TL;DR
- Most sleep scores are a weighted mix of duration, sleep timing, wake after sleep onset, and one or more signals like heart rate and HRV.
- Two devices can give different scores on the same night because they use different sensors, different sleep stage models, and different weightings.
- Use your sleep score as a trend signal and a behavior feedback loop, not a diagnosis.
- The most actionable split is usually: sleep duration + sleep consistency + signs of strain (HR and HRV).
- Century is building a Whoop-like layer on top of Apple Health so you can get better guidance without switching wearables.
What a sleep score is trying to do
A sleep score is an attempt to compress a complicated night into a single number.
That number is not the same thing as sleep quality.
In practice, sleep quality is influenced by:
- how much sleep you got
- how consistent your sleep schedule is
- how often you woke up
- how your body handled recovery (autonomic nervous system, stress)
- your training load, illness, and alcohol
Most consumer wearables cannot measure sleep directly. They estimate it.
The ingredients most sleep scores use
Different brands use different names, but the underlying inputs tend to be similar.
1) Sleep duration
This is the biggest driver for most scores.
Typical scoring logic:
- you get rewarded for getting close to a target duration
- you get penalized for sleeping too little
- you might get slightly penalized for sleeping far longer than usual (often because it correlates with illness or sleep debt)
Actionable takeaway:
- If your sleep score is low and your sleep duration was low, do not overthink it. You found the reason.
2) Sleep timing and consistency
Many algorithms reward consistency because it predicts recovery, mood, and performance.
They look at:
- bedtime and wake time regularity
- whether you slept at your usual circadian window
This is one reason a night of 7.5 hours can score lower than 7 hours if it was far off your normal schedule.
3) Fragmentation and wake time
A common metric is WASO (wake after sleep onset).
Even if you slept long enough, lots of micro-wakes can reduce the score.
Wearables infer this from motion and heart rate patterns.
4) Heart rate (HR)
Many sleep scoring models include overnight heart rate, often focusing on:
- average heart rate during sleep
- the lowest heart rate
- whether heart rate stayed elevated compared to your baseline
Elevated overnight HR often correlates with:
- late meals
- alcohol
- dehydration
- heat
- illness
- heavy training load
5) Heart rate variability (HRV)
Some devices incorporate HRV directly. Others use it indirectly through a recovery score.
Key nuance:
- HRV is naturally variable day to day.
- The useful signal is typically your baseline and trend, not a single reading.
6) Sleep stages (light, deep, REM)
Stages are the most tempting part of sleep tracking and also the part where devices disagree most.
Wearables estimate stages based on patterns in:
- motion
- heart rate
- sometimes skin temperature or SpO2
They do not measure brain activity (the gold standard in sleep science).
Practical takeaway:
- Treat stage breakdowns as "directionally useful".
- Do not try to optimize a single stage in isolation.
Why two devices can disagree on the same night
If you have ever worn two trackers and gotten two different sleep scores, you are not crazy.
Common reasons:
Different sleep detection
- one device calls you asleep earlier or later
- one device labels still time in bed as sleep
Different stage models
- staging is a model output, not a direct measurement
Different weightings
- one score prioritizes consistency
- another prioritizes duration
- another heavily weights HRV
Different baselines
- many scores are relative to your recent weeks
- if you changed training or schedule, the baseline can lag
This is why the best use of a sleep score is within one ecosystem over time.
A simple mental model: sleep score = sleep opportunity + recovery signals
If you want a practical way to interpret any sleep score, split it in two:
Part A: sleep opportunity
- duration
- timing and consistency
- fragmentation
Part B: recovery signals
- overnight HR
- overnight HRV trend
If your score drops, ask:
- Did I give myself enough sleep opportunity?
- Did my body show signs of strain overnight?
What to do when your sleep score suddenly drops
Use this checklist in order.
Step 1: check duration first
If you slept 5 to 6 hours, the score will be low.
Fix:
- go to bed earlier for 2 to 3 nights
- if you cannot, reduce training intensity temporarily
Step 2: check the "usual suspects"
If duration was normal, look for:
- alcohol (especially within 3 to 4 hours of bed)
- late, heavy meals
- dehydration
- sleeping in a hot room
- unusual stress
- a ramp in training
Fix:
- move the last meal earlier
- keep the bedroom cooler
- hydrate earlier in the day
- consider an easier day if HRV is trending down
Step 3: check your baseline, not the single night
One bad score can be noise.
A more useful rule:
- if the 7-day average score is down and HR is up, take it seriously
- if one night is down but trends are stable, do not spiral
Step 4: sanity check the data quality
Sometimes the score drops because the device had a bad night.
Look for:
- gaps in heart rate data
- unusual spikes
- the watch being too loose
Two useful YouTube explainers (not ours)
If you want a deeper dive on how sleep tracking works and why stages are hard:
Where Century fits
Most wearables give you raw sleep data and a score. What people actually want is guidance:
- should I train hard today or go easy?
- is my sleep debt catching up?
- what habit will move my score the most?
Century is building that layer on top of Apple Health so you can get Whoop-like insights while keeping the watch you already wear.
Disclaimer
This article is educational and does not provide medical advice. If you suspect a sleep disorder or persistent fatigue, consider speaking with a clinician or a sleep specialist.