.
The bottom line
Are fitness trackers accurate? For overnight heart measures, close to lab grade. For calories, not on any brand. No wearable is the most accurate on every metric: Oura ranked first on everything measured overnight, Apple Watch, Garmin and Fitbit / Google split the daytime metrics, and skin temperature has no independent test to rank. The best device for you depends on the metric you care about most, and the most useful thing you can do with any of them is watch your own trend over time.
Key sources: Dial 2025 (Physiological Reports, 13 adults, independent, Air Force Research Laboratory funded); Robbins 2024 (Sensors, Oura-funded †); Svensson 2024 (Sleep Medicine, Oura-funded authors †); Miller 2022 (Sensors, 53 adults, lab receives WHOOP support †); Walzel 2023 (Sensors, induced hypoxia); Kostrna 2026 (PLOS ONE, 58 adults); Yun 2026 (Bioengineering, 70 men); Lee 2026 (Sensors, 62 men, SOLUM Health Care grant); Perez 2025 (conference abstract, about 20 adults); Kristiansson 2023 (BMC Medical Research Methodology, corrected September 2023, independent ‡); Paluch 2026 (Journal of Sport and Health Science, 59-study meta-analysis); Engel 2026 (European Journal of Applied Physiology); Jamieson 2024 (Computing in Cardiology). Every other study, with its funding, is listed in the report.
Disclaimer: Kygo is a personal data aggregation and insights platform designed for informational purposes only. The information provided by Kygo, including correlations, patterns, and trends identified in your data, does not constitute medical advice, diagnosis, or treatment. Always consult a licensed healthcare provider with any questions regarding medical conditions.
I'm curious: which metric does your own wearable get right, and which one have you stopped trusting? Tell me at Ryan@kygo.app, I read every reply.