Epix 2 vs Sleep Study: SpO2 off, wake & deep undercount

For being a watch it did great in my opinion.

-Total duration is matching pretty much perfectly and light sleep is pretty close.

-It detected very well the longest awake periods(this is something subjective but i prefer to know when i really woke up instead of knowing that each night it's common to have 1 hour of awake time that is composed of super short periods that i will never remember about). 

-it detected well more than half of rem stages and the same for the biggest deep part

Didn't expect an accurate average Spo2 like that.

I would really like to see more nights if you have any chance. Thanks 

Given this WatchPAT is not a medical device (there are different types of FDA approval and this is for home use not medical/clinical) it provided results that were not overly different from the Epix. A 2020 study (https://pubmed.ncbi.nlm.nih.gov/31965862/) found the "WatchPAT slightly overestimated total and REM sleep time, and sleep efficiency." 

However, from this one sample compared to Epix one could assume that the Epix overestimated REM sleep. The total duration of sleep recorded by both is virtually identical. As usual, the different stages of sleep are identified differently as is often the case when doing comparative studies.

Watches and devices like this WatchPAT use algorithms to determine sleep phases making it very hard to determine which is better. Too difficult to draw any firm conclusions about which of these produces the better outcomes from just this one sample. 

Thankyou for sharing

How many data points do you have and on how many data points per average participants will the Studie be done?

only one night seems a bit odd. 

Forgive me, I need to understand.
I have had experience with some sports watches, 2 huawei, 2 Polar and 1 Garmin Venu.
Why do the huawei (from 200 €) correctly recognize sleep phases and naps in a precise way, while Garmin (much more expensive) has to be set to an "approximate" time and is not able to recognize naps?
How can Garmin look reliable in the Bodybattery, with values that seem random?
thank you

I understand why we need to set an approximate time, it is using that also for the DND mode/ sleep mode. I think that is useful.

If you used Fenix 3 or older, there you needed to start the sleeping mode manually. It was also able to track naps because you just started it for the nap like you did for the night sleep. That was changed in the F3 someday and for people working in shifts its a PIA since then. After a while, they even brought back the manual mode to the F3 because of the uproar.

When you take naps, body battery will recharge because heart rate and stress levels goes down.

The only thing that doesn't change with naps is the sleep score. 

as others have stated it does recognize naps, but correctly not as sleep - it recognizes them as part of your body battery i.e. a recharge.

As part of sleep tracking your percentage estimate of overall duration for Deep and REM sleep is quite important. Adding naps to your overall duration will impact/lower the percentage score for deep and REM due to how sleep cycles traditionally work (especially as naps are normally considered to be under 40 mins and thus very likely only light sleep - and thus you would minimize the percentage duration of deep and REM to overall sleep if additional light sleep was incorrectly added)..

So based on this I am very happy that naps do not impact my sleep stages percentage of duration, and am quite happy to see if a nap was beneficial (i.e. did it recharge my body battery or not) via checking body battery.

I understand but, apart from naps, how can it be reliable if I was in the car at 7.30 this morning and for the Garmin I was still in REM sleep because it was set to 00:00 - 08:00?

Sleep detection doesn't start/end based on the sleep schedule.
Sleep schedule is used more for activation/deactivation of the DND mode and for the "night" watchface/brightness level.

If you have issues with sleep detection try a soft reset(hold top left button until the watch turn off and then turn it on) or wait few more days(if the watch is new) and see if the watch get used to your behaviors.

Is the watch updated to the latest version(7.35)?

For example, my sleep schedule is 10pm-6:30am but this was the sleep detection of the last night