Sleep tracking fixed in 4.20?

I think the root cause of the majority of the problems on the 235 was the decision on Garmin's part to reduce the all day sampling rate. They released a device with an advertised battery life that was 5 times more than reality. To "rectify" this they basically broke the watch by reducing the sampling frequency.

Still too soon to say all is good, but my sleep time last night was spot on. That being said, it has been spot on before. The issue for me is days when I relax in the evening/early night and it thinks I'm asleep. My Fitbit Charge HR somehow managed to differentiate TV watching from sleep. My Garmin seems to place any "lazy time" after my regular configured bedtime in the sleep column.

My Charge HR also caught almost every nap I took. My 235 has NEVER managed to detect a nap for me.

The Forerunner 235 tracks movement throughout the day and while users are sleeping. Garmin Connect processes the data based on the sleep time that user enter when setting up the device along with accelerometer data in order to display the sleep time information. The website teams are working on an issue were Garmin Connect will display odd or irregular sleep data. This is an issue with how Garmin Connect is displaying sleep data, not the way that the watch is detecting sleep. As the website team continues to work on this issue we expect more users to see better automatic sleep detection. I do not have a time line on when the issue will be fully resolved, but this is something that is actively being worked on.

Since the Forerunner 235 does not have a manual sleep mode you will not see any nap times displayed in Garmin Connect.

I also have a sleep issue. GC is displaying sleep before actual sleep is taking place.

Also, its not showing much hr monitoring during sleep. It looks like only once or twice per night, then it starts up with movement on waking at a regular interval.

Also, its not showing much hr monitoring during sleep. It looks like only once or twice per night, then it starts up with movement on waking at a regular interval.

WHR version 2.60 is sampling RHR every 15 minutes. See thread: OHR Polling frequency when asleep/not moving (vs visualization on 4hr-graph). My HR graph appears to flat line and not sample during sleep, but after viewing my files I found my HR fluctuated by 2-3 bpm while sleeping. This would be normal and such a minor change wouldn't appear on the graph. I suspect that's what you're experiencing also.

To "rectify" this they basically broke the watch by reducing the sampling frequency.

I disagree. Changing operational parameters in a manner that is perceived by some individual users as handicapping particular features and/or reducing the device's capability, or even taking away something to which they have become accustomed, is not the same thing as “basically broke the watch” from the perspective of meeting or not meeting product specifications as marketed/sold.

It is those individual users' prerogative to be disappointed, but they're in no position to validly claim their expectations and ‘standards’ of acceptable operational performance are representative of the broader base of product owners and paying customers.

Still too soon to say all is good, but my sleep time last night was spot on.

It was spot on for me, when I reviewed my sleep data for last night just now. However,

GC is displaying sleep before actual sleep is taking place.

when I was looking at it last night, Garmin Connect reported that I was sleeping at 7PM, when I'd simply taken off my FR235 to protect it from mishaps while I cooked dinner. It appears that the system has elected to re-interpret the captured data since and present a different set of conclusions.