Sleep tracking issues with 10.10 - increased (unrealistic) awake time

Do you have other metrics such as HR, movement, respiration, stress, PulseOX (if it's enabled) recorded in these periods?

Same issue for me. 

Watch recorded 45min of (unrealistic) awake time last night. Seems, that the algorithm is too sensitive currently. 

I am also having sleep tracking issues and it hasn't been working properly for the last few days. It is unacceptable for software issues like this to cause such a disruption in people's data that they rely on.

Yes. I have them all, except pulseOX which is disabled.

For an update, tonight, July 16, it recorded 59min of awake time.

As for the other sensors:

- Respiration rate hasn't changed after the update - I'm still getting similar values

- Stress hasn't changed much after the update. Even when the watch recorded 1h44m of "awake" time on Jul. 15, the stress was recorded as being in the "rest" range for the whole sleep, except around 8 minutes of "medium", so that doesn't correlate to the awake time being recorded.

- Body battery, which correlates strongly to stress while sleeping, also hasn't really changed after the update, and still recharges similar amounts to what it did before.

- Movement seems to be the problem. This varies each night, but usually stays under 1/3. Before the update, it would only pass or get close to the halfway point when I actually got out of bed to get water or go to the bathroom, and would add a few minutes of awake time. In the 5 nights since the update, though, I have movement spikes reaching 100% on several occasions, without getting out of bed. This leads me to think that the first problem is that the sensitivity of the sensor was tweaked, and it's now treating normal sleeping movement as if I were walking or getting out of bed for minutes or hours every night.

For example, here's July 4, before the update, in which I did get out of bed for water. You can see the 4 minutes of awake time recorded when I did that, and the spike in movement. This is realistic.

And here's July 15, post update, a night in which I didn't get out of bed or even recall waking up, and the watch added 1h44m of awake time. The same weird 100% spikes can be seen on Jul 12, 13 and 14, all after the update.

- In addition of the suspected sensitivity change of the sensor, I think Garmin tweaked the algorithm that interprets the data. For instance, comparing my sleeps on Jul 8 before the update and Jul 16 after the update, you can see very similar sleep patterns. Yet, on Jul 8, the watch calculated 2 minutes of awake time, which is realistic, but on Jul 16, it added 59 minutes.

Here's July 8:

And here's July 16, where you can see the "awake" minutes in pink.

So at least from my experience, I believe that there are two possible causes for the increased, incorrect awake time:

- the sensitivity of the sensor was changed, so movements that are relatively normal while sleeping are now being recorded as if I were walking around, for up to hours. This referes to the data coming from the sensor itself.

- the software that interprets the data was also changed and had its sensitivity tweaked, so even if the sensor happens to record my normal movement data, the interpretation adds a lot of "awake" minutes it wouldn't before the update.

This could explain the jump from 2-10 minutes "awake" each night to 30-120 minutes. And trust me, I didn't suddenly start sleepwalking or drinking before bed the same night the sensor was updated. 

Same here, definitely getting more awake time than I used to and I'm a pretty heavy sleeper.

Agree with your theory, i have the same. Usually at he bathroom 1-2 times every night, and 1 hour 43 awake, they have changed something, for me i feel the «old» sleeptracking was much better, now i wake up and with the score and following comment from the watch i am «supposed» to feel ***, wich i don’t, but well rested. With the old tracking it was much more like i it showed up *** data, i felt the same, so that miraculous firstbeat is not so magic after all. 

Agree with all the comments.  I saw this happening when they introduced the sleep changes in the 9.XX betas that included the added sleep widget.   Lots more awake periods than before and I certainly don't seem to recall them.  

Question: previously did the sleep calculations take place within the app after the watch synced data like movement/heart rate/etc rather than on the watch?  It now looks like the calcs are done on the watch because I have immediate access to the info on my watch through the widget when I wake up but the data is not in the app until after it performs a sync with the watch.

Same here, definitely getting more awake time than I used to and I'm a pretty heavy sleeper.

I'm chiming in here as well to record another user with similar experiences with the 10.10 update on a 6X Sapphire .  Sleep tracking seemed much more reflective of how I felt prior to the update.

Seems broken now, too sensitive and recording far too much awake time.    Not cool to be a beta tester when I'm not installing beta software.  This is an expensive toy to have Garmin break what already worked.

I love this watch, but come on. 

Same issue for me sleep tracking doesn’t work one night it said 7, the next night was 4 hours, I went to sleep at the same time and hit up at the same time nothing changed.

Last night my 4 hours of sleep recorded 3 rem sleep cycles but no deep sleep, Garmin’s sleep tracking is absolutely useless.

considering I went to bed at 9pm there is no chance I was awake for almost 3 hours.