Garmin - Chris said:

sleeping vs. napping

That's not the point. The point is weather or not the last night gets overwritten. Neither a real nap, nor misinterpretation of any other state should override the "main" sleep. Ask the user what to do or categorize it as a separate nap (that gets added). Anything, but DON'T OVERRIDE the last night. How hard is this??

Z said:

Nor completely agree with the suggestions on the link

So what is your suggestion then and what parts do you not agree with? In the perfect world the sleep algorithm will work perfectly and we might have naps that contributed to the sleep not replacing it; even though I do not take naps unless sick. But improving the algorithm as Chris said will take time and we may also end up with an algorithm that still does not work for all when in place some day far far away in the future. I just want to be able to wear my watch on daily basis again and I do not think it is too much to ask for to give the user some more control when the sleep detection kicks in. At least then I can tell the watch myself when start to listen for sleep in case it fails badly as it does today.

The big problem here at least for me is that naps that are not naps replace my nights sleep and then mess up with daily suggestions, training readiness, and recovery times. Thus if I want to use these features I can’t wear watch during daytime due to regular fake sleeps. That is not what I expect from the watch and I upgraded from fenix 6 only for the race widget daily suggestions and training readiness. If those does not work due to destroyed by fake sleep then I could just had kept my old watch

I said not completely, which means I partly agree :) Anyway, I repeated my suggestion(s) so many times. For example, user Michael above your post explained and my post on the top of this topic, just below Chris' one.

To quote my self: "Moderator, users are repeatedly requesting to not replace prior night's sleep, regardless how long nap is."

Federico G. said:

In my opinion a good idea would be to improve software to register more "sessions" per day, one called "primary sleep" which is the nearest to the sleep window

For most people probably a good idea but not without challenges. Some people (I am not one of those) work irregular times and for this to work for them they have to change the sleep window every time their work schedule changes. Also if a nap is too long it does not contribute to your sleep quality, instead it worsen.

This is a tricky one as Chris mentioned and that is why I suggested that maybe the user can get some more control of when the sleep detection is active by some selectable option in the sleep menu under system; the options may be e.g., as today, only during sleep window, or when manually activated via sleep widget

Another option is to compare the sleep score for every sleep and only save the best sleep every day. Thus a worse sleep will not overwrite a better sleep; however that will change the current behavior of how it works today and maybe need some more investigating on the impact for users.

Hi Chris,

I understand that your engineers need more time to fix the algorithm, but the major complaint here is that sleep gets overwritten by a nap. 

Why not make a simple workaround by adding "Save sleep/nap Yes/No", immediately after the "End sleep Yes/No"? That should not affect your algorithm at all, and we will get an opportunity to discard fake naps/sleep.

This should be done immediately and included in the next 13.xx beta, while your engineers investigate this and work on improving the algorithm in the future.

Z said:

Moderator, users are repeatedly requesting to not replace prior night's sleep, regardless how long nap is

That I completely agree with and that is the root issue here. One way for the Garmin engineers to achieve this is to use the sleep score and only save the sleep with best sleep score for that day; I.e., if already a sleep saved that day then do not replace if new sleep, fake sleep, nap, or whatever has a worse sleep score.

Often these fake naps contains a lot of awakening time which result in a low sleep score. Most, if not everyone of my fake sleeps has 0 sleep score.

With that said, since it seems such a big thing to change the algorithm then I suggested to at least give us/me some option to define myself when sleep algorithm should be active. This does not prohibit any future improvement in the sleep algorithm itself to avoid overwriting of sleep, but meanwhile we wait for that to happen why not give us som more control, that doesn’t hurt.

90de6a656fe046f0976b71c0d1c6d954 said:

Why not make a simple workaround by adding "Save sleep/nap Yes/No", immediately after the "End sleep Yes/No"? That should not affect your algorithm at all, and we will get an opportunity to discard fake naps/sleep.

That would be an acceptable solution for me, but only for sleep outside sleep window. I do not want to risk press the wrong button when maybe still a bit sleepy in the morning causing me to loose that nights sleep

However one thing is that when these fake sleep has happened I have never got any option to end sleep unless actively press end sleep when find out in widget that I obviously seems to sleeping. More often the fake sleep has ended and just replaced the sleep without any question at all about ending sleep.

Can understand how frustrating it is to have sleep overwritten or have the watch detect sleep while watching television.  Fortunately I have only had it happen once.  For many of us, the auto sleep detection works great and adding steps to say yes it's still working great would also be frustrating.  What I would prefer is they offer the option to turn off auto sleep detection so that those of you who want to manually choose which sleep to record can.  Those of us who don't can just keep carrying along with these amazing watches

OnTheEdge said:

What I would prefer is they offer the option to turn off auto sleep detection so that those of you who want to manually choose which sleep to record can.

Yes, a simple enable/disable automatic sleep detection would be a big improvement. When disabled the sleep detection has to be enabled/disabled from the sleep window. This would let all of us that have the issue with fake naps wear the watch 24/7 again while the engineers continue to work on a better sleep detection algorithm