# Are sleep trackers accurate?

> Reasonably accurate for how long you slept — though most err on the generous side — less so for how long you lay awake, and least of all for sleep stages. A watch, ring or mattress sensor measures movement and heart rhythm and infers sleep from them; the stages are defined by brain activity, which a sleep laboratory records and a tracker does not. Over weeks a tracker can be useful. Night by night it can mislead.

Source: https://dreamsheeps.space/sleep-trackers
Published: 2026-10-14 · Updated: 2026-10-14

## How accurate are sleep trackers?

Good at telling that you were asleep, weaker at spotting when you were awake, and least reliable about sleep stages. That is the pattern in every study below, across different devices and different years. Each compared consumer devices with polysomnography — the overnight recording a sleep clinic uses — on the same night, in the same bed.

- *Seven devices, 2021.* Chinoy and colleagues tested four wearables and three bedside or under-mattress devices on 34 healthy young adults over three nights in a sleep laboratory, one night deliberately disrupted. All caught real sleep well; wakefulness was the weak point. Five of the seven matched or beat a research-grade actigraph, the wrist monitor sleep studies have long used, at telling sleep from wake; the two Garmin watches did worse. Most devices missed between a third and a half of deep and of REM sleep, usually calling it light, and tended to do worse on disrupted nights. No device company funded the study.
- *Six wrist devices, 2025.* Schyvens and colleagues tested a Fitbit Charge 5, Fitbit Sense, Withings ScanWatch, Garmin Vivosmart 4, WHOOP 4.0 and Apple Watch Series 8 for one night on 62 adults, some of them patients with suspected sleep apnoea. Each recognised more than 9 in 10 of the laboratory’s 30-second stretches of sleep but only 3 to 5 in 10 of its stretches of wakefulness. On average all six credited people with more sleep than the laboratory recorded: four of them significantly, by 20 to 40 minutes; the two Fitbits by 6 and 11 minutes, within chance. On a single night the error could run past an hour.
- *A ring and a mattress pad among five devices, 2024.* In 53 healthy adults aged 18 to 30, a Fitbit Inspire HR, Fitbit Versa 2, Garmin Vivosmart 4, second-generation Oura Ring and the Withings Sleep Tracking Mat, which goes under the mattress, all showed the same pattern: on nights with little wakefulness they overestimated it, and on nights with a lot of it they counted too little (Kainec and colleagues).
- *People with sleep problems, 2025.* In 65 adults referred to a Paris hospital sleep centre, two older trackers, the Withings Activité and Fitbit Alta HR, both overestimated total sleep and agreed with the laboratory on stages barely better than chance. The authors concluded that such trackers are not reliable for the structure of sleep in people with sleep complaints (Frija and colleagues).
- *The healthy-adult studies together, 2026.* A review of 16 studies, 10 of them pooled in a meta-analysis, found that wearables overestimated total sleep and underestimated time awake in the night, with no device consistently best: reasonable for overall sleep, not for diagnosis or detailed stages (Agostinho and colleagues).

### So do sleep trackers work?

For how long and how regularly you sleep, broadly yes — most now match the research actigraphs that sleep science relied on for years. For what happened inside the night, stage by stage, much less, and least of all on bad nights, which is exactly when the number is most likely to worry you.

## What does a sleep tracker actually measure?

Not sleep, directly. A sleep stage — light, deep or REM — is defined by the brain’s electrical activity. Polysomnography records it with electrodes on the scalp, beside the eyes and on the chin, along with the heart, and trained technologists score the night by eye in 30-second stretches. A tracker records what the body does instead, and an algorithm infers the stage from it:

- *Movement*, from an accelerometer. Apple’s sleep stages come from this alone, including the small movements of breathing.
- *Heart rate and heart rate variability*, from an optical sensor that shines light into the skin. Google says its Fitbit and Pixel devices read stages from tiny changes in heart rhythm.
- *Temperature and breathing rate* — skin temperature on Oura’s ring, wrist temperature on newer Apple Watches, breathing rate on several devices — and *blood oxygen* on some.
- *Heartbeat, breathing and movement through the mattress*, for pads that go under it and covers that go on it.

Google’s help page puts the difference plainly: “In clinical labs, experts use sensors on the brain and muscles to track sleep”, while its own devices “estimate sleep stages” from the wrist. Each company’s algorithm is its own and unpublished, and it changes with software updates — which a 2019 review lists among the things to weigh before relying on device data (de Zambotti and colleagues). Apple’s paper on its method records one change to its sleep tracking in 2024 and another to its staging in 2025.

What actually happens in the body as you drop off, and what each stage is, is the subject of [what happens when you fall asleep](https://dreamsheeps.space/what-happens-when-you-fall-asleep); the heart in particular, of [heart rate when falling asleep](https://dreamsheeps.space/blog/heart-rate-when-falling-asleep).

## How accurate is Apple Watch sleep tracking?

Good at whether you were asleep, moderate at sleep stages, and least reliable on deep sleep. Apple estimates stages from the watch’s accelerometer alone and labels them Awake, REM, Core and Deep, Core being what other trackers call light sleep (Apple’s paper on its method, updated October 2025).

- *Apple’s own validation*, run by Apple, reports that the watch identifies nearly all sleep, and that a 2025 update got better at catching “quiet wake” — lying awake but very still. It is useful detail from a company testing its own product.
- *An independent laboratory study (2025)* found the Series 8 agreed with polysomnography on stages better than the five other wrist devices tested, but still only moderately, and credited people with about 20 minutes more sleep than the laboratory recorded, on average (Schyvens and colleagues).
- *A study funded by Oura (2024)*, in 35 healthy adults, found the Series 8 underestimated deep sleep by about 43 minutes and overestimated light sleep by about 45 (Robbins and colleagues).
- *A living review of Apple Watch accuracy (2026)*, covering 82 studies of everything the watch measures, found three that tested sleep stages: good at separating sleep from wake, moderate to poor at telling similar stages apart, with deep sleep tending to be labelled light (Lambe and colleagues).

So: trust its record of when you slept and roughly how long, read its stage chart as an estimate, and expect the deep sleep figure to be the shakiest number on the screen.

## How accurate is the Oura ring?

Close to the laboratory on several measures in healthy adults, on the evidence so far — but the most favourable study was paid for by Oura, and none of them tested people whose sleep is disturbed. The ring reads movement, skin temperature, resting heart rate, heart rate variability and breathing rate, and Oura’s help pages cite its own validation of its staging, which is the company testing itself.

- *A study funded by Oura (2024).* In 35 healthy adults aged 20 to 50, the third-generation ring did not differ significantly from polysomnography on time awake, light, deep or REM sleep. Oura paid for the study, and its first author sits on Oura’s medical advisory board (Robbins and colleagues).
- *An independent study of the older, second-generation ring (2024)* found it underestimated light sleep on every kind of night (Kainec and colleagues).
- *The 2026 meta-analysis* noted that in individual studies the third-generation ring came closest to the laboratory on how long people took to fall asleep and on sleep efficiency (Agostinho and colleagues).

What these studies cannot tell you is how the ring does when nights are genuinely broken: they were run in healthy sleepers, and the one study above of people with sleep complaints tested other devices — and found them unreliable.

## What do Fitbit, Pixel Watch, Garmin, WHOOP, Samsung and Eight Sleep measure?

As each company’s own pages described them on 14 October 2026. Sheeps has not tested any of these devices and does not rate them. Sheeps — which is ours — is not in the table, because it is not a sleep tracker.

**Sleep trackers as their makers describe them, 14 October 2026**

| Device | What it uses at night, by its maker’s account | What it gives the night |
| --- | --- | --- |
| Apple Watch | The accelerometer, for sleep stages; heart rate and breathing rate shown alongside, and wrist temperature on Series 8 and later | A Sleep Score from 0 to 100: duration (up to 50 points), bedtime consistency over the last 13 nights (30) and interruptions (20) |
| Oura Ring | Movement, skin temperature, resting heart rate, heart rate variability and breathing rate | A Sleep Score from 0 to 100, from seven contributors: total sleep, efficiency, restfulness, REM sleep, deep sleep, latency and timing |
| Fitbit and Google Pixel Watch | Movement and heart rate, with heart rate variability for the stages | A Sleep Score in the Google Health app, from sleep duration, time to sound sleep, sound sleep, restlessness, full awakenings and interruptions, measured against targets for your age and gender |
| Garmin | Heart rate, heart rate variability and movement; breathing rate and blood oxygen on some models | A sleep score from 0 to 100, from duration and quality — stages, time awake, restlessness and an overnight stress reading from heart rate variability |
| WHOOP | Its developer documentation lists what a night returns: light, slow-wave and REM sleep, disturbances and breathing rate | Sleep performance, against a nightly “sleep needed” that adds sleep debt and recent strain to a baseline and takes naps off; sleep consistency and efficiency alongside |
| Samsung Galaxy Watch | The length of each sleep phase and movement during sleep; blood oxygen; snoring, from recorded sound | A sleep score that can be compared with other users of the same age and gender |
| Eight Sleep Pod | Sensors built into a mattress cover: heart rate, heart rate variability, breathing, sleep stages and snoring | A Sleep Fitness Score |

Pads that go under the mattress, and bedside sensors, need nothing worn at all. The under-mattress device in the 2021 study used a pressure-sensitive (piezoelectric) sensor beneath the chest to pick up heartbeat, breathing and movement, and two bedside devices used low-power radio waves; all three did as well as or better than the research actigraph at separating sleep from wake (Chinoy and colleagues). A cooling cover like Eight Sleep’s raises a separate question — whether temperature helps at all — which [does a cooling mattress help you sleep?](https://dreamsheeps.space/blog/cooling-mattress) takes on.

## What is a sleep score?

One company’s summary of the night, built to its own formula — not a clinical measure, and not the same thing from one device to the next. Apple’s counts duration, the regularity of your bedtime and interruptions, and leaves sleep stages out; Oura’s includes deep and REM sleep; Google’s compares you with targets for your age and gender; Garmin’s folds in an overnight stress reading; WHOOP’s measures you against a sleep need it works out each day.

So the same night can earn different numbers on two devices, and the same device can change its number when the formula changes — Google says it has “redefined how the Sleep Score operates”. It helps to know what ordinary looks like on these scales: Google says most of its users average between 72 and 83, and Garmin that its users averaged 71 in 2024.

Most scores also lean on the stage estimates that the studies above found least reliable. And a score is not a diagnosis: the American Academy of Sleep Medicine’s position, in 2018, was that consumer sleep technology, unvalidated and not cleared by the US Food and Drug Administration, cannot be used to diagnose sleep disorders — though it can be a useful thing to bring to a doctor (Khosla and colleagues).

## How should you use a sleep tracker?

As a rough record over weeks, not a verdict on last night. The researchers who test these devices draw the same line: the 2025 laboratory study concluded that the better devices could track “prolonged and significant changes” in sleep, and the 2026 meta-analysis that wearables may suit “longitudinal self-tracking” but not diagnosis.

- *Look at weekly averages, not single nights.* An average of many nights smooths out much of the random error — though not a device’s habit of erring on the generous side.
- *Trust duration and timing more than stages.* Deep and REM sleep are where devices disagree most with the laboratory, and with each other: in the hospital study above, one tracker overestimated deep sleep and the other underestimated it.
- *Keep the set-up the same.* Same device, same wrist or finger, worn snugly: Apple says a loose watch registers too much movement, and Google suggests wearing its devices two or three finger-widths above the wrist bone. A software update can shift the numbers on its own.
- *Look once, in daylight.* Not in bed, and never in the middle of the night; [orthosomnia](https://dreamsheeps.space/blog/orthosomnia) explains why watching sleep tends to disturb it.
- *Read a slow night against what is normal.* If the app says you took half an hour to fall asleep, see [how long it should take to fall asleep](https://dreamsheeps.space/blog/how-long-to-fall-asleep) before deciding anything is wrong — and remember that lying awake but still is what movement-based devices find hardest to spot.

The things that help sleep happen before you lie down, and a [bedtime routine](https://dreamsheeps.space/bedtime-routine) is the place for them. A tracker’s job is to record; looking can wait until the morning, or the end of the week.

## When does a sleep tracker make sleep worse?

When the number starts to matter more than the night. There is a name for that, [orthosomnia](https://dreamsheeps.space/blog/orthosomnia), and part of the reason for it is that what you are told about your sleep changes how you feel. In a 2018 experiment, 63 people with insomnia wore a watch that told them the next morning how well they had slept — except that the verdict was assigned at random. Those told they had slept badly felt less alert and more tired by the evening than those told they had slept well, though the two groups did no differently on an objective test of reaction time (Gavriloff and colleagues).

The other part is that watching sleep closely works against it: [why trying to sleep keeps you awake](https://dreamsheeps.space/learn/trying-to-sleep) explains how.

## When should you see a doctor rather than a dashboard?

When the problem is in how you feel or how you breathe, not in a number. A tracker cannot diagnose a sleep disorder, and a quiet one is not an all-clear.

### Signs of sleep apnoea

The NHS lists breathing that stops and starts during sleep, gasping, snorting or choking noises, loud snoring and waking up a lot, with tired days, poor concentration or headaches on waking. It advises seeing a GP if you have the main symptoms, and bringing along anyone who has seen them. Diagnosis is a sleep study, which can usually be done at home with devices that record breathing and heartbeat overnight.

Some watches now look for it. Apple’s sleep apnoea notifications use the accelerometer to watch for breathing disturbances over 30 days, and Samsung’s sleep apnoea feature for Galaxy Watch received authorisation from the US Food and Drug Administration in 2024. Apple says its feature is not intended to diagnose, and that not everyone with sleep apnoea will get a notification; Samsung says its feature should not replace diagnosis by a qualified clinician. If you have the symptoms, see a doctor whatever the watch says.

### Insomnia

If falling or staying asleep is hard on most nights, has gone on for months and is affecting your days, talk to a doctor. The first treatment the American Academy of Sleep Medicine recommends for chronic insomnia is cognitive behavioural therapy for insomnia, CBT-I (Edinger and colleagues, 2021), and a few weeks of tracker data is a reasonable thing to take along. If the tracker itself has become the worry, [put it away for a while](https://dreamsheeps.space/blog/orthosomnia).

## How this page was put together

What the devices use and how they score comes from each company’s own pages on 14 October 2026; how accurate they are comes from the published studies below, and where a study was paid for or run by the company whose device it tested, this page says so. We have not tested these devices and do not rank them, and this site carries no affiliate links. Devices change faster than studies are published; if something here has gone out of date, the address in the [Imprint](https://dreamsheeps.space/imprint) reaches a person.

## Sources

1. [Chinoy et al. (2021) — Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep 44(5), zsaa291](https://doi.org/10.1093/sleep/zsaa291)
2. [Schyvens et al. (2025) — A performance validation of six commercial wrist-worn wearable sleep-tracking devices for sleep stage scoring compared to polysomnography. Sleep Advances 6(2), zpaf021](https://doi.org/10.1093/sleepadvances/zpaf021)
3. [Kainec, Caccavaro, Barnes, Hoff, Berlin & Spencer (2024) — Evaluating accuracy in five commercial sleep-tracking devices compared to research-grade actigraphy and polysomnography. Sensors 24(2), 635](https://doi.org/10.3390/s24020635)
4. [Frija et al. (2025) — Metrology of two wearable sleep trackers against polysomnography in patients with sleep complaints. Journal of Sleep Research 34(2), e14235](https://doi.org/10.1111/jsr.14235)
5. [Agostinho, Borges, Pereira, Borges & Soares (2026) — Are wearable sleep-tracking devices reliable alternatives to polysomnography? A systematic review and meta-analysis. Behavioral Sleep Medicine, published online](https://doi.org/10.1080/15402002.2026.2673893)
6. [Robbins et al. (2024) — Accuracy of three commercial wearable devices for sleep tracking in healthy adults. Sensors 24(20), 6532 (funded by Oura Ring Inc.)](https://doi.org/10.3390/s24206532)
7. [Lambe, Baldwin, O’Grady, Schumann, Caulfield & Doherty (2026) — The accuracy of Apple Watch measurements: a living systematic review and meta-analysis. npj Digital Medicine 9, 63](https://doi.org/10.1038/s41746-025-02238-1)
8. [de Zambotti, Cellini, Goldstone, Colrain & Baker (2019) — Wearable sleep technology in clinical and research settings. Medicine & Science in Sports & Exercise 51(7), 1538–1557](https://doi.org/10.1249/mss.0000000000001947)
9. [Khosla et al. (2018) — Consumer sleep technology: an American Academy of Sleep Medicine position statement. Journal of Clinical Sleep Medicine 14(5), 877–880](https://doi.org/10.5664/jcsm.7128)
10. [Gavriloff, Sheaves, Juss, Espie, Miller & Kyle (2018) — Sham sleep feedback delivered via actigraphy biases daytime symptom reports in people with insomnia: implications for insomnia disorder and wearable devices. Journal of Sleep Research 27(6), e12726](https://doi.org/10.1111/jsr.12726)
11. [Edinger et al. (2021) — Behavioral and psychological treatments for chronic insomnia disorder in adults: an American Academy of Sleep Medicine clinical practice guideline. Journal of Clinical Sleep Medicine 17(2), 255–262](https://doi.org/10.5664/jcsm.8986)
12. [NHS — Sleep apnoea (page last reviewed 11 May 2026; checked 2026-10-14)](https://www.nhs.uk/conditions/sleep-apnoea/)
13. [Apple — Estimating sleep stages from Apple Watch, updated October 2025 (checked 2026-10-14)](https://www.apple.com/health/pdf/Estimating_Sleep_Stages_from_Apple_Watch_Oct_2025.pdf)
14. [Apple — Track your sleep on Apple Watch and use Sleep on iPhone (checked 2026-10-14)](https://support.apple.com/en-us/108906)
15. [Apple — Sleep apnea notifications on your Apple Watch (checked 2026-10-14)](https://support.apple.com/en-us/120031)
16. [Oura — Sleep Stages, Oura Member Care (checked 2026-10-14)](https://support.ouraring.com/hc/en-us/articles/11752397946003-Sleep-Stages)
17. [Oura — Sleep Contributors, Oura Member Care (checked 2026-10-14)](https://support.ouraring.com/hc/en-us/articles/360057792293-Sleep-Contributors)
18. [Google — Learn about Google Health sleep stages, Google Health Help Center (checked 2026-10-14)](https://support.google.com/googlehealth/answer/14236712?hl=en)
19. [Google — What’s the Sleep Score in the Google Health app, Google Health Help Center (checked 2026-10-14)](https://support.google.com/googlehealth/answer/14236513?hl=en)
20. [Garmin — How Garmin watches track your sleep, calculate sleep score, 30 January 2025 (checked 2026-10-14)](https://www.garmin.com/en-US/blog/fitness/how-garmin-watches-track-your-sleep-calculate-sleep-score/)
21. [WHOOP — Sleep, WHOOP for Developers (checked 2026-10-14)](https://developer.whoop.com/docs/developing/user-data/sleep/)
22. [Samsung — What is Samsung Health? Samsung Australia support (checked 2026-10-14)](https://www.samsung.com/au/support/apps-services/samsung-health/)
23. [Samsung — Samsung’s sleep apnea feature on Galaxy Watch first of its kind authorized by US FDA, Samsung Global Newsroom, 10 February 2024 (checked 2026-10-14)](https://news.samsung.com/global/samsungs-sleep-apnea-feature-on-galaxy-watch-first-of-its-kind-cleared-by-us-fda)
24. [Eight Sleep — How Eight Sleep Pod 6 works (UK site, checked 2026-10-14)](https://www.eightsleep.com/uk/pod-cover)

---

Sheeps shows you no sleep data and no score: no sleep score, no streaks, no badges, and nothing asks how you slept. It plays a new plotless story, written for you, with sentences that shrink and pauses that grow, and ends in silence — something to listen to with the lights off, not something to check in the morning. Sheeps on the App Store: https://apps.apple.com/app/id6749510901