How to read your HRV score: what Garmin, WHOOP, and Elite HRV actually show
Three devices, three numbers, none of them comparable. What each screen actually measures, why the numbers disagree, and how to build a personal baseline that means something.
Same morning, two devices, two numbers that would not agree. My Garmin told me my overnight variability had settled into its usual band, somewhere between 77 and 94 ms. Then I sat up, ran my Elite HRV reading lying in bed, and got 65. Lower. One screen looked like a normal recovery, the other looked like I was down, and for a while that gap genuinely confused me.
Neither number was wrong. I had been reading them as if they answered the same question, and they do not. The Garmin figure is a report on the night that just finished: how well my system recovered while I slept. The Elite HRV reading is a snapshot of the present moment: whether that recovery has actually left me ready this morning. Same metric, same milliseconds, two different questions, because they are measured at two different moments.
That is the whole reason your devices disagree, and it is the first thing to get straight before you trust any of them.
If you have read what your HRV actually measures and what moves your variability score, you know what the number tracks and which inputs drive it. If you have read Garmin vs WHOOP, you know which device to buy.
This article answers the question that comes next: you have opened the app, there is a number on the screen, and you need to know what it is actually telling you. Because the honest starting point is that the number is more confusing than it looks.
Here is the trap that catches everyone with more than one device. Garmin shows you a figure in milliseconds. WHOOP shows you a recovery percentage. Elite HRV shows you a score out of 100. Three different numbers for what feels like the same thing, and none of them line up. The instinct is to decide one device is wrong. That instinct is the mistake.
None of them is wrong. They are measuring on different windows and reporting on different scales. Once you understand that, the numbers stop competing and start being useful. This article is how to read your screen, and then how to build the one thing that makes any of these numbers mean something: your own baseline.
At a glance: what the three main devices actually show
The single most useful thing to understand up front is that these are not interchangeable numbers. Read this table before you ever compare two screens.
| Garmin (HRV Status) | WHOOP (Recovery) | Elite HRV (Morning Readiness) | |
|---|---|---|---|
| What it measures | Your variability score (RMSSD) | A composite: variability score, resting heart rate, respiratory rate, sleep | Your variability score (RMSSD), shown as a score |
| When it measures | Averaged across the night, during sleep | During your deepest sleep, overnight | A single reading you take on waking |
| What the number is | Raw milliseconds (e.g. 67 ms) | A 0 to 100% recovery figure | A 0 to 100 score built from ln(RMSSD) |
| What it is really answering | How variable was my heart overnight? | How ready does the whole system look? | How variable am I in this one morning reading? |
| Comparable to the others? | No | No | No |
The right-hand row is the one to internalise. A “67” on Garmin, a “67%” on WHOOP, and a “67” on Elite HRV are three unrelated quantities that happen to share a digit. Comparing them tells you nothing. (This table describes what each platform displays; for which device suits your training, see the Garmin vs WHOOP breakdown.)
What does Garmin HRV Status actually show?
Garmin shows you the raw variability number in milliseconds, averaged across your night’s sleep. It records your variability score (RMSSD, the beat-to-beat variation that reflects your recovery system) through the watch’s sensor while you sleep, then compares last night against your personal reference. The status words, Balanced or Unbalanced, are just that comparison in plain language.
The reference is built slowly, and that is by design. Per Garmin’s own documentation, the watch needs roughly 19 nights of sleep data before it will show a HRV Status at all, and it reads each night against a seven-day rolling average sitting on top of a three-week baseline.
So the word on the screen is relative, not absolute. “Balanced” does not mean your number is high. It means last night sat inside your normal range. The same 60 ms can read as Balanced for one athlete and Unbalanced for another. The status is always about you versus you, never you versus a chart.
What does WHOOP recovery actually show?
WHOOP does not show you your variability score as a number. It folds it into a recovery percentage. Per WHOOP’s published methodology, the 0 to 100% figure is a proprietary blend of your variability score (measured during your deepest sleep), resting heart rate, respiratory rate, and sleep quality. You see the output and the inputs, but not the weighting.
This is the key difference from Garmin. A Garmin number is your variability, full stop. A WHOOP recovery score is a judgement that already has variability, heart rate, and sleep baked into it. That makes it easier to read at a glance and harder to interrogate.
The practical consequence: a low WHOOP recovery does not tell you why on its own. Two mornings at 40% can have completely different drivers, one a suppressed variability score, the other a poor night’s sleep with variability intact. To read a WHOOP number properly you have to open the underlying metrics. The headline percentage is a starting question, not the answer.
What does Elite HRV show, and why is it a different number again?
Elite HRV takes a single reading in the morning and turns your variability score into a 0 to 100 score. You measure for a minute or two on waking, and the app applies a natural log to your variability number (RMSSD) before scaling it. Per Elite HRV’s knowledge base, the raw log value “typically ranges from 0 to 6.5,” which is why they expand it into a friendlier 0 to 100 range.
Two things make this number incompatible with Garmin and WHOOP. First, the window: this is one spot reading after you wake, not an average across the whole night. Second, the maths: a log transform compresses big numbers and stretches small ones, so the score does not move in a straight line with the milliseconds.
Why the log transform at all? Because your variability score is lopsided. A handful of very high mornings would otherwise drag any simple average around. Taking the natural log tames those outliers so the trend behaves, which is exactly why the research standard uses the log of the number, not the raw millisecond value, to track training (Plews et al., 2013). The score you see is that statistical smoothing, made readable.
Why you cannot compare the number between two devices
The reason three devices disagree is almost never the sensor. It is the window and the scale. This is the point most people get backwards, so it is worth proving.
On raw signal quality, the hardware is not the problem. Plews et al. (2017) put wrist optical, a chest strap, and a clinical ECG (electrocardiogram, the hospital-grade gold standard for heart signals) head to head on the same people. According to Plews et al. (2017, Int J Sports Physiol Perform), the wrist and strap measures had “almost perfect correlations with ECG (R = 1.00).” In plain terms: for the underlying variability number, a good wrist reading and a hospital-grade trace land in the same place.
So if the sensors agree, why do the screens disagree? Two reasons, both already on this page. One, the sampling window: Garmin and WHOOP average across the night, Elite HRV takes a morning spot reading, and your variability is genuinely different at 3am deep sleep than at 6:30am awake. Two, the scale: milliseconds, a recovery percentage, and a log-derived score are three different rulers.
My own two devices show the window effect directly. My Garmin overnight baseline sits between 77 and 94 ms; my Elite HRV reading, taken lying in bed on waking, comes in around 65 ms. Same metric, same millisecond scale, and the morning number is simply lower, because a supine spot check the instant you wake is not the same thing as an average across a full night of sleep. One tells me how the night went; the other tells me where I am right now.
Never compare a number from one device against a number from another. Pick one system, and only ever read today's figure against your own history inside that same system. Cross-device comparison is not a tie-breaker, it is a category error. The number is only ever meaningful relative to your own baseline on the same measurement.
How to build a baseline that actually means something
A device number is useless until it has something to be compared against. That something is your baseline, and you have to build it deliberately. Here is the three-step version that works on any of the three platforms.
The research is blunt about which unit carries the information. Plews et al. (2013) showed in elite endurance athletes that the weekly trend, not the daily reading, is what tracks training adaptation. Day-to-day swings of 10 to 20 ms are normal and driven by dozens of inputs unrelated to readiness. Flatt and Esco (2015) confirmed that a rolling average “provides a more accurate reflection of overall training adaptation” than any single number.
So the baseline you are building is a moving average, not a personal record. The goal is not to chase your highest reading. It is to establish the band your normal mornings live in, so that when a number falls outside it, you know the move is real rather than noise.
One more practical note on the scale you are averaging. If you use a raw-milliseconds device, your daily numbers will look jumpier than a log-based score, because the log transform in apps like Elite HRV is doing that smoothing for you (Plews et al., 2013). Neither is more correct. Just know which one you are reading, and do not be alarmed by the raw number’s larger swings.
What to do once you have your baseline
Building the baseline is the setup. The payoff is what you do with a reading once you have a reference to judge it against. That is a separate skill, and it has its own article.
The short version: a number below your baseline is information, not permission to skip. It tells you what kind of work your body can absorb today, not whether to train. Turning a morning figure into a specific session decision, push, adapt, or start easy and reassess, is covered step by step in how to turn your HRV into a training decision.
Read together, the two articles are the full loop. This one gets you a baseline you can trust. That one turns it into a call you can act on, without ever using the number as an excuse to do less.
Want your own data read this way?
The framework here gets you a clean baseline. Applying it to your device, your sessions, and your training block is a different step. If you want someone to look at your actual numbers and tell you what they mean for how you should train this week, that is exactly what the HRV follow-up service does.
A few places available. Get in touch.
Why is my HRV number different on Garmin, WHOOP, and Elite HRV?
Because they are not the same measurement. Garmin and WHOOP average your variability across a full night of sleep; Elite HRV takes a single reading in the morning after you wake up. On top of that, each shows it on a different scale: raw milliseconds, a recovery percentage, and a 0 to 100 score from the log of the number. The sensors agree closely; the windows and scales do not. Pick one system and stay in it.
Which device gives the most accurate HRV reading?
For the underlying variability number, the sensor is not where accuracy is won or lost. Plews et al. (2017) found wrist and chest-strap readings correlated almost perfectly with a clinical ECG (R = 1.00). What differs between platforms is the window they sample and the algorithm they wrap around it, not the raw signal. Accuracy for you comes from measuring under identical conditions every day, not from owning a specific brand. Consistency beats hardware.
How long does it take to build a reliable HRV baseline?
Plan on three to four weeks of daily readings under identical conditions. Garmin needs roughly nineteen nights before it shows a HRV Status at all. Below three weeks, one bad week of travel, illness, or a peak block can drag your reference far enough to mislead you. Your first two weeks are directional, not definitive. By week four the rolling average settles into something you can read against. Start now; the data only sharpens with time.
Should I look at the single reading or the trend?
The trend, every time. A single morning reading swings 10 to 20 ms under stable conditions, driven by inputs that have nothing to do with readiness. The signal lives in the rolling average, not the daily dot. Track the seven-day average and how it moves week to week. One reading outside your normal range is noise; a multi-day move in the same direction is the first thing worth reading as information.