Sleep trackers

The morning numbers are persuasive. People check the sleep score before deciding how they feel, and some wake feeling fine only to be told by a ring that they slept badly. Those numbers are worth understanding before you trust them, because a ring or watch is measuring something quite different from what it reports.

What the device can and can't see

The stages of sleep (light, deep and dreaming sleep) are defined by the brain's electrical activity. The only way to measure them properly is polysomnography, an overnight recording in a sleep laboratory of brain waves, eye movements and muscle activity alongside breathing and heart rhythm.

A ring or watch can't see any of that. It estimates sleep from two things it can measure at the wrist or finger: movement, through a motion sensor, and heart rate, through a small light shining into the skin. From those it infers when you were asleep and guesses which stage you were in. Readiness, recovery and sleep scores are then each company's own calculation built on top of those estimates, and they aren't measurements in themselves.

The weak spot follows from the method. Lying still and awake in the dark looks a great deal like sleep to a motion sensor, so devices tend to be good at spotting sleep and poor at spotting time awake.

How well they do against the laboratory

Chinoy and colleagues (2021) tested seven consumer devices against polysomnography in 34 healthy young adults over three nights in a sleep laboratory, one of them deliberately disrupted. Every device caught at least 93% of the time people were asleep, but they correctly picked out only 18 to 54% of the time they were awake. Their sleep stages were inconsistent, and they tended to do worse on the disrupted night.

Newer devices show a similar pattern. Schyvens and colleagues (2025) compared six wrist-worn trackers, including the Apple Watch Series 8, Whoop 4.0 and two Fitbits, with polysomnography in 62 adults, most of them men in middle age, over one laboratory night. All caught more than 90% of sleep, spotted wake only 29 to 52% of the time, and agreed with the laboratory on sleep stages only fairly to moderately.

Rings are no different in principle. In patients at a Berlin university sleep laboratory with a mix of sleep and other medical conditions (45 in all, with usable Oura data from 31), the Oura ring's average total sleep came within 12 minutes of the laboratory's figure across the group, yet on individual nights the difference was often large (Herberger et al., 2025). The authors concluded that these individual errors rule the rings out for use in clinical sleep medicine.

So we're reasonably confident a current device will tell you roughly when you slept and how long, averaged over several nights. We're much less confident in any single night, in time awake, or in the split between light, deep and dreaming sleep, and the bad nights you most want explained are the ones devices measure worst.

When the numbers become the problem

Baron and colleagues (2017) described patients who sought treatment at their sleep clinic because of what their tracker showed, and whose efforts to perfect the numbers had become part of the sleep problem. They called it orthosomnia. Those were individual case reports, so we don't know how common it is, and it fits a well-cited account of insomnia in which close attention to sleep, and effort to make it happen, get in its way (Espie et al., 2006).

Why your plan reads the diary first

Your plan is built on a sleep diary of the kind sleep researchers standardised as the Consensus Sleep Diary (Carney et al., 2012): when you went to bed, how long you think it took to drop off, how long you were awake and how you felt in the morning. It records your experience of the night, and that's what the plan sets out to change. When the diary and the device disagree, the diary decides; the device is a second opinion.

Keep wearing it if you like it. We compare its heart rate trends at week six, but we don't grade or band its scores, and if it starts to make you anxious about sleep, tell us: the plan works without it.

A consumer device can't diagnose a sleep disorder. If someone has noticed you snore loudly or stop breathing in your sleep, or you're very sleepy in the day despite a full night, see your GP whatever the device says.