Apple Watch Nails Resting Heart Rate and Misses on HRV

Research comparing Apple Watch readings to lab-grade reference devices found that its resting heart rate numbers closely tracked validated equipment, while its HRV readings did not hold up the same way. One trial found the watch underestimated HRV by a meaningful margin even as its heart rate numbers matched a chest strap almost exactly, and statistically ruled out the heart rate gap as the explanation for the HRV gap. The two metrics on the same device, measured at the same time, do not share one accuracy story.

What people expect when a watch gets their heart rate right

If your Apple Watch lines up with a manual pulse check or a second device, it's tempting to assume that the HRV number sitting right next to it on the same screen is built on just as solid ground, since the watch is clearly counting beats correctly. That leap from one metric to the other is exactly what researchers have gone and tested directly, rather than assumed.

The answer they found does not confirm the leap, not quite.

Where the accuracy diverges: RHR versus HRV

Despite using different watch models, these two trials tell a consistent story. Resting heart rate error came in small, in the range of a few beats per minute or a low single-digit percentage. HRV error was an order of magnitude larger in relative terms, with one trial reporting a mean absolute percentage error near 29 percent for HRV against a chest strap paired with Kubios HRV software, compared to under 6 percent for resting heart rate on the same wrists. That's a big gap.

Part of why that gap exists comes down to how the watch actually detects a heartbeat in the first place. Counting pulses reliably over a stretch of time is basically all resting heart rate requires, and optical sensors handle that part well. HRV needs the exact timing between individual beats, down to the millisecond. A much harder ask for a wrist-based optical sensor, a distinction covered in more depth in how wearables measure HRV using ECG versus PPG.

None of this tells you whether a given HRV reading is high or low for a person. It only speaks to how closely the number on the watch matches what a reference device recorded at the same moment, which is a separate question from what a good HRV number even means.

Why one number is more forgiving than the other

A broader review of short-term HRV reliability found that coefficients of variation for HRV measures ranged enormously, from under 1 percent to over 100 percent, depending on the measurement conditions. Reliability was consistently worse when measurements were taken during some kind of intervention or challenge to the body rather than at quiet rest, and worse in clinical populations than in healthy volunteers. That pattern of instability under changing conditions is separate from device accuracy, but it stacks on top of it. It's part of why a single overnight number can move around more than people expect, something explored further in why HRV can drop sharply overnight.

One systematic review looked specifically at wearable devices measuring HRV against ECG and PPG technology, and described the overall evidence on wearable HRV accuracy as still inconclusive, rather than settled in either direction. That framing lines up with what the Apple Watch specific trials found: heart rate counting is a comparatively solved problem for these devices, HRV extraction is not.

Across brands, wrist optical heart rate holds up at rest and asleep. A meta-analysis of 44 studies and 15 brands found wrist readings within about half a beat per minute of an ECG or chest strap at rest, during sleep and on a treadmill, while they read 7.26 beats per minute low during resistance training and 4.55 low during cycling, with the gap growing as heart rate rose (Journal of Sports Sciences, 2020).

What this evidence doesn't establish

Both of the Apple Watch validation trials recruited healthy adults for relatively short windows, one lab session in one case, 14 days in the other. Neither tested people with diagnosed cardiovascular or autonomic conditions. And the broader reliability review found that clinical populations show worse HRV reliability than healthy volunteers, so there's no basis here for saying how these accuracy figures would look in someone managing a heart condition.

Common questions

KnowYourPrime Evidence Registry: Finding KYP-ER-0003: Wearable HRV Accuracy Against ECG · Finding KYP-ER-0009: Wearable Heart Rate Accuracy · Finding KYP-ER-0025: Apple Watch HRV Accuracy Against ECG and Chest Straps · Finding KYP-ER-0004: Chest Strap HRV Accuracy Against ECG