Your sleep score changed: what wearable ratings actually measure

A lower sleep score can reflect a change in the scoring system, rather than a sudden deterioration in sleep. Google’s current help page explicitly describes this possibility after an algorithm update. It does not follow that every change has that explanation. Google’s explanation of its new score

A sleep score is a summary built from selected inputs. Understanding those inputs is more useful than treating every wearable’s number as the same measurement.

What sits underneath a sleep rating

There are several layers between a sensor and the number on an app’s home screen. Apple describes using the watch’s accelerometer, which detects small movements, to estimate sleep duration, sleep stages and wakefulness. Its score then combines selected sleep information into a separate rating. Apple’s sensor explanation, Apple’s score documentation

That separation matters: a study or technical description of stage detection answers a different question from how a manufacturer decides to reward duration or consistency. Evidence for one layer should not automatically be treated as evidence for every label above it.

What Apple and Google put into the score

These are two documented examples, selected to explain different approaches. They are not a survey or ranking of all sleep trackers.

Sources: Apple Sleep Score, Google Health Sleep Score.

Apple’s rating runs from 0 to 100. In its published breakdown, the interruption component considers both the duration and quality of awake periods. The headline number therefore represents more than hours spent asleep. Apple’s current explanation

The shared word “score” does not make the systems interchangeable. Comparing two brands’ ratings as though they were the same unit would obscure their different definitions.

An update can change the baseline

Google says its new algorithm changes how sleep is characterised and may lower a user’s score during the transition. It also says recent historical scores and stages may change. Google’s claims of improved accuracy are manufacturer claims; the help page is not an independent comparison study. Updated-score explanation

This creates a practical reading question for a trend chart: did the apparent change occur under one scoring method, or across an update? A discontinuity between algorithms is different from variation within the same system.

Missing or edited data can change the picture

Apple says a watch that runs out of battery cannot record the night’s sleep, and a loose fit can affect movement recording. Its Health app provides duration and stage histories alongside the rating. Apple’s recording and history guidance

Google says manually editing a sleep log can recalculate the score; extended entries may contain gaps or fail to generate a fully accurate score. A revised number can therefore involve a revised record, rather than new sensor observations. Google’s sleep-log explanation

A score is not a diagnosis

In its 2018 statement, the American Academy of Sleep Medicine distinguished consumer sleep information from validated diagnostic testing. It also recognised that such information could help discussions between patients and clinicians during an appropriate evaluation. This is a dated principle, not a claim that every present-day wearable lacks a regulated health feature. AASM’s statement summary

The useful contribution of a score is its summary; the cost is that the underlying detail is compressed. The device, data quality and algorithm version help explain what changed. The number alone does not identify a sleep disorder or establish that one tracker is clinically better than another.