Deep sleep accuracy?

@Elbo, i think the real situation is WORSE than you say. Proper sleep stage determination also requires independent observers . Google: polysomnography

As a result no watch will EVER have accurate sleep stage information.

And anyone who says well Garmin/Fitbit/Whoop et etc 'feel' right to me should just take solace in their feelings rather than the science.

Why does anyone want  to know sleep stages? if you believe that deep sleep gives muscle recouperation then, fine, but just take waking HRV readings instead to see how your body is recovering from stressors.

the5krunner said:

Why does anyone want  to know sleep stages?

That's exactly what I always ask. And what if my watch reports I have not enough deep sleep, how do I change this? 

For me, it's quite simple. Ask yourself in the morning, am I ready to get up or do I want to sleep some more. In case of the second, you have to change your sleep routine. No watch can help you there.

But it's the same with the new stamina feature on the mega expensive fenix 7 series. I'm perfectly capable of feeling when I hit the wall. Don't need a watch to confirm. 

But people want more data. Just to collect it and not really use it (or know what to do with it). Garmin (and other brands) sell a lot of new watches thanks to that.

So the problem is that judging if you are ready or how you feel is anecdotal data (i.e. personal point of view) rather than quantifiable data. My problem is that the inaccuracy of the Garmin sleep metrics impact other scores such as Training Readiness. You should check out the Quantified Scientist YouTube channel if you want to know the science of how accurate wearables are. I put up with these inaccuracies as the other data I get from my Fenix 7 is more valuable to me. I wear a Fitbit Charge 5 on my other wrist solely for sleep data. Compared to EEG devices, the Fitbit is more accurate at tracking sleep stages than most other wearables. Obviously it would be great if Garmin would get it together and fix this so I can just wear one device but this obviously isn't going to happen.

I regularly go into a performance lab to get my ventilatory thresholds (VT1 & VT2) through testing such as gas exchange, etc. to understand how I am responding to training, when my body starts producing lactic acid, VO2Max, etc.. This is very helpful as I then use the data to monitor my heart rate so I know during a training session if I am accomplishing my goal of low aerobic, high aerobic or anaerobic training. I wear a Polar H10 connected to my Fenix 7x to gather data and do trend analysis over time to track performance improvement. So if you ask me if my performance is improving I can answer yes and show you the data to prove it. I am a high altitude mountaineer so my training is important. I understand how my body reacts to altitude and I adapt my training to optimize my performance when climbing.

I noticed the same. If I check it right after I wake up there's deep sleep and then it vanishes later. I printed screens to make sure of this. Disappointed...

Yes, the sleep algorithm seems to be heavily weighted toward light sleep and sometimes REM at the expense of deep and awake times. I noticed this when I had the 6x as well and it consistently gave me low awake and deep sleep every night. The 7x is better but it is still an issue. This of course throws off the sleep score and training readiness as well since a component of the calculation is time in each stage. I have had to call support on quite a few occasions as the sleep system seems buggy between the watch and the app. Since there are a lot of devices that do a better job of sleep tracking you would think that Garmin could fix this relatively easy with some adjustments via a software update. I agree with the disappointment.