I have had many conversations with coaches struggling to use data from wearables and determining what data can be trusted and what not. Dismissing it all is too simplistic and fails to recognize how data - in combination with other tools at our disposal - can be used effectively.
Here is an important question to ask before you decide to invest any more time or energy in looking at one particular parameter, and more importantly, before you choose to act and make any adjustments based on the data:
is this specific parameter measured or estimated?
If something is estimated, be very skeptical. Regardless of how transparent is the company providing the data, an estimate typically involves larger errors. This is nothing more than an educated guess based on other parameters (possibly measurements). Examples of estimates are: readiness and recovery scores, sleep time, sleep stages, and sleep quality.
If something is measured, it’s a good starting point. We still need to establish what is the error of measurement, and make sure the sensor is accurate (e.g. through validations with respect to reference systems). However, a measurement is typically more reliable. Examples of measurements are: heart rate, HRV, and temperature.
A simple test you can do to determine when something is measured accurately or estimated is to look at the data from more than one wearable. For accurate measurements, the data will be very similar and very highly correlated between devices. For estimates, the data will be all over the place. Do not waste your time on something that is not consistent and that cannot be reproduced between multiple devices.
No company is superior in this space (they all have many smart people and plenty of money). Lack of consistency when estimating a parameter most likely highlights our inability to estimate it at all, more than the superiority of one product over another.
And indeed, when looking at recovery or readiness scores, as well as all sleep-related parameters (time, or stages), data from multiple wearables is extremely inconsistent and all over the place. This is a huge red flag.
On the contrary, looking at what is measured, for example resting heart rate and HRV, provides a very consistent picture at this point (see here).
Make sure to spend your time and energy looking at accurate measurements, as opposed to inaccurate estimates.
I hope this was informative, and thank you for reading!
Marco holds a PhD cum laude in applied machine learning, a M.Sc. cum laude in computer science engineering, and a M.Sc. cum laude in human movement sciences and high-performance coaching.
He has published more than 50 papers and patents at the intersection between physiology, health, technology, and human performance.
He is co-founder of HRV4Training, advisor at Oura, guest lecturer at VU Amsterdam, and editor for IEEE Pervasive Computing Magazine. He loves running.
Twitter: @altini_marco
I’ve seen very different estimates on sleep stages & readiness scores between Oura and my Polar Ignite. I’m back to “do I feel rested?” The measurable, like HR are consistent so it’s nice to see this topic being discussed.
Thank you.
Nice. I my personal case Oura 3 have much higher correlation with ECG than Garmin Fenix 7S.