Sleep tracking on Forerunner 745 does not work for me

Hello, the current status with me and the problem of the topic is in short: it still does not work, but there is an explanation :-)

Here are the records of the last days, mostly completely unusable or non-existent:

A halfway usable like yesterday (image 2 above) entry then looks like this:

If I then change the sleep time to the real time of falling asleep, it's almost ok: 

Unfortunately only in about one out of ten cases, max 10% ok. As you can see in picture 1, the trackings are mostly unusable.

I have been very busy with the topic, because the Forerunner is really a very good sports watch and fitness tracker 24/7: accurate, comfortably geared to real sports, long battery life. I have also tried other watches, example Samsung Galaxy Watch 4 (completely impractical for sports).

These current candidates from the higher price range all had the same problem with me, with sleep tracking and other, such as HRV measurement.

Sleep tracking is very important for me, because I have problems with sleeping and want to have an overview of my sleep times, with shift work, preferably also with naps for example, I'll write about that later below. The Forerunner 735 XT with its basic tracking software was already really very useful.

That in the Forerunner 745 since version 5.20 now in addition FirstBeat is used, one can derive for example also here: forums.garmin.com/.../basic-sleep-tracking-disabled-hr-sensor-stop-working-after-5-20

Following about the FirstBeat software, which is now built into the newer Forerunners: you can use the deeple.com or Google Translation or the English version of the pages:
www.firstbeat.com/.../
www.navigation-professionell.de/.../
discover.garmin.com/.../

Here more about HRV, in short and german (use deepl or g  translate): sportuhrenguru.net/.../

Translated with www.DeepL.com/Translator (free version)

More about me: I have had chronic cardiac arrhythmias, atrial fibrillation (AFIB), for a few years, which do not bother me much in everyday life and sports. Any watch software that now recently adds heart rate variability for sleep and stress evaluation is overwhelmed with it.

I noticed the whole thing when in the Forerunner 745 also the activity "HRV stress" did not work. After that, I dealt for the first time with the unfocused topic of athletic stress control with heart rate variability (HRV):

After all, the normal heart doesn't beat regularly, but with slightly different intervals - when it's relaxed, it's loose, so to speak. And under stress it beats much more monotonously. This is the basis why HRV can be used to control stress during sports and recovery, and it is used a lot nowadays, especially by professionals. And the Forerunner offers a good and practical solution here.

But with AFIB, the differences in the intervals between heartbeats are much larger and become uniform only with heavier sport load, that's why HRV software in default setting can't handle it. An RMSSD value of 332 ms on a normal Wednesday afternoon is of course completely out of range for such a small piece of software:

Possible solutions: the software (Firstbeat works on the topic AFIB too, as far as I know) recognizes AFIB, that is nowadays already possible. The software issues a message "AFIB" and sets the threshold for the beat correction higher, so that a usable range comes out, see picture two and three in the previous posting.

Possible practitioner solutions for the near future:

1) A way to manually start a sleep recording from the sleep widget, just like you can already manually stop it today. Then you could at least record your naps with, or a more accurate sleep time, see my manually corrected recording from yesterday. (A sleep "activity" would also be imaginable :-) )

2) The possibility to set the sleep detection back to "Basic", so only motion detection. That would also help Filip from the topic forums.garmin.com/.../basic-sleep-tracking-disabled-hr-sensor-stop-working-after-5-20 and probably others.

1) or 2) would be great and if Garmin needs me as a test person, just contact me !!!!

Many greetings,
Arno

Hello Arno,

Thank you for providing so much content regarding the topic. I just bought the smartwatch and I see the same pattern happening to me, but just the opposite: last night I woke up once around 2:00 and once around 4:00. So the full sleep tracking happened only from 4:00 to 8:00.

Any suggestions on how to approach improving the experience? And, did Garmin try to update/fix this?

Hi Bojan,

unfortunately Garmin has never responded here. In addition to the larger problems that I have also otherwise discovered with me, this simple BUG is also still there.

I once read something that an interruption from 5 minutes leads to termination of sleep tracking. But that's not true, because with me individually also wake times of one hour have not led to interruption (Since I have brought my daughter to school in my bedtime and then put me back to sleep).

The simplest solution that Garmin could offer for this would be to manually start and stop tracking in a sleep activity that overwrites the automatic start and stop detection. So that only the quality of sleep is analyzed by Firstbeat.

Garmin would probably also be relatively independent of Firstbeat to develop that. I think that is here with the topic namely the other problem that Garmin does not support at all for FirstBeat.

Some observations I would still like to write down here. They are of course only guesses, since you get here apparently no explanations or help from Garmin, unfortunately.

Maybe another user still has ideas or information about it?

1. starting and stopping the tracking is also strongly related to the stress index. In the sleep score widget, this is also a named variable.If the stress index is calculated from the HRV with the Baevsky formula, regeneration variations (or AFIB as in my case) could lead to irregularities in starting and stopping. Firstbeat does not, to my knowledge, describe anywhere exactly whether they use the Baevsky formula or a variation.

(Baevsky: www.kubios.com/.../)

(Firstbeat: assets.firstbeat.com/.../How-to-Analyze-Stress-from-Heart-Rate-Variability.pdf)

2.) There are algorithms in the software that use previous user behavior or data, sometimes more or less well. This is only a guess due to the poor information available. But, by a simple reset of the clock everything can be set to zero again and then also the measurement changes quite strongly, first.

I'm pretty sure sleep is only recorded from stress level recovery.

I found the original F745 Firstbeat Sleep function on my watch was pretty bad.

On the "Quantified Scientist" youtube page, I saw that Garmin sleep is pretty bad. He found that all Fitbits were much more accurate with sleep monitoring. So, I bought a cheap Fitbit 11 months ago and started wearing it at night to track sleep. It does pretty well. I have about six months of solid sleep tracking, and I understand my sleep in a much more profound way (tracking is educational).

With the new F745 firmware update to 7.6, Garmin pronounces that the sleep algorithm is improved and polished. Okay.

I'm going to take off my cheap Fitbit, put on my not-so-cheap F745 with the latest firmware, and give it a week or two of sleep tracking to see what Garmin can now do.

Does anyone else have experience with the new 7.6 firmware update and its impact on sleep tracking?