No wearable measures HRV perfectly, and the best-known ones do not even report the same HRV measure. In an independent overnight study of 13 adults over 536 nights, measured against a chest-strap ECG ECG A recording of the heart's electrical activity, taken with electrodes on the skin. ECG-based RR intervals are the reference standard for measuring HRV. More about ECG ,, the Oura Ring agreed most closely with ECG-based RMSSD RMSSD Root mean square of successive differences: an HRV metric built from how much each heartbeat interval differs from the next. It mainly reflects vagal activity. More about RMSSD , followed by WHOOP 4.0 and the Garmin Fenix 6, while Apple Watch was not tested and reports SDNN SDNN The standard deviation of all NN intervals in a recording, in milliseconds. It measures overall variability, and its value depends strongly on how long the recording lasts. More about SDNN , a different measure (Dial and colleagues, 2025).
Evidence and device documentation checked: 13 October 2026. Makers update their software often, so details can change.
The devices at a glance
| Device | Sensor | Measure shown | When it measures | Independent overnight evidence |
|---|---|---|---|---|
| Oura Ring | Optical ( PPG PPG An optical method that shines light into the skin and detects the pulse of blood with each heartbeat. Most watches, rings and phone-camera apps use it to estimate HRV. More about PPG ), finger | RMSSD | During sleep: the mean of 5-minute samples | Dial 2025: Gen 4 and Gen 3 |
| WHOOP | Optical (PPG), wrist | RMSSD | During sleep, weighted towards deep sleep | Dial 2025: WHOOP 4.0 |
| Garmin | Wrist heart rate sensor | HRV in ms; formula not published | During sleep; HRV Status uses a 7-day average | Dial 2025: Fenix 6 |
| Apple Watch | Optical (PPG), wrist | SDNN | Separate samples recorded automatically | Not in Dial 2025 |
| Polar | Optical (PPG), wrist | RMSSD, inside the ANS Autonomic nervous system The part of the nervous system that runs the body without conscious control, including heart rate. HRV mostly reflects how it adjusts the heart from moment to moment. More about Autonomic nervous system charge score | Roughly the first 4 hours of sleep | Dial 2025: Grit X Pro |
| Fitbit | Optical (PPG), wrist | RMSSD | Longest sleep of the past 24 hours, over 3 hours | Not in Dial 2025 |
| Samsung Galaxy Watch | Optical (PPG), wrist | HRV; formula not published | During sleep and at rest | Not in Dial 2025 |
| Chest strap (reference) | ECG, chest | RMSSD and others, depending on app | When you start a reading | Used as the reference in Dial 2025 (Polar H10) |
The table describes what each maker publishes. “Not in Dial 2025” means the device was not tested in that study, not that it is inaccurate.
Oura Ring
Oura measures HRV only during sleep. According to Oura, the ring takes 5-minute samples through the night, and the average HRV shown is the mean of all those samples. Oura also shows a maximum value for the night, and it advises comparing your HRV only with your own trend, not with other people (Oura Member Care). The ring reads your pulse with light.
The Dial study compared Oura’s overnight RMSSD with a chest-strap ECG. The Gen 4 ring had the closest agreement of all devices tested, and the Gen 3 was close behind (see the accuracy table below).
Because Oura averages the whole night, its number is usually not the same as a WHOOP value from deep sleep or a 5-minute morning reading. See why Oura and Whoop show different HRV.
WHOOP
WHOOP measures RMSSD from an optical sensor on the wrist during sleep. Its calculation gives extra weight to slow-wave (deep) sleep. The exact weighting is not public, so the Dial study compared WHOOP’s value with an all-night ECG average instead (Dial and colleagues, 2025). Even so, WHOOP 4.0 agreed well with the ECG.
A value taken mostly from deep sleep can differ from an all-night average, even when both use RMSSD. That alone can make WHOOP and a ring disagree on the same night.
Garmin (HRV Status)
Garmin watches analyse wrist heart rate readings while you sleep to work out your HRV. After about three weeks of consistent sleep data, the watch shows an HRV Status. It compares your seven-day average with your own baseline Baseline Your own usual HRV level and range, built from weeks of readings taken the same way. Comparing a reading with your baseline says more than comparing it with other people. More about Baseline range (Garmin owner’s manual):
- Balanced: your seven-day average is within your baseline range.
- Unbalanced: your seven-day average is above or below your baseline range.
- Low: your seven-day average is well below your baseline range.
- Poor: your HRV values are averaging well below the normal range for your age.
- No status: not enough data for a seven-day average.
Garmin’s manual does not name the formula behind its millisecond value, so it cannot be matched exactly to RMSSD or SDNN norms. In the Dial study, the Garmin Fenix 6 agreed reasonably well with ECG RMSSD, though less closely than Oura or WHOOP. For what the labels mean in daily life, see what unbalanced HRV status means.
Apple Watch
Apple Watch stores HRV in Apple Health as SDNN: the standard deviation of the intervals between normal heartbeats, in milliseconds. Apple’s documentation says the watch records these samples automatically (Apple Developer Documentation). They are separate readings through the day and night, not one overnight average.
On Apple Watch Series 12 and Apple Watch Ultra 4, Apple also shows a Recovery HRV. Apple says it is sampled while you are still, so the number of measurements varies with how active you are (Apple Support).
SDNN and RMSSD are different measures. SDNN includes slower swings in heart rate, and it depends heavily on how long the recording lasts. So an Apple Watch value cannot be compared directly with an Oura, WHOOP or Fitbit value, or with RMSSD norms. Apple Watch was not part of the Dial study. Read why Apple Watch HRV is different.
Polar
Polar’s Nightly Recharge uses beat-to-beat data from roughly the first four hours of sleep. It combines three things into one ANS charge score: heart rate, HRV measured as RMSSD, and breathing rate. Heart rate has the biggest influence and breathing rate the smallest. The score runs from minus 10 to plus 10, where zero is your usual level over the past 28 days (Polar Support).
Because the score mixes three signals and uses only part of the night, it is not a plain HRV number. In the Dial study, the Polar Grit X Pro had the weakest agreement with ECG RMSSD of the devices tested. Polar also makes the H10 chest strap that the same study used as its ECG reference.
Fitbit
Fitbit uses RMSSD. Its HRV value comes from your longest sleep in the past 24 hours, and only sleep periods longer than 3 hours count (Google Health Help). Fitbit devices read the pulse with an optical sensor on the wrist.
Like other sleep values, it is not directly comparable with a 5-minute reading taken while awake. Fitbit was not tested in the Dial study.
Samsung Galaxy Watch
Samsung says its watches record HRV automatically during sleep and while you are resting, by analysing variations in the intervals between heartbeats (Samsung Newsroom). The watch uses Samsung’s BioActive sensor, which includes an optical heart rate sensor. Samsung does not publish which HRV formula it uses, so its number cannot be matched exactly to RMSSD or SDNN norms. Samsung was not tested in the Dial study.
How accurate are wearables for HRV?
Validation studies usually report two numbers:
- Concordance (CCC) shows how closely the device’s values match the reference, from 0 (no agreement) to 1 (perfect).
- Mean absolute percentage error (MAPE) shows how far off a typical reading is, as a percentage.
Here are the overnight RMSSD results from the Dial study, against a Polar H10 chest-strap ECG:
| Device (model tested) | Concordance (CCC) | Average error (MAPE) |
|---|---|---|
| Oura Ring Gen 4 | 0.99 | 5.96% |
| Oura Ring Gen 3 | 0.97 | 7.15% |
| WHOOP 4.0 | 0.94 | 8.17% |
| Garmin Fenix 6 | 0.87 | 10.52% |
| Polar Grit X Pro | 0.82 | 16.32% |
Overnight RMSSD agreement with ECG in 13 healthy adults (Dial and colleagues, 2025).
To put the errors in everyday terms: at an RMSSD of 50 ms, an average error of about 6% means a typical night is off by about 3 ms. An error of about 16% means about 8 ms. If your device’s typical error is close to the size of the change you are looking for, wait for a trend over several nights before drawing conclusions.
The study had limits. It was small, the people were healthy, and it tested specific models. The makers’ algorithms are private and can change with updates. Accuracy at night, when you are still, also says little about readings during movement. For a plain comparison of wrist sensors and chest straps, read is wrist HRV as accurate as a chest strap? and how HRV is measured.
Why two devices show different HRV
Wear two devices on the same night and you will usually get two numbers. There are four main reasons:
- Measure. RMSSD and SDNN give different values from the same heartbeats.
- Window. Whole night, first four hours, deep sleep only, or spot samples.
- Sensor site. Finger, wrist or chest, each with its own noise.
- Processing. Each maker filters out bad beats in its own way.
The lesson on why your devices disagree works through an example.
Scores and labels on top of HRV
Several makers turn HRV into a score or label, and each one combines different inputs:
- Garmin HRV Status compares your seven-day average with your own baseline range. Only the “Poor” label refers to normal values for your age.
- Polar ANS charge mixes heart rate, RMSSD and breathing rate, and compares them with your past 28 days.
- Oura shows nightly average and maximum HRV and recommends comparing them only with your own trend.
A score that moves does not always mean HRV moved. Check the raw millisecond value, where the app shows it, before reading much into a change in the label.
Switching or combining devices
If you change devices, treat it as a fresh start:
- Build a new baseline over a few weeks before reading much into the numbers.
- Do not mix values from two devices in one trend.
- If you wear two devices, compare the direction of change, not the values. Both should rise and fall on roughly the same days.
What to look for if HRV matters to you
If you are choosing a device, look at what it does rather than at rankings:
- When it measures. Overnight tracking takes no effort; on-demand readings give you more control.
- Which measure it shows. RMSSD lets you use more published research and norms.
- Access to your data. Some apps let you export raw beat-to-beat intervals or nightly values.
- Independent validation. Look for published studies of the model you plan to buy, not only the maker’s claims.
- Comfort and battery. Overnight tracking only works if you wear the device every night.
Measuring without a wearable
A chest strap with a free app gives ECG-quality readings for a morning check. A phone camera app can also measure HRV at rest. See can a phone camera measure HRV?