Are sleep trackers accurate? What the numbers can and cannot tell you

What a tracker actually measures

A wrist or ring tracker does not measure sleep. It measures things that change during sleep: movement (from an accelerometer), heart rate and its beat-to-beat variability (from an optical sensor) and, on many newer devices, overnight blood oxygen. An algorithm then guesses, minute by minute, whether you were awake or asleep and, often, which sleep stage.

A clinical sleep study, called polysomnography, works differently. According to the National Heart, Lung, and Blood Institute, it records brain waves, eye movements, muscle activity, heart rate, breathing effort and blood oxygen through the night. Sleep stages are defined by brain-wave patterns, so a device without a brain-wave sensor can only estimate them.

Where trackers do reasonably well

The honest answer to "are sleep trackers accurate?" is: it depends on the number, and on who is wearing it. A 2021 study in the journal Sleep compared seven popular consumer devices with polysomnography in 34 healthy adults aged 18 to 35 with no sleep disorders. Most detected sleep about as well as, or better than, the research-grade wrist monitors (actigraphy) used in sleep research for decades. That makes them reasonably useful for:

• Total sleep time on a typical night.
• Bedtime and wake time, and so schedule regularity.
• Long-term patterns, such as consistently sleeping less on work nights.

Two cautions. Those figures come from young, healthy sleepers over three laboratory nights that included one deliberately disrupted night — and the devices performed worse on the disrupted night. Expect lower accuracy in older adults, people with insomnia, sleep apnea or otherwise broken sleep, and shift workers, which is to say many people reading this. Even there, the weak spot was wakefulness: devices recognized sleep far better than wake, so time lying still and awake is often counted as sleep. With insomnia, your tracker may report more sleep than you got.

Where they fall short

Sleep staging is where the numbers become unreliable. In the same study, devices showed mixed and often poor results for light, deep and REM sleep, frequently mislabelling one stage as another. The sleep specialists who first described orthosomnia put it plainly: consumer wearables cannot accurately distinguish sleep stages and have poor accuracy for detecting wake after you first fall asleep.

So "1 hour 12 minutes of deep sleep" is an estimate built on an estimate, with no way to check it. The same goes for the "sleep score": each company uses its own private formula, so one night can earn very different scores on different wrists, and none is a medical measurement.

Orthosomnia: when the data keeps you awake

In 2017, a group of sleep specialists coined the term orthosomnia for a pattern seen in clinic: people so focused on perfect tracker numbers that the pursuit itself was harming their sleep. One woman with normal deep sleep on a laboratory study still asked why her device said she slept badly; others spent longer and longer in bed to raise their totals, a reliable way to make insomnia worse.

The milder version is common: checking your score before deciding how tired you feel, or dreading a low number more than a bad night. Our guide to insomnia and what actually helps explains why watching the clock, or the app, backfires.

How to use the numbers without losing sleep

A tracker can still help, treated as a weather log rather than a report card.

• Look at weeks, not nights. One night is noise; a month shows whether your sleep is drifting later, getting shorter or holding steady.
• Trust the trend, question the detail. Total sleep and schedule regularity are the most dependable figures; stage breakdowns and scores are rough hints.
• Let how you feel be the tiebreaker. Refreshed and functioning well? A mediocre score does not change that. Wrecked despite a "great" score? Believe your body.
• Do not chase deep sleep. You cannot will yourself into a stage, and it is the least accurate number reported.
• Take breaks. If the app makes you tense, leave it off for a couple of weeks.

What a tracker cannot diagnose

The American Academy of Sleep Medicine's position statement on consumer sleep technology says these devices are no substitute for a clinical evaluation, and that anything intended to diagnose or treat a sleep disorder must be cleared by regulators and tested against the clinical gold standard. Most trackers are neither, and even the few cleared features are screening aids, not diagnoses.

Sleep apnea is the clearest example. Diagnosis means measuring breathing and oxygen overnight and counting how often breathing pauses or becomes shallow. A wrist tracker sees only the downstream ripples, such as a restless night, a raised heart rate or a dip in blood oxygen, and cannot confirm what caused them.

Since 2024, a few mainstream smartwatches and rings have carried regulator-cleared features that flag signs of possible sleep apnea. The US Food and Drug Administration classifies them as over-the-counter devices for assessing sleep apnea risk in undiagnosed adults. They target moderate-to-severe apnea, can miss cases, and their labelling says they do not diagnose. A notification is a reason to see a doctor about a proper sleep study, not an answer; no notification does not rule apnea out, and your symptoms count for more than your watch.

When a real sleep study is the answer

Talk with a doctor, whatever your tracker says, if you snore loudly, someone has seen you gasp or stop breathing in your sleep, you regularly wake unrefreshed after a full night in bed, or you are sleepy enough to nod off while sitting still. Morning headaches and getting up often at night to urinate are less obvious signs. If you have nodded off, or nearly nodded off, at the wheel, do not wait: stop driving until you have been assessed and see a doctor promptly.

That may mean a night in a sleep laboratory or a simplified home sleep test that mainly measures breathing and oxygen. The STOP-Bang self-check can gauge your risk first, and our overview of sleep apnea symptoms, risks and treatments covers what comes next.

In summary

Consumer sleep trackers estimate sleep from movement, heart rate and sometimes blood oxygen, not from the brain. They are reasonably good at total sleep time and schedule patterns in healthy young adults, poor at sleep stages and quiet wakefulness, and likely worse when sleep is broken. A few can now flag possible sleep apnea; none can diagnose it or any other sleep disorder. Trends over weeks can help; nightly verdicts feed the anxiety that keeps people awake. If the real question is your breathing at night, the answer is a sleep study, not a wrist.