Overview
Sleep is the single heaviest input to how you feel, and it is also the one your watch is most confident and least precise about. This page separates the two: what the stage labels on your wrist actually mean, and which parts of the night Grio’s scoring relies on.
The short version: duration is the reliable part, and Grio weights it accordingly. Efficiency, timing regularity, and the restorative-stage mix are used as small refinements, in that order of influence.
How it’s calculated
What the stage labels mean
| Stage | What it is | How a wrist device infers it |
|---|---|---|
| Awake | Time in bed, not asleep | Movement and heart-rate spikes |
| Core (light) | The majority of a normal night | The default when nothing else fits |
| Deep | Slow-wave sleep, physically restorative | Stillness plus a low, stable heart rate |
| REM | Dreaming sleep, cognitively restorative | Heart-rate variability patterns with low movement |
A wrist device has movement and optical heart rate. Clinical sleep staging uses brain activity, eye movement, and muscle tone. Consumer devices agree with a clinical study well on whether you were asleep and considerably less well on which stage you were in. That gap is not a defect in any one product; it is the limit of what a wrist can observe.
What Grio uses
The sleep contributor to your Condition Score is built primarily from duration, anchored to an absolute curve rather than only to your baseline. Sleep need is individual, but a short night is short however little you usually sleep, so your personal baseline adjusts that grade rather than defining it. The plateau sits at roughly seven and a half to nine hours.
Three optional refinements can adjust that grade, and their weights are ordered deliberately:
- Efficiency: the share of time in bed you were actually asleep. The largest of the three refinements.
- Regularity: how close your sleep midpoint was to your usual one. Smaller.
- Restorative mix: the share of staged minutes labelled deep or REM. The smallest, precisely because it is the least reliable measurement.
Each refinement needs enough baseline history before it applies, and any that is unavailable returns its share to duration rather than to an unrelated contributor.
Edge cases and fallbacks
- No night recorded: the Check-In’s reported sleep takes over the contributor entirely.
- A device that reports sleep but not stages: duration and efficiency still work; the restorative refinement simply does not apply.
- Naps: what reaches Grio is what Apple Health recorded as sleep.
- Raw samples never leave your device: What Grio’s servers store is a privacy-reduced daily projection, not the sleep session itself.
How to read your score
- Total sleep time is the number to act on: it is the most accurate figure on the screen and the one with the largest effect on the score.
- Efficiency below the mid-eighties, night after night, usually means time in bed awake, pointing to a genuinely practical pattern.
- Stage percentages are for trends, not targets: “Get more deep sleep” is not an executable instruction, and chasing a stage number that your device infers indirectly is a reliable way to sleep worse.
- Regularity is underrated: a consistent midpoint is often easier to change than total duration, and it is one of the refinements Grio reads.
What to do with it
- Fix duration first: every refinement is smaller than the thing it refines.
- Then fix consistency: a stable bedtime moves both the regularity refinement and, usually, duration.
- Then look at efficiency: Long wake periods point at the room, the schedule, or a late meal, which are factors you can change.
- Ignore the stage chart on any single night: Read it monthly if at all.
If your watch missed a night, do the Check-In. It is the difference between a score built on your report and a score built on two metrics with no sleep at all.
Troubleshooting & FAQs
- My watch says 40 minutes of deep sleep. Is that number reliable?
- Treat it as a relative indicator rather than a measurement. Wrist devices infer stages from movement and heart-rate patterns, and stage-level agreement with clinical polysomnography is far weaker than their agreement on total sleep time. Grio reflects that by weighting duration heavily and the restorative-stage mix very lightly.
- I forgot to wear the watch. Is the night lost?
- No. Report the night in the Condition Check-In and it takes over the sleep contributor. Missing values stay missing rather than becoming zero, so a night nobody recorded is not scored as a night of no sleep.
- Why did a nine-hour night not score better than an eight-hour one?
- The sleep curve plateaus: roughly seven and a half to nine hours sits at the top, and both shorter and much longer nights fall away. An unusually long night is not penalised for being unusual, but it is not treated as extra recovery credit either.