Sleep stages and cycles accuracy

I haven't looked at how it scores sleep but I did spend 3 months with a Garmin 255 on one wrist and a Fitbit charge 5 on the other. The results were fairly stark - the Fitbit was far worse at tracking sleep in any useful way. 

I wouldn't say the 255 is perfect by any stretch but after 18 months I would say the score tends to match how I feel (with the occasional anomaly). I've been through a period of giving up smoking, which destroyed my sleep for months. Giving up alcohol which really helped my sleep. And finally giving up Nicotine/Nicorrette, which destroyed my sleep again for months. Through all these phases the sleep score has generally reflected how I feel and the REM/deep values looked r about right.

Saying all that - I got a software update yesterday and a deep sleep value of 0 last night which is patently nonsense

Thank you for the input.

Oly said:

and the REM/deep values looked r about right.

How do you evaluate the amount of deep/REM sleep you actually have had in a given night ?

I only can recommend the https://www.youtube.com/watch?v=-OBCwiPPfEU HubermanLab podcast with Matthew Walker about sleep.
Further I found a study https://pubmed.ncbi.nlm.nih.gov/36414751/ where Garmin tracker were used.
We can't expect that a tracker like a F 255 is comparable with a sleep laboratory, but I guess it's similar like this BMI body fat scales, the tendency is visible. And I remember a quote from a Garmin developer who said the if you're using your sleep laboratory report and give it 3 different people to interpret you'll get 3 different interpretations. And that the Garmin tracker is right by 75 - 80%.
I by myself def feel it if my sleep score is 80 and higher and I also feel like I had an accident with a bus when my score is 60 or lower.
This watches can be boone and bane ... we should use the as a tool

Thanks. Yes I have watched the first first 5 parts of the Matthew Walker's podcast with Uberman and they are an expansion of Walker's book I suggested in my first post in this thread. An excellent read.

I had read the study you link above but regettably, the "Conflict of Interest" statement makes me cautious about its findings.

I agree I think overall myself if my sleep score is over 80, I can estimate I had an ok night. Above 90, I had a great night. And in the 70s or lower, something went wrong.

But... what I personnally come to realise is that this can be evaluated with the overnight stress score alone. Because its weighting in the calculation of the sleep score appears to be so prominent, the other factors matter a lot less. As in, I only look at overnight stress score and I already know if it's consistent with how good I feel in the morning.

Matthew Walker in his book (and the various podcast he participated into) goes way beyond that obviously and explains of the importance of deep and REM sleeps, my interest in "how to use my watch to this effect" is the direct consequence of this. Walker is the reason why I am looking into this.

Well, I also watched the 5 parts of the podcast and what I use as a take away is QQRT. I think that is key!
I agree also with your argument about the over night stress level.
Well, but consider that as more deep sleep you get over night as lower the stress level should be.
We should also consider that the sleep score from my point of view is a cane. It's an interpretation from overnight HRV, Rest HF, Bodybattery and so on.... And that then is presented to us as Sleepscore. Similar to this morning report, that also only a thing for our convenience, isn't it. I like to create my own morning routine and there I shut the report down.
But I really like it to see that there are other people interested into this topic.
Well, I guess my point of view is clear. I wanna use the watch as tool especially for my common health. Because Stress, Sleep, HRV and Nutrition, especially to avoid sugar and proceed food .....
I'm *** impressed by the HRV and how it is detecting infections an other abnormalities....

unimatrixzero said:

I'm *** impressed by the HRV and how it is detecting infections an other abnormalities....

I share your view. Let me tell you a story.

Mid november last year out of the blue Garmin started recording high levels of overnight stress for me, in the 30s, for no apparent reason. Obviously very low sleep scores, and despite getting copious amount of time in bed, I was restless all the time. I thought it was some kind of winter bug, may be covid or whatever, so took anti-inflammatory and aspirin... but it wouldn't go away.

Two weeks later (early december), I started having more visible symptoms: night sweat, mood swings, weight loss, and a few other I forgot.

Then mid-december, I felt a lump on my thyroid. Looking up the symptoms for thyroid cancer, I discovered that I had almost the entire list of symptoms. I rushed to the Dr. An ultrasound and a biolpsy later, it was discovered it was in fact a benign thyroid inflammation that the Dr treated with hyperthyroidism medication. Early February, blood test showed normal levels of thyroid hormones, we stopped the medication and my stress level was back to normal.

I was very impressed that my watch knew 2 weeks in advance that something was coming.

JRKingOfBeer said:

This is, of course, totally impossible.

Here is a logical trap - you suppose Garmin did calculation alike for a healthy person. But not for you ;) Whole measurement was beyond the limits. It was a data garbage. Funny but not reliable. 

And yes, garmins sleep tracking works for me pretty accurate too.

An extremely inaccurate method:

* REM - mood, state of mind, temper, dreaming

* Deep - you tend to notice when you aren't getting much deep sleep as it feels like you are awake a lot of the night. Physical recovery - I do a lot of running and notice how recovered my legs are

I've been wearing various sleep monitors for enough years I can pretty much guess my sleep score factors.

global_local said:

Whole measurement was beyond the limits. It was a data garbage. Funny but not reliable. 

I'm not sure I understand the trap in which you think I fell but what we're saying is the exact same thing:

JRKingOfBeer said:

I am probably ok ignoring sleep reports after a night of drinking, as these have become quite rare now, and focus on the "normal" days. I presume Garmin's algos are considering this as a pre-requisite anyway.

I don't think it's accurate and I've stopped wearing mine during sleep.