HEILRATH are you on SW 11.28 and did you had a nap or fake sleep lasting shorter than 3 hours that replaced your nights sleep? Or was it longer than 3 hour? Real or fake nap?

My mistake, I know… but kinda funny. I took my watch off this afternoon and let it “sleep” on a table. Indeed it slept! For a little longer than 3 hours, and got a bad score that replaced the one of my sleep, hence messing with the training recovery hours. Next time I will take it off, I will make sure to switch it off or simply to turn it upside down, so it won’t collect fake HR and sleep data. 

MarEvaCre said:

My mistake, I know… but kinda funny. I took my watch off this afternoon and let it “sleep” on a table. Indeed it slept

️

MarEvaCre said:

a little longer than 3 hours, and got a bad score that replaced the one of my sleep

We were told that this is intentional and "this is by design":

https://forums.garmin.com/outdoor-recreation/outdoor-recreation/f/fenix-7-series/314681/night-sleep-data-gets-replaced-by-daytime-sleep#pifragment-1292=11 

Moderator Update: If the watch perceives sleep for more than 3 hours, it will replace the prior night's sleep. If there is sleep recorded less than 3 hours, I believe there remains an issue that still needs to be resolved. Less than 3 hours should not be recorded or synced to Garmin Connect replacing the prior night's sleep.

MarEvaCre said:

Next time I will take it off, I will make sure to switch it off or simply to turn it upside down, so it won’t collect fake HR and sleep data. 

Interesting. I take mine off and put it on the table and it doesn't show any heart rate data. It certainly does not record any sleep.

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

Z said:

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

I could not agree more.
Garmin - Chris
It is to early to say if the do not save sleep if sleep less than 3 hours now works as it should or not. With that said it is still not a good solution as this example shows.

I went to bed at about 23:30 but watch thought I started to sleep already at 23 when my sleep window started when watched TV. I did not had the best of sleeps during that night but stil it was 6 hours and 11 minutes. Due to this I was a little bit tired so I decided to take a shorter nap, but it become a little bit longer than planned. This long nap that lasted 3 hours and 7 minutes replaced my longer and better sleep. That is not what I want or expect. I think the algorithm works bad due to this.

I manage to take a screen shot in Garmin connect App just before the watched completed the sync from watch and overwritten the existing sleep in GC App as well

Happened at another occurrence as well when I went for a 2,5 hour nap, but since nap was followed directly after watching some TV from Sofa, the watch had started to register sleep 70 minutes before that and I got a total sleep of 3:40 despite that the actual sleep only was 2:30. Result, the better nights sleep was replaced by this 3:40h sleep.

Please let the developers know that they have a bad design and that there is still a problem, for us the users, even if sleep lasted more than 3 hours.

SETM said:

Please let the developers know that they have a bad design and that there is still a problem, for us the users, even if sleep lasted more than 3 hours.

I could not agree more.

SETM said:

Please let the developers know that they have a bad design

I do agree also. It is pretty clear that the watch needs to have "a night of sleep" data to calculate the training readiness (sleep and sleep history). So they need to put a threshold to make a distinction between a "nap" and a "night of sleep". From that perspective, 3 hours is not an absurd threshold.

But the bad design is making training readiness depending on sleep data, which will remain unreliable, even if all the "bugs" or limits of the algorithm are ever addressed.

I am suggesting that training readiness be based on body battery instead.

Sleep data should still be fixed, but only for people (like me) who care about sleep analysis per se.

And body batter is based on sleep, etc. Then training readiness on the body battery... Basically without a good sleep/rest all metrics will be messed up ;)

Anyway, it's ridiculous that we are talking about the same for how long? Months? Years? And all the time moderator is coming back with 3 hours.