Most accurate sleep tracker? What lab tests show

The short answer

• ≥93%

  Of minutes asleep that each of 7 consumer trackers got right in a lab test

• 18–54%

  Of minutes awake that the same trackers spotted

• 8min

  Gap between Apple Watch Series 8 and a lab sleep test for total sleep, on average

• −43min

  Deep sleep the Apple Watch Series 8 missed, compared with the lab test

Key findingTrust your tracker most for how long and when you slept. Treat time awake and deep or REM sleep as rough guesses, especially after a bad night.

How do scientists test a sleep tracker?

They compare it with polysomnography: a full overnight sleep test in a lab, which records brain waves, breathing and more. Experts score each 30 seconds of the night as awake, light sleep, deep sleep or REM sleep (rapid eye movement sleep).

Even experts don’t agree perfectly. Two trained scorers reading the same night agree only about 80% of the time.

Two numbers matter most. One is how much real sleep the tracker calls sleep. The other is how much real awake time it calls awake. Most trackers are good at the first and weak at the second.

The reason is simple. Lying still while awake looks a lot like sleep to a device on your wrist. Older trackers used movement alone. Newer ones add heart rate, which may help, but a 2019 review said that still had to be proven.

Are sleep trackers accurate? The seven-device test

This was a careful experiment in a lab. No device company paid for it or helped with it. Each device was tested on 11 to 20 people.

Missing awake time means counting too much sleep. Over 8 hours in bed, the two Garmin watches added about 45 minutes of sleep on average. The Fitbit was off by under 3 minutes on average. Yet single nights could still be about 42 minutes under or 47 minutes over.

Sleep stages were weaker still. Most devices missed 30% to 50% of deep sleep and REM sleep. The stages they got wrong were mostly called light sleep. All devices did worse on nights with broken sleep.

Is the Apple Watch sleep tracker accurate?

Neither the seven-device test nor the 2019 review covered an Apple Watch, so we added a 2024 lab study that did. It compared an Apple Watch Series 8, a Fitbit Sense 2 and an Oura Ring Gen3 with a full sleep test. It covered one night in 35 healthy adults aged 20 to 50.

The Apple Watch caught 97% of minutes asleep. It agreed with the lab test 93% of the time on asleep versus awake. Its total sleep was only 8 minutes off on average (our calculation). But it found only about half (50.5%) of the deep sleep, and called much of the rest light sleep.

The Oura ring wasn’t clearly different from the lab test on any stage. The Fitbit counted 15 minutes too little deep sleep. Data were also lost: the Apple Watch recorded nothing for 6 of the 35 people, even though it was charged and set up. Keep in mind that the lead author advises Oura and is paid by the company.

A 2026 review of Apple Watch accuracy found the same picture in three sleep studies with 221 people. The watch was good at telling sleep from wake, but poor at deep sleep. Across studies, its authors noted, the Apple Watch matched lab tests less well than WHOOP, Fitbit and Garmin devices.

An older Apple Watch (Series 6) did worse at spotting awake time in a 2022 test of six devices. It caught only 26%, against 56% to 57% for WHOOP, Oura and a forehead patch. But that test read the watch through a separate app, in 5-minute blocks, not with Apple’s own sleep stages. The authors say coarse data like this may lower agreement.

Apple’s own sleep score, and how it differs from Bodyvault’s, is explained in our guide to the Apple Watch sleep score.

So which is the most accurate sleep tracker?

There isn’t one clear winner. In the seven-device test, the Fitbit Alta HR came out best overall and the two Garmin watches worst. In the six-device test, the Oura ring, WHOOP and the Somfit patch did best on sleep stages. In the three-device test, the Oura ring came closest on stages.

Each test used different models, people and methods, so you can’t rank them all on one list. Companies also change their software without notice. A model is often replaced by the time a study about it comes out. So a good result may not apply to the one you buy.

For most people, the best sleep tracker is one you will wear every night, the same way, so you can follow your trends. A watch, a ring and a bedside device all did well at telling when people were asleep.

Claims and evidence

• ClaimMy tracker knows exactly how long I slept.

  EvidenceOn average, many were within minutes. But single nights could be 40 minutes or more off, and errors grew on nights with broken sleep.

• ClaimThe deep sleep number on my watch is accurate.

  EvidenceThe Apple Watch Series 8 counted 51 minutes of deep sleep against 94 in the lab. In the seven-device test, most devices missed 30% to 50% of deep sleep.

• ClaimA sleep tracker can tell me if I have a sleep disorder.

  EvidenceNo. A 2019 review found too little evidence. It quotes the American Academy of Sleep Medicine, which says consumer sleep devices can’t be used to diagnose or treat sleep disorders.

• ClaimIf my tracker says I slept well, I did.

  EvidenceTrackers tend to count quiet awake time as sleep, so they count too much sleep on bad nights. A 2019 review warned that false “good sleep” results can delay getting help.

What this means for you

What the research can’t tell you yet

• Short tests. The studies lasted one to three nights in a lab, not weeks at home.
• Who was studied. Mostly healthy adults under 50. People with insomnia or sleep apnoea were rarely tested.
• Old models. Many tested models are no longer sold, and software updates can change results.
• Company links. Some researchers had ties to Oura or WHOOP. The seven-device test was independent.
• Naps and timing. Trackers were poor at catching naps. In one study of one tracker, the time it said people fell asleep was more than 20 minutes late on about 1 night in 4.

How Bodyvault uses this research

Bodyvault reads your nights from Apple Health or Health Connect. Its Sleep score puts the most weight (40%) on how long you slept compared with what you need. Time awake in bed, a regular schedule and your overnight heart rate count for 15% each. Deep and REM sleep together count for only 15%, and after 28 nights with stages they are judged against your own usual levels.

General information about fitness and nutrition, not medical advice. If you have a health condition, are pregnant or take medication, talk to your doctor before changing your diet or training.