Disappointing Sleep tracking after new update

pilleus

@Iggge said in Disappointing Sleep tracking after new update:

Any idea how to test this?

If you have a WearOS watch, then you can capture your HRV 24/7 and regard the values as graph in a web interface. Correlatioins to heart rate and stress are shown very good and transparent.

Just install the app Stila out of the PlayStore and have fun (available only for Android). It is a project of the Ludwig-Maximilian-University in Munich.

Stoke80

@Iggge

For me the HRV values from my SV do make sense. I state this because of the following reasons:

I checked on Apple Health with the average values I had when I was wearing an Apple Watch (October 2020 until November 2022). And they are about the same (bit lower now)
They are in line with (even a bit higher) than the values on the link from @SuperFlo75.

Sure that doesn’t tell that the Suunto values are 100% accurate, but for me they make sense, because of the mentioned reasons.

Iggge

@pilleus Thanks! Done that now!

Iggge

@Stoke80 Thats good! I want to trust my Suunto, but I don’t. All data is showing different. On my SV I have 140 BPM peaks, My fitbit none.
My steps altso differ alot

duffman19

@Stoke80 Apple reports HRV using SDNN readings. Almost every other platform uses rMSSD. These two statistics almost never match, with SDNN typically being higher than rMSSD, so they cannot be compared to one another.

I don’t think there’s any documentation as to how Suunto is calculating HRV, but I’d bet on rMSSD as it’s the more common and accepted number used by others such as Garmin and HRV4Training.

Here’s some info on the difference between the two stats: https://welltory.com/rmssd-and-other-hrv-measurements/

DrSilverthorn

Since I haven’t had much luck tracking sleep on other devices, it’s not so important to me. Plus, I know when I slept well and when I didn’t. Garmin probably worked the best in this regard, but it chronically underestimated my sleep.

The bigger issue is that you need to track sleep to get HRV estimates. I wouldn’t mind having HRV, since it can be indicative of a few things. Unfortunately since the update, its frequent that my SV tracks no sleep whatsoever. So I don’t get any HRV estimate either. Funnily enough, I had one night where is seemed to do pretty well for estimating sleep, but all the other nights have been terrible.

Yes, the watch is snug on my wrist, but not to the point of discomfort. I think the watch is tracking my min HR properly, but for running, the HR sensor doesn’t work well at all. It seems that the poor sleep tracking may also be linked to the poor HR measurements.

Really strange that some users have such issues with inconsistent data. I can only say for myself that I can compare it with another device (whoop) and the data seems to be correct. I also have low HRV, but I am out of shape at the moment and also whoop measures approximately the same values. I am 37 and my HRV is around 38, so yes it can be low also when you’re “young” (not sure if I am). Plus it’s also genetics. My girlfriend has super low resting HR I can only dream of and her HRV ist always (a lot) hihg(er) than mine. Her Zone 1 ends where I am having already a heart attack. So the best would be a second device to double check and compare. I think it’s super individual and what works for one doesn’t have to for others… unfortunately.

pilleus

Mean RR Duration, Mean Heart Rate: Indicator of stress. Is typically measured during training and interpreted as a measure of the intensity of the load.

SDNN: Shows how well the autonomic nervous system can regulate the processes in the body.

RMSSD: Indicator of ability to recover, fitness and health. Shows how quickly the body can react to stress. This allows conclusions to be drawn about the level of stress.

pNN50: Well-founded measure of parasympathetic activation.

https://knowledge.time2tri.me/de/articles/die-wichtigsten-hrv-parameter-teil-1-zeitbezogene-parameter

Brad_Olwin

@pilleus said in Disappointing Sleep tracking after new update:

Mean RR Duration, Mean Heart Rate: Indicator of stress. Is typically measured during training and interpreted as a measure of the intensity of the load.

SDNN: Shows how well the autonomic nervous system can regulate the processes in the body.

RMSSD: Indicator of ability to recover, fitness and health. Shows how quickly the body can react to stress. This allows conclusions to be drawn about the level of stress.

pNN50: Well-founded measure of parasympathetic activation.

https://knowledge.time2tri.me/de/articles/die-wichtigsten-hrv-parameter-teil-1-zeitbezogene-parameter

Although these values can be useful, I would not over interpret them.
https://uphillathlete.com/aerobic-training/why-we-stopped-relying-on-hrv-apps/

pilleus

@Brad_Olwin said in Disappointing Sleep tracking after new update:

Although these values can be useful, I would not over interpret them.

I know and I don’t …

AnonVerticalOwner

@AnonVerticalOwner said in Disappointing Sleep tracking after new update:

Another night, another report back: HR was tracked properly for the duration of the night this time, and the sleep data is consequently pretty spot-on.

Another two nights, same setup: Still having the strap tighter than usual (i.e. beyond comfort), both nights the wake-up time was off by 2 hours. Interestingly enough, one of the nights the HR graph was blank exactly where I had to loosen the strap (by 1) due to numbness, which is from where the watch thought I woke up, despite falling asleep again. For the other night it showed a perfect HR graph, but was still off by 2h.

I’m not sure what to test next, having done all recommended plus what I could think of myself. The HR sensor and/or sleep algorithm definitely seem more sensitive, at least. This morning I woke up with my hand having gone numb, like, that’s how tight it needs to be to avoid gaps and be accurate right now. I’ve had the SV since release without the need for this and with pretty accurate tracking.

I suppose I’ll await some acknowledgement from Suunto (…) or — if I start losing sleep over it (ha) — bother customer support.

Over and out!

wmichi

@AnonVerticalOwner Did you try to shave the hair off where the watch is? For me OHR works a little bit better this way.

Iggge

Its so strange! My fitbit and Suunto shows exactly same time for when I fell asleep, and when I did wake up. But the Suunto always shows like 2-3 hours awake every night.
Can I turn this off or something? Clearly it don’t work

Mff73

Since two days, my SV tracked my sleep correctly
Coincidence or it needed some learning from previous nights (HRV or anything else ? ), let see.
As i am not exercising these days, i have just these data to look at

AnonVerticalOwner

@wmichi You’d hope that wouldn’t be necessary! For what it’s worth, I’m not particularly hairy there, so I should be good not testing that

I’ll give it a few more days. Currently, my best guess is that the new algorithm is just very quick to judge whenever HR values are missing (or changing), deeming the wearer awake, and not really considering the bigger picture.

E.g. you’d expect when before and after a data gap the HR is 40, that it would assume nothing happened in between, rather than make up events.

peegee

@Mff73 I also got an accurate sleep reading last night. This is a first and I’m extremely happy with it
I just can’t understand what gave me an exact reading since I haven’t made any changes to watch’s definitions.
Will keep following and try to understand what causes off tracking…

Patrick Löffler 0

Which has currently the better sleep tracking Race or Vertical? And second question. An even better one. Do you know what factors the sleep score is calculated on?

wakarimasen

One general comment which (I think) would benefit sleep tracking, is the ability to actually edit the start and end times after recording. We all know that this metric can be a little hit and miss, so being able to ‘correct’ the times might help

AnonVerticalOwner

@AnonVerticalOwner said in Disappointing Sleep tracking after new update:

E.g. you’d expect when before and after a data gap the HR is 40, that it would assume nothing happened in between, rather than make up events.

Definitely this ️ Another night with approx. 2h short, and the (awake and wake up) times match the two brief gaps in the HR tracking. The algo hasn’t been this quick to draw to (incorrect) conclusions before.

Iggge

Just to correct myself. ButI think i had wrong, and the watch was correct. My HRV is about 40, as other watches is showing same results. No idea why its so low, but what I understands is that is quite low (bad?) for a 32 years old male. I exercise alot (Gym, running and some walking), and Vo2 is between 55-60 most of the time.

However, sleeptracking is still not good enough and never shows over 5h(usually between 3-4 hours), when in reality i have been aslep for much longer (6h and over)

Overall, still like the update very much! Good work Suunto!