Why Your Wearable's Sleep Score and How You Feel Often Disagree
Sleep scores are algorithm-derived composites from accelerometer and heart rate data. Research consistently shows that objective sleep metrics and subjective sleep quality often don't match — and this disagreement is normal, not a sign your device is broken.
On this page
TL;DR
Wearable sleep scores are proprietary estimates from motion and heart rate data, not direct measurements of restorative sleep quality. Research consistently shows that objective sleep metrics and subjective sleep experience often disagree — this is a known and well-studied gap in sleep science, not a sign your device is broken. Both the number and how you feel are useful signals, but neither tells the whole story on its own.
What a sleep score actually is
When your Apple Watch or wearable presents a sleep score or stage breakdown in the morning, it is easy to treat the number as an authoritative assessment of how well you slept. It is not.
A sleep score is a proprietary composite figure derived from algorithm-interpreted sensor data — primarily accelerometer readings (to estimate movement and stillness) and heart rate patterns (to infer physiological state). Different devices use different algorithms, different weightings for sleep stages, different thresholds for what counts as "good" sleep, and different approaches to handling the inevitable ambiguities in the data.
Apple Watch estimates sleep stages using a combination of wrist accelerometry and heart rate from its optical PPG sensor. Apple has published a technical white paper describing this approach and its validation against polysomnography (Apple, 2023). What the validation shows is instructive: Apple Watch is very good at detecting whether you are asleep or awake (sensitivity ~97–98%). But its ability to distinguish which sleep stage you are in is considerably more limited — deep sleep is correctly classified roughly 62% of the time; REM around 81% of the time. The watch struggles most with the boundary between light sleep and quiet wakefulness, which is a challenging estimation problem for any wrist-based device.
The sleep score synthesises these imperfect stage estimates into a single number. The number is useful for tracking trends over time — consistently low scores that correlate with how you feel are meaningful. A single night's score, taken alone, should be interpreted cautiously. The related question — whether sleep duration, the other commonly tracked sleep metric, tells you what you think it does — is explored in why sleep duration alone is a misleading health metric.
Why objective and subjective sleep often disagree
The gap between how a device scores your sleep and how you feel is not a technology failure. It is a fundamental finding of sleep research.
Sleep science has long distinguished between objective sleep measures — what polysomnography (PSG) records about brain waves, eye movements, and muscle tone — and subjective sleep quality — how rested, refreshed, or restored a person reports feeling. These two things do not map onto each other as neatly as intuition suggests.
A 2024 study published in PNAS used in-home EEG to measure sleep objectively in participants' own beds and compared those measures to subjective assessments. The researchers found consistent and systematic discrepancies — objective sleep metrics reliably predicted only a modest fraction of variance in subjective sleep quality. People could sleep very similarly on the objective measure and report dramatically different levels of restedness, and vice versa.
The reasons for this gap are multiple:
Individual variation in sleep needs. Population-average sleep stage benchmarks (e.g. "adults need 90 minutes of deep sleep") do not apply uniformly. Some people genuinely feel rested with less slow-wave sleep than the average; others feel unrestored even after a night that looks textbook on paper.
Mood and expectation effects. How you feel on waking is influenced by what time you woke, recent stress, expectations, and the psychological effect of looking at a number before you've had time to assess how you actually feel. A controlled study using sham actigraphy feedback in people with insomnia found that participants who were told their sleep had been poor (regardless of how they had actually slept) reported significantly lower daytime alertness and more sleepiness than those told their sleep had been good — a kind of nocebo effect from their own data (Gavriloff et al., J Sleep Res 28(1), 2019, doi:10.1111/jsr.12726).
Emotional and mental restoration. Sleep is restorative in ways that no sensor currently measures — memory consolidation, emotional processing, hormonal regulation. A night that scores modestly on movement and heart rate metrics may have been deeply restorative in dimensions the device cannot observe.
Sleep inertia: the morning grogginess that isn't about sleep quality
One of the most common experiences of objective-subjective mismatch happens at the moment of waking. You open your eyes feeling genuinely foggy and fatigued — and then you check your sleep score and see a number that looks reasonable. Are you broken? Is the watch wrong?
Very often, neither. What you may be experiencing is sleep inertia — a well-characterised physiological state of impaired cognitive and sensory-motor performance in the minutes immediately following waking. Sleep inertia is driven by the brain transitioning out of sleep, and it is most pronounced when you wake from deep slow-wave sleep (N3).
Research on sleep inertia (PMC 6710480) shows that it typically lasts 15–30 minutes and can involve subjective grogginess, difficulty with complex reasoning, slower reaction times, and a general sense of being unrefreshed — all regardless of total sleep duration or sleep quality during the night. A high sleep score the previous night does not prevent sleep inertia if your alarm happened to fire during a deep sleep episode.
Wearables do not track sleep inertia. They measure your sleep until you wake up. How you feel in the first 30 minutes of your day is a post-sleep experience that no sleep score captures.
When the disagreement matters and when it doesn't
Not all score-versus-feeling gaps are equally meaningful.
When to trust how you feel over the score: A single morning on which the score looks fine but you feel rough is most likely explained by sleep inertia, alarm timing, or the device's stage estimation inaccuracy. If you feel genuinely energised, focused, and functional through the day, the low score was probably noise.
When to take a persistent low score seriously: If your sleep score trends notably lower across several consecutive weeks — and this correlates with increasing fatigue, worse concentration, lower exercise performance, or a rising resting heart rate — the alignment of objective and subjective signals makes the pattern more meaningful. The combination of device data and how you feel tells a richer story than either alone.
When to look beyond the score: Sleep scores do not screen for sleep disorders. If you habitually feel unrested despite seemingly adequate sleep, snore loudly, or have been told you stop breathing during sleep, those are symptoms that warrant professional evaluation — not further optimisation of your wearable data.
How to use sleep data more effectively
A few practical habits shift sleep tracking from a daily verdict to a genuinely useful information source:
Track your subjective feeling alongside the score. Many sleep apps allow you to log how rested you feel on waking. This creates paired data — subjective and objective — that is far more interpretable over time than the device score alone. Consistently misaligned scores and feelings across multiple weeks is a real signal.
Look at trends, not individual nights. Seven-day or monthly averages of sleep duration and HRV during sleep carry far more information than any single night. Individual nights are noisy; trends are meaningful.
Note your sleep environment and behaviour. Alcohol, late caffeine, screen exposure before bed, and room temperature all affect both objective sleep metrics and how you feel on waking. Annotating your data with these factors makes it possible to distinguish a bad night driven by a glass of wine from a pattern that needs attention.
Give yourself 30 minutes before judging your day. Sleep inertia is real and time-limited. Your subjective assessment of how rested you feel is more accurate mid-morning than the moment your alarm fires.
Where Sam Health fits in
Sam tracks your sleep data as a trend rather than a verdict. Rather than surfacing a single night's score as a readout to react to, Sam shows how your sleep duration, HRV during sleep, and resting heart rate are moving across your recent nights relative to your own baseline — which is the approach the research supports.
When your sleep metrics and your subjective energy align — both consistently lower than your norm — Sam surfaces the pattern. When they diverge — a low score on a night you feel fine — Sam's context helps you recognise that divergence as the expected noise it usually is. Both the number and how you feel remain yours to weigh. For a complete overview of the wearable metrics Sam works with, see the wearable biomarkers that actually matter.
Try Sam HealthSources
- Apple Inc. Estimating Sleep Stages from Apple Watch. September 2023. apple.com
- Zerouali Y, et al. Discrepancies between subjective and objective sleep assessments revealed by in-home electroencephalography during real-world sleep. PNAS. 2024. doi.org/10.1073/pnas.2412895121
- Hilditch CJ, McHill AW. Sleep inertia: current insights. Nat Sci Sleep. 2019;11:155–165. PMC 6710480
- Jaworski B, et al. Apple Watch Sleep and Physiological Tracking Compared to Clinically Validated Actigraphy, Ballistocardiography and Polysomnography. IEEE EMBC. 2023. PMID 38083143
- Gavriloff D, Sheaves B, Juss A, Espie CA, Miller CB, Kyle SD. Sham sleep feedback delivered via actigraphy biases daytime symptom reports in people with insomnia: Implications for insomnia disorder and wearable devices. Journal of Sleep Research. 2019;28(1):e12726. doi:10.1111/jsr.12726 (PMID 29989248)
Frequently Asked Questions
Why does my sleep score look good when I feel tired?+
Several things can produce this gap. If you woke from deep sleep (slow-wave sleep) when your alarm fired, you may experience sleep inertia — a physiological grogginess that can persist 15–30 minutes regardless of how your total sleep looked. Your wearable also cannot measure how mentally or emotionally restorative sleep felt. A high score reflects measured patterns, not your subjective experience.
Why do I feel rested after a sleep my wearable scored poorly?+
Sleep score algorithms weight stages and duration in fixed ways that don't account for individual variation. Some people genuinely feel rested after less total sleep or less slow-wave sleep than population averages assume. Your device doesn't know how you feel — it infers sleep quality from motion and heart rate patterns.
How accurate are Apple Watch sleep stages?+
Apple Watch reliably detects whether you are asleep or awake (sensitivity around 97–98%). Stage classification is less precise: per-epoch classification accuracy is around 62% for deep sleep and 81% for REM in Apple's own validation. The watch struggles most with distinguishing light sleep from periods of quiet wakefulness. Independent studies that compare total time-per-stage to polysomnography report different-looking numbers (e.g. a 43-minute mean underestimate of deep sleep) because they measure duration error rather than per-epoch agreement — both are valid views of the same underlying limitation.
What does a wearable sleep score actually measure?+
Sleep scores combine estimates of sleep duration, sleep stage distribution, and sometimes heart rate or HRV during sleep into a single number using a proprietary algorithm. The algorithm and weightings differ by device and brand. The score is an interpretation of sensor data, not a direct physiological measurement.
Should I worry if my sleep score is consistently low?+
A consistently low score is more informative than an occasional low score, and it becomes more meaningful when it aligns with how you actually feel — persistent fatigue, difficulty concentrating, or low motivation over several days. A low score that doesn't correspond to any subjective symptoms is less concerning; a low score that correlates with how you feel is worth paying attention to.
Can wearables detect sleep apnea?+
Some Apple Watch models with the Sleep Apnea Notifications feature can notice breathing disturbances during sleep that may be consistent with moderate-to-severe sleep apnea, but this feature is not a diagnostic tool and requires clinically validated follow-up. A sleep score alone cannot diagnose sleep apnea or any other sleep disorder.
What is sleep inertia and why does it affect how I feel in the morning?+
Sleep inertia is a short-lived state of cognitive and physical grogginess that follows waking, particularly when waking from deep slow-wave sleep. It typically lasts 15–30 minutes and reflects the brain transitioning from a deeply restorative state to wakefulness. It has nothing to do with how long or well you slept overall.