Farnflower Newsletter
Sleep

Are sleep trackers accurate?

Rings and watches matched a sleep lab 86 to 89 percent of the time on asleep versus awake, but only 50 to 65 percent on sleep stages. What to trust.

Which sleep stealers apply to you? Take the one-minute test

Sleep trackers in the lab

Trackers know fairly well when you sleep, much less well how.

Average agreement of six consumer sleep trackers with a sleep lab over the night, in percent

  1. Asleep or awake 88 %
  2. Which sleep stage 57 %
Farnflower Source: Miller et al., 2022 · 53 adults, one night

For when and how long you sleep, reasonably; for sleep stages, not very. In sleep-lab comparisons, current rings and watches detected sleep well but often missed short periods of being awake, so they tend to flatter a restless night. Their deep, light and REM sleep figures matched the lab only a little more than half the time.

Key takeaways

  • Trackers are good at recognizing sleep and weak at recognizing wakefulness, so they tend to overestimate sleep.
  • They performed worse on nights with disrupted sleep, the nights you most want to understand.
  • Trends over weeks are more useful than any single night’s score.
  • Some people start sleeping worse because they worry about their tracker data, which sleep doctors call orthosomnia.
  • Their heart readings at night react clearly to things like a glass of wine.

How accurate are sleep trackers compared with a sleep lab?

Sleep trackers are quite accurate at telling sleep from wake. Australian researchers had 53 young adults spend a night in a sleep lab wearing six popular trackers at once, among them a smartwatch, two sports watches, a smart ring, a fitness band and a forehead sensor, alongside full polysomnography, the clinical gold standard. Moment by moment through the night, the devices agreed with the lab on whether the person was asleep or awake 86 to 89 percent of the time. On the exact sleep stage, agreement fell to between 50 and 65 percent (Miller et al., 2022). The authors concluded that all six are valid for the timing and duration of sleep and all six need improvement for sleep stages.

Do sleep trackers overestimate how much you sleep?

Yes, because they miss many short periods of being awake. The high agreement for sleep versus wake hides this weakness. Most of a night is sleep, so a device that guesses “asleep” most of the time scores well even if it misses many wake periods. A US study of seven consumer devices in 34 young adults found exactly that: the devices caught at least 93 percent of the minutes spent asleep, but only between 18 and 54 percent of the minutes spent awake. They also did worse on nights when the researchers deliberately disrupted sleep (Chinoy et al., 2021).

So if your tracker says you slept seven hours after a night you remember as broken, it is probably counting some of your wake time as sleep.

Why are sleep stages so hard to measure from the wrist?

Sleep stages are hard to measure from the wrist because they are defined by brain waves, and a ring or watch can’t see them. A sleep lab records brain activity, eye movements and muscle tone with electrodes on the head. Consumer devices infer the stages indirectly from movement and heart signals, and in the lab comparisons those guesses were inconsistent (Chinoy et al., 2021).

This is the part of the app I’d take least seriously, because a “deep sleep” score that dropped by a few minutes tonight is within the device’s error and tells you nothing about your brain.

Can a sleep tracker make your sleep worse?

A sleep tracker can make sleep worse for some people. Sleep researchers in Chicago described patients who came to their clinic convinced they had a sleep problem because of light or restless sleep on their tracker, and who started chasing a perfect score. They named it orthosomnia. For these patients, the tracker data often felt more convincing than the clinic’s own measurements (Baron et al., 2017).

What are sleep trackers actually good for?

Sleep trackers are most useful for patterns over time. They are reliable enough to show when you go to bed and get up, and whether those times drift, which can matter for your health (the regular bedtime article explains why). They show trends in sleep duration across weeks. And their heart readings at night react clearly to things like alcohol, as the wine and sleep article explains. What your resting heart rate says about health is covered in What is a good resting heart rate?

What I’d do

  • Look at weekly averages and bedtimes, not at last night’s score.
  • Ignore small changes in deep or REM sleep.
  • Trust your own sense of how rested you feel when it disagrees with the app.
  • Take the tracker off for a week if checking it makes you anxious about sleep.
  • See a doctor about persistent sleep problems instead of trying to fix them with app scores; a sleep lab measures what a ring can’t.

Frequently asked questions

Are rings better than watches? In the Australian comparison, the ring and two other devices agreed somewhat better with the lab on sleep stages than the watches did, but none was good enough for stages (Miller et al., 2022). Each new model needs its own test.

Why does my tracker say I slept when I was awake? Trackers recognize sleep far better than wakefulness, so short wake periods often get counted as sleep, especially on restless nights (Chinoy et al., 2021).

Can a sleep tracker detect sleep apnea? Some devices now offer features that look for breathing problems, but a tracker can’t replace a medical diagnosis. If you snore loudly or wake up gasping, ask your doctor about a sleep study.

Is a bedside device as good as a wearable? In the US study, several devices that sit beside or under the bed performed as well as a research-grade wrist monitor for sleep versus wake (Chinoy et al., 2021).

Sources

  • Baron, K. G., Abbott, S., Jao, N., Manalo, N., & Mullen, R. (2017). Orthosomnia: Are some patients taking the quantified self too far? Journal of Clinical Sleep Medicine, 13(2), 351–354. https://doi.org/10.5664/jcsm.6472
  • Chinoy, E. D., Cuellar, J. A., Huwa, K. E., Jameson, J. T., Watson, C. H., Bessman, S. C., et al. (2021). Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep, 44(5), zsaa291. https://doi.org/10.1093/sleep/zsaa291
  • Miller, D. J., Sargent, C., & Roach, G. D. (2022). A validation of six wearable devices for estimating sleep, heart rate and heart rate variability in healthy adults. Sensors, 22(16), 6317. https://doi.org/10.3390/s22166317

One question a month, answered properly