PNN50
Rick Lamb
Considering the impact of artifacts (e.g., Apple Watch measurements artifacts due to movement or poor contact) on SDNN and RMSSD, I suggest you add the pNN50 calculation with the ability to track over time. With the short time length of the Apple Watch samples, the pNN50 may be a more reliable measure of variability.
Byron
Just for context: we do run a cleaning algorithm over Apple's beat-to-beat data to account for artifacts due to movement, and based on our testing, Apple appears to do something similar when calculating HRV using SDNN.
That said, it shouldn't be much of a lift for us to give you the option to select pNN50 if preferred. Will add it to the roadmap.