Field note
A watch should compare you to your own recent weeks because people differ so much that group averages often fail as a personal rule.
There is no single chart of normal people hiding behind a good overnight score. Fisher, Medaglia, and Jeronimus (2018) argued that lack of group-to-individual generalizability is a real threat to human subjects research. Patterns that hold for averages can miss what happens to one person. Wearable scores that ignore that gap will misread you.
Quer and colleagues (2020) studied nearly a hundred thousand watch wearers and mapped inter- and intraindividual variability in daily resting heart rate. Two perfectly healthy people could sit dozens of beats apart. Age, sex, sleep, BMI, and season all mattered. The hopeful flip side of that same data: each person's own pattern is steadier and more readable than the spread between people.
A reading that would be ordinary for your neighbor can be a real event for you. Only your own recent weeks can tell the difference. Altini and Plews (2021) analyzed large free-living resting heart rate and HRV series and tied within-person changes to training, alcohol, sickness, and sleep in ways population cutoffs never see.
| Comparison | What it answers |
|---|---|
| You vs your last few weeks | Is today unusual for you? |
| You vs population average | Where you sit among strangers (often noisy for daily calls) |
| You vs age/sex ranges | Whether a long-term level looks healthy at all |
What a recovery morning is doing when it uses that within-person band is in What does recovery actually measure? Why a run of high mornings beats one spike is in Why is my resting heart rate up?
A baseline sounds fancy. It is mostly this: take your recent weeks of readings and find the middle. Using the middle value rather than the average does real work. An average gets yanked around by one wild night. The middle barely moves. Leys and colleagues (2013) showed why absolute deviation around the median beats standard deviation around the mean for spotting outliers without letting one extreme redefine normal.
Fast-moving signals like overnight heart numbers are often read against a couple of recent weeks, because your body genuinely changes that fast. Slow-moving ones like fitness look back over a month or more. Each day should be judged only by the days that came before it. Peeking ahead invents a future you have not lived.
Plews and colleagues (2013, 2014) showed in endurance athletes that HRV monitoring needs enough compliant mornings to be valid, and that adaptation shows up in the relative pattern, not in a universal cutoff. A few good readings a week can carry a trend. Missed nights do not have to break the picture.
What HRV itself is, as a relative overnight read, is in What is HRV?
A baseline that moves with you can normalize a slow drift. If a number worsens gently for months, each day looks ordinary against an already-shifted normal. That is one reason some longevity views also look outward at published ranges for age and sex. And a rough patch leaves a mark. Weeks of illness or heavy living teach the baseline a worse normal for a while, so the first good nights afterward can read better than they should. It relearns, but it takes the same few weeks it always takes.
Early days should stay cautious. When an app knows nothing about you yet, fake confidence is worse than saying it is still learning.
Your own usual is the band that makes daily signals useful. Population charts answer a different question.
I write these because I build Aera, an iPhone app that reads Apple Health against your own baseline and puts the research and the error bars underneath every number it shows you, including this one. You do not need it to use anything above.