What Is Your Health Score Actually Compared To?

Published 12 September 2026 · 10 min read

Two of these four wearable makers tell you in their own documentation. Two do not.

Your ring or watch hands you a number every morning. Readiness 71. Recovery 44. Body Battery 38. Before that number can mean anything, you need one more fact: what it was compared against. A score is a comparison, and a comparison needs something on the other side.

There are only really two candidates. Your own history, or other people. Those produce different verdicts from the same night, and they call for different responses from you.

So this guide does something narrow and checkable. It goes to each vendor's own published documentation and asks a single question: do they say? Every quote below was read on the vendor's own page, and every page is linked. One of them, WHOOP's, sits behind a WHOOP account login, and that is flagged where it appears. Where a vendor does not say, that is recorded as not saying rather than guessed at.

The short answer, in a table

The two that disclose both turn out to be mixed, part personal and part population. That is the more interesting finding, and it is the one nobody tells you at setup.

WHOOP says it plainly, and says it changes

WHOOP is the clearest of the four. Their support article on the Recovery calibration period (support.whoop.com, which requires a WHOOP account to open) describes the reference switching over as the app learns you:

"The first time you record a Sleep on WHOOP, your data is compared against typical Recovery metrics (like Heart Rate Variability (HRV) and Resting Heart Rate (RHR)) based on your age, gender, and fitness level. As time goes on and more of your personal data is stored/measured, your Recovery 'calibrates', and your personalized baseline is used to generate your Recoveries. After 4 consecutive Recoveries, your Recovery is considered to be calibrated, or measured to your own physiology."

Read that again, because it answers a question people ask constantly. For your first few days the score is not about you. It is about people of your age, gender and fitness level. WHOOP tells you this, and it is the honest thing to publish.

WHOOP also defines the calibration period as the "time period where WHOOP learns about you (the user) and sets an educated baseline that's personalized to you."

Oura says it too, and names where the population creeps back in

Oura's Readiness Contributors page (support.ouraring.com, last updated 14 July 2026) leads with the personal reference:

"Readiness contributors are based on your personal averages. It can take up to two weeks for the Oura App to learn your average values for each contributor listed below. Once average ranges are established, your daily readings are evaluated against them."

Two weeks, and then it is your averages. Clear enough. But the same page then describes individual contributors that reach outside you:

"The Sleep Balance contributor assesses the amount of sleep and possible sleep debt over the past two weeks compared to your baseline and general recommendations for people of your age."

"Previous Day Activity measures all of your activities and sedentary (inactive) time from the day before, and compares them with your long-term averages and recommended daily amounts of physical movement."

So Sleep Balance is measured against your baseline and against a recommendation for your age group. Previous Day Activity is measured against your long-term averages and against a recommended amount. The headline is personal. Two of the contributors underneath it are hybrids.

Oura's Readiness Score page (support.ouraring.com) adds the window: "Readiness Contributors with 'balance' in their name, such as Activity Balance, HRV Balance, and Sleep Balance, use 14-day weighted averages." The same sentence goes on to say those averages are compared to your long-term average over the past two months, so that particular comparison is personal on both sides.

That last detail matters more than it looks, and it is the subject of how long until my baseline is accurate.

Garmin does not say, and the interesting part is that they say it elsewhere

Garmin's owner's manual for Body Battery (garmin.com) is specific about inputs and bands:

"Your watch analyzes your heart rate variability, stress level, sleep quality, and activity data to determine your overall Body Battery level. Like a gas gauge on a car, it indicates your amount of available reserve energy. The Body Battery level range is from 5 to 100."

Four named inputs, four named bands. Nothing about what any of it is compared against. The same is true of Garmin's all-day stress score (garmin.com):

"Your watch analyzes your heart rate variability while you are inactive to determine your overall stress. The stress level range is from 0 to 100."

Again, inputs and bands, no reference.

Now the part that makes this an omission rather than an oversight. On HRV Status, in the same vivoactive 5 manual that documents Body Battery, Garmin does state a reference, and states two different ones:

Balanced: "Your seven-day average HRV is within your baseline range."

Poor: "Your HRV values are averaging well below the normal range for your age."

Your baseline for one status, the normal range for your age for another. So Garmin discloses references when they choose to. For Body Battery and the stress score, the two numbers a Garmin owner looks at most, they do not.

This is not a hypothetical gap. On Garmin's own vivoactive 6 forum in August 2026, an owner posted four weeks of data showing a Body Battery that never rose above 50 while HRV Status stayed Balanced, and asked directly whether the Body Battery scale is normalised against population values rather than a personal baseline (forums.garmin.com). In his words: "I'm trying to understand the algorithm's normalization model."

Three weeks later the thread has one reply, and it is from the same person adding a further question. Nobody answered. That is the state of the public record on what a Body Battery is compared to.

Apple has just entered the same category, with the same silence

Apple announced Readiness on 9 September 2026, for Apple Watch Series 12 (apple.com/newsroom) and for Ultra 4 in a same-day release:

"Apple Watch Series 12 introduces readiness, a new feature that analyzes a user's recent activity, training load, vitals, and sleep score to offer insight into their capacity for the day ahead. Readiness provides a single 0-10 score with a clear, actionable recommendation: Recover, Pace Yourself, Ready, or Go For It."

Four inputs, a 0 to 10 scale, four recommendations. No statement of what any of it is compared against. Apple says the algorithm "was developed using data from the Apple Heart and Movement Study", which describes how the model was built, not what your individual score is weighed against day to day.

Apple does use personal-reference language elsewhere in the same release, for a neighbouring feature: "Overnight vitals now includes recovery HRV measurements analyzed against a user's personal baseline." That sentence is about Overnight Vitals. It is not about Readiness, and reading it across would be exactly the kind of inference this guide is trying to avoid.

Apple's watches reach stores on 18 September 2026 by Apple's own schedule, so a large number of people are about to hold a second readiness-style score alongside one they already own. When those two disagree, the reference is the first thing worth asking about, and for both of them it is currently undocumented.

Why the reference decides the verdict

Take one ordinary night and score it four ways.

Nothing about the night changed. Only the question did. This is the mechanism behind a complaint that turns up in every wearable community: the score disagrees with how you feel and you conclude the device is broken. Often the device is measuring competently and answering a question about somebody else. That is the same structural problem covered in why Oura and WHOOP give different scores and in why health apps disagree.

It also explains a quieter failure. A score referenced to a short recent window will drift with you. Have a genuinely rough month and the window comes down too, so readings that should stand out start reading as ordinary. Nothing gets flagged, because by then there is no outlier left to flag.

What you can do with this today

You cannot make a vendor document its reference. You can stop reading an undocumented score as though it were a measurement of you.

1. Find out whether yours discloses. For Oura the answer is in the pages linked above and opens for anyone. WHOOP's is behind a WHOOP account login. For Garmin Body Battery and Apple Readiness there is currently no answer to find, and knowing that is worth more than assuming either way.
2. Discount the first two weeks. WHOOP says four recoveries. Oura says up to two weeks. During that window your score is at least partly about other people, by the vendors' own account.
3. Ask "different for me" rather than "good". A change against your own record survives an unknown reference. A grade does not, because a grade needs a class and you do not know whose class you were put in. This is what a personal health baseline is for, and it is the same reasoning behind what a normal range on a blood test means.
4. Read a flat result as information. If you changed something and your own numbers did not move, that is a finding about you rather than a missing answer. The same logic runs through how to tell if a supplement is working.

What this guide does not establish

Stating the limits, because the whole point was to separate what is documented from what is assumed.

• Not saying is not the same as being population-based. Garmin and Apple may well reference your own history for these scores. Their published documentation does not state it, and that is the entirety of the claim here.
• This is documentation, not behaviour. Nobody tested these scores. A vendor could document one thing and compute another, and this guide would not detect it.
• Documentation moves. Every page here was read on 12 September 2026 and every one is linked so you can check it yourself. Apple's Readiness had no support page at the time of writing, which is consistent with the feature not having shipped.
• The Garmin forum thread is a user's unanswered question, not evidence of a mechanism. The poster proposes population normalisation as a hypothesis and asks for confirmation. He did not get one. It is cited as an open question, which is what it is.
• Four vendors is not the market. Fitbit, Samsung, Polar, Ultrahuman and others are not covered.
• None of this is medical advice. These are consumer wellness scores, not clinical measurements, and no score here diagnoses anything.

The short version

A score is a comparison, and the comparison is the part that decides the verdict. WHOOP and Oura publish theirs, and both turn out to be mixed: personal baselines with age-based recommendations underneath. Garmin publishes the inputs to Body Battery and not the reference, while publishing a reference for HRV Status in the same manual. Apple has just launched a readiness score and published the inputs and not the reference.

Until a vendor tells you, the safest reading of any score is the weakest one: it went up, or it went down, against something you cannot see.

That is the argument NUVARD is built on. One model of one person, reading every signal against your own record rather than against a range you were never in, and reporting no measurable signal as a result when something you tried did nothing for you. NUVARD is coming soon. Join the waitlist at nuvard.ai

Sources

All pages read 12 September 2026.

• WHOOP. What is the Recovery calibration period? Requires a WHOOP account to open. support.whoop.com
• Oura. Readiness Contributors, last updated 14 July 2026. support.ouraring.com
• Oura. Readiness Score, last updated 14 July 2026. support.ouraring.com
• Garmin. Body Battery, vivoactive 5 Owner's Manual. garmin.com
• Garmin. Heart Rate Variability and Stress Level, vivomove Trend Owner's Manual. garmin.com
• Garmin. HRV Status, vivoactive 5 Owner's Manual. garmin.com
• Garmin Forums, vivoactive 6 Series, August 2026. Owner question on Body Battery normalisation, unanswered. forums.garmin.com
• Apple. Introducing Apple Watch Series 12, with the all-new Health Sensing System, 9 September 2026. apple.com/newsroom