Sleep
An 82 on one wrist and an 82 on another are not the same 82. Every manufacturer built its own scale out of its own inputs, and none of them publishes the recipe.
Broadly, 80 and above is good and 90 and above is excellent on most devices — but the scales are not interchangeable, because each brand weights duration, timing, interruptions and heart-rate data differently. Your own trend over weeks is far more informative than any single night’s number.
| Device | Bands | What it weights |
|---|---|---|
| Garmin | 90–100 excellent · 80–89 good · 60–79 fair · <60 poor | Folds in overnight stress derived from HRV |
| Oura | 85–100 optimal · 70–84 good · 60–69 fair · <60 pay attention | Built from seven separate contributors |
| Fitbit | 90–100 excellent · 80–89 good · 60–79 fair · <60 poor | Duration, deep and REM, overnight heart rate, restlessness |
| Apple Watch | 96+ very high · 81–95 high · 61–80 OK · <40 very low | Duration 50 points, bedtime consistency 30, interruptions 20 |
| Samsung | 90–100 very good · 70–89 good · 50–69 fair · <50 poor | Duration, efficiency, wake time, physical and mental recovery |
| WHOOP | Sleep performance as a percentage of need | Not a like-for-like score; hours versus need is only one input |
Note how different the same number is. An 82 is comfortably “good” on Garmin and Fitbit, merely “high” on Apple, sits in Oura’s “good” band below optimal, and is mid-table on Samsung. Comparing scores with a friend on another brand is a conversation about two different measurements.
Almost every sleep score is built from four ingredients in some proportion:
None of this is a clinical measurement, and no manufacturer fully discloses its weighting. Treating the score as a precise fact about your night is the main way people get misled by it.
The finding that surprises people most: a stable schedule raises the score more reliably than extra hours do. Going to bed and getting up within about an hour of the same time, seven days a week, produces higher scores and lower morning stress readings than an erratic schedule with the same average sleep duration.
This is why a long weekend lie-in often scores worse than the Tuesday night that was an hour shorter. The device is not malfunctioning; it is picking up the circadian disruption.
Three numbers under the score are more actionable than the score:
And the one nobody mentions: eating far too little. A steep calorie deficit reliably fragments sleep and raises overnight heart rate. If the score fell off a cliff the week the diet started, that is the explanation — and it is invisible to any device that cannot see what you ate.
There is a documented pattern of people becoming anxious about their sleep data, which then damages the sleep. If you find yourself checking the score before you have got out of bed and letting it set your expectations for the day, the useful intervention is to look at the weekly average once, on a Sunday, and ignore the daily figure entirely.
On most devices, 80 and above is good and 90 and above is excellent. Garmin and Fitbit both treat 80 to 89 as good and 90 plus as excellent, Oura calls 85 and above optimal, Apple Watch classes 81 to 95 as high, and Samsung treats 70 to 89 as good. The scales are not directly comparable between brands.
Yes on most scales. An 80 sits at the bottom of the good band on Garmin and Fitbit, comfortably inside good on Samsung, and just under high on Apple Watch. On Oura it is good but below the optimal threshold of 85. A run of 80s week after week is a healthy pattern.
Sleep scores weight duration and continuity heavily and estimate sleep stages from movement and heart rate rather than measuring them. A short but solid night, a late bedtime or a restless tracker reading can all lower the score without matching how you feel. How you feel is the better guide.
Duration and timing are reasonably accurate on modern devices. Sleep stage estimates are much less so, since consumer wearables infer stages indirectly from movement and heart rate. No sleep score is a clinical measurement, and each manufacturer keeps its formula proprietary.
Increase time in bed, keep bedtime and wake time consistent within about an hour, avoid alcohol within three hours of sleep, move the caffeine cut-off to eight hours before bed, and keep the room cool and dark. A steep calorie deficit also lowers sleep quality noticeably.
A sleep score is only useful next to what caused it. NutriMacro puts sleep, resting heart rate and recovery on the same day as the training you did and the food you ate, which is where the explanation usually is — and it reads your sleep from Apple Health, Health Connect or Samsung Health, so you do not need to buy another ring or band.
NutriMacro is a nutrition and training tracker for iPhone and Android. It logs food by barcode, photo or search, counts food in slices and units rather than grams, and keeps strength training, GPS activities, sleep and recovery in the same daily view. It runs in the language of your phone across 108 countries.
Read next: Why is my sleep score low after 8 hours? How much deep sleep do you need? Why am I always tired?
