Apple Watch, readiness & HRV
HRV explained: SDNN vs RMSSD, and why your numbers disagree
Apple Health says your HRV was 38 ms. Polar’s Nightly Recharge says 62 ms for the same night, and a friend’s WHOOP shows 70. None of these devices is broken. They calculate heart rate variability with different formulas, over different stretches of time, and the numbers were never meant to be compared.
This guide explains what HRV is, what SDNN and RMSSD each measure, why your devices disagree, and how to read your own numbers so they actually tell you something.
Short answer
HRV is the variation in time between heartbeats. Apple Watch stores it as SDNN, the spread of all beat intervals, from short readings taken through the day and night. Polar, WHOOP, Oura and Garmin report RMSSD, the average change from one beat to the next, measured during sleep. Different formulas and different time windows give different numbers, and there is no reliable way to convert one into the other. Compare each device only with itself, as a trend against your own baseline.
What HRV is
Your heart doesn’t beat like a metronome. At 60 beats per minute the gaps between beats might be 0.96, 1.04 and 0.98 seconds. Breathing in shortens them slightly, breathing out lengthens them, and your nervous system adjusts them all the time. When you’re rested and relaxed, the variation is usually larger. Hard training, short sleep, alcohol, stress or feeling unwell tend to make it smaller. HRV puts a number on that variation, in milliseconds.
SDNN and RMSSD in plain words
Both start from the same raw material, the list of intervals between beats. They then ask different questions about it.
- SDNN (standard deviation of normal-to-normal intervals) asks: how spread out are all the intervals in this recording? It captures slow swings as well as fast ones, reflects both branches of the nervous system, and grows with the length of the recording.
- RMSSD (root mean square of successive differences) asks: how much does each interval differ from the one before it? It captures the fast, beat-to-beat changes driven mainly by the vagus nerve, the “rest and recover” side of the nervous system, and stays stable over short windows.
A small example shows how far apart they land. Take five intervals: 1,000, 1,040, 980, 1,020 and 960 ms. Their SDNN is about 32 ms. Their RMSSD is about 51 ms. Same heartbeats, two correct answers, and neither can be turned into the other without the raw data.
| SDNN | RMSSD | |
|---|---|---|
| Measures | Overall spread of beat intervals | Change from one beat to the next |
| Mostly reflects | Both branches of the nervous system | Vagal (“rest and recover”) activity |
| Typical use | 24-hour recordings in medicine; Apple Watch spot readings | Overnight recovery tracking |
| Used by | Apple Watch (Apple Health’s HRV field) | Polar, WHOOP, Oura, Garmin |
Why your devices disagree
The formula is only the first difference. Each device also measures at a different moment:
- Apple Watch: short readings, taken only when you’re still, by default about every four hours around the clock (more often with irregular rhythm notifications or AFib History on), plus one after every Breathe session. Apple Watch Series 12 measures HRV far more often and adds an overnight “Recovery HRV”.
- Polar Nightly Recharge: RMSSD over roughly the first four hours of sleep.
- WHOOP: RMSSD during sleep, weighted towards deep (slow-wave) sleep.
- Oura: an average of five-minute samples across the whole night.
- Garmin: an overnight average for HRV Status. Garmin’s manuals don’t name the formula; it is widely reported to be RMSSD-based.
A reading at 3 p.m. after two coffees and one at 3 a.m. in deep sleep describe two different states of your body. Add different sensors (wrist, finger, chest strap) and different ways of filtering out bad beats, and gaps of tens of milliseconds are normal. The review most often cited on HRV metrics, Shaffer and Ginsberg (2017), warns that short and 24-hour measurements “cannot substitute for one another”, even when the formula is the same.
iOS 27 gives RMSSD its own place in Apple Health
Until now Apple Health had one HRV field, and it expects SDNN. Apple’s developer documentation now lists a second type, heartRateVariabilityRMSSD, new in iOS 27 and watchOS 27, next to the existing heartRateVariabilitySDNN. That means RMSSD values can be stored as what they are, in their own history, instead of being squeezed into the SDNN field. Older Apple Watches keep recording SDNN, so for most people both histories will exist side by side.
How to read your own HRV
- Compare with yourself. HRV varies widely between people and falls with age. A friend’s number, or a chart of population averages, tells you little about yours.
- Use a rolling average. Single nights jump around, sometimes a lot, for no clear reason. A 7-day average against your normal of the past several weeks is far more stable.
- Look for context. Alcohol, late meals, short sleep, stress, hard training, travel, heat and feeling unwell all lower HRV. Often you can name the cause.
- Keep conditions steady. Same device, same wrist, measured overnight. With Apple Watch, the night’s readings are more comparable than daytime ones. A Breathe session raises HRV through slow breathing, so it isn’t a like-for-like reading.
- Read HRV with resting heart rate. HRV below your normal for several days while resting heart rate sits above it is a clearer signal than either one alone.
None of this is a diagnosis. HRV is a useful personal signal for training and recovery, and a poor tool for judging your health from one number.
How Keho handles SDNN and RMSSD
Keho keeps the two measures apart everywhere. It never writes Polar’s RMSSD into Apple Health’s SDNN field, because that would mix two different measures into one history. Polar HRV from Nightly Recharge stays RMSSD, stored on your iPhone with its own trend. Apple Watch HRV stays SDNN. On iOS 27, RMSSD from Apple Health is a third, separate series.
For the morning readiness score, Keho picks one HRV series, the one with a reading from last night and the longest history, and compares it only with its own baseline from the last 49 nights. It never blends measures or switches measure inside a baseline. You then see your HRV as Favorable, Typical, Slightly off or Notably off against your own normal, next to resting heart rate, breathing and temperature. If you use a Polar watch, here is what Polar Flow does and doesn’t send to Apple Health.
Common questions
Is SDNN or RMSSD better?
Neither in general. RMSSD measured during sleep is the common choice for daily recovery tracking. SDNN over 24 hours is the standard in medical long-term recordings. For you, sticking with one measure on one device matters more.
Why is my Apple Watch HRV lower than my WHOOP or Polar HRV?
Apple Watch reports SDNN from short spot readings; WHOOP and Polar report RMSSD from sleep. Different formulas and moments give different values. A gap between them says nothing about which is right.
Can I convert SDNN to RMSSD?
No. Both come from beat intervals, but the relationship between them differs from person to person and from night to night. Any conversion factor would be a guess.
What is a good HRV?
One that is normal for you. Age, sex, fitness, genetics and the measurement method all shift the number. Watch your own trend rather than a target value.
Sources: Shaffer F, Ginsberg JP, “An Overview of Heart Rate Variability Metrics and Norms”, Frontiers in Public Health 5:258 (2017); Apple, “Using Apple Watch to measure heart rate, calorimetry, and activity” (November 2024); Apple developer documentation, “heartRateVariabilityRMSSD”; Polar support, “Nightly Recharge”. All checked on 28 September 2026. Personal guidance, not medical advice. Keho is an independent app and not affiliated with Apple or Polar.