Related
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

HRV bug followed me to new watch

7714909
over 1 year ago

I discovered an HRV bug after the February update when I had Instinct 2s. I then thought that was what did it.

Over time I've discovered it tends to strike maybe 4-5 days after I come down with a virus or other illness. HRV nightly average goes up way above the actual average when looking at the graph for that night. Sometimes the average is higher than the highest 5m value.

Then I got the 255s Music. Right after I started getting the dotted white line I got a really dramatic spike. Because of how Garmin watches calculate nightly average, you will get a very wrong average for that night if you happen to get a spike of that kind (which can happen if you're up for a while).

Now, I believe the bug struck again at that point because the watch didn't have more than a few days of averaging to look back on instead of weeks. And with that spike, even a normal HRV looked to the watch as similar to when HRV went down because of a virus.

This bug has continued to crop up regularly since, and spikes too.

That makes status and dotted line completely unusable, because few of the nightly averages are actual averages.

I need the ability to edit nightly average to get out of this and not have to create my own graph to get reliable data.

And it would be good to figure out why this happens.

I had a long discussion with a support specialist at the Instinct forum, but now this is spanning two watches. He said then that it wasn't a wide spread fault. To which I replied that most people don't check the graph, so how would they even discover this bug? So it could be wide spread and nobody would know.

If you get a virus or other reason for HRV to go down for more than three days, could you check your nightly graph and see if the average it calculates is the actual average?

Egregious example of how maddening it can get when the bug strikes. These kinds of results are rare, but wrong average is more common.

  • 0 over 1 year ago

    I had covid earlier this month, so I checked the nightly averages for the covid period along with the week after recovery. My numbers seem to match up well with the graphs. No strange readings like yours. There was a positive HRV spike the first day I felt recovered, but the entire graph was high. 

  • 0 over 1 year ago

    I've had a look at the last ~2 weeks, and my averages seem reasonable. I'll pay closer attention to this going forward, and report back if something weird happens

  • 0 over 1 year ago

    Are you really sleeping from 7 till midday? Was it your normal sleep behavior? Have you set it up in GC?

  • 0 over 1 year ago in reply to global_local

    That was a night I couldn't sleep. The bug struck last night as well. I have just adjusted sleep. But that's not the reason for the bug. It strikes even if sleep falls within my schedule.

  • 0 over 1 year ago in reply to 7714909

    I think sports watches are aimed at people who are healthy enough to lead a well-ordered lifestyle. So cases like yours were unlikely to be taken into account by the algorithms. A watch simply cannot take into account such complex things as the body’s fight against a virus, etc. They only cover basic behavior.
    I myself checked my data for the last 3 weeks - everything is normal and understandable - heavy load - drop in HRV, enough recovery - growth and fixation in the middle of the scale. The average is always 2-3 times less than the maximum.

  • 0 over 1 year ago in reply to global_local

    Correct me if I'm wrong, but I don't think the exact value of the HRV is the issue here. For me it looks like the average HRV is wrong.
    On the screen  provided, the chart suggest a lower HRV average than the one show by the watch. Here, maybe this modified screenshot will illustrate the point:

    I've marked the "average" value of 39 with a red line. Almost the whole chart is below it, which doesn't seem right

  • 0 over 1 year ago in reply to Cooad

    We can discuss possible errors ve-e-ery long (but they are (may be) absent). But it is still a black box for us. My point is - the measurement is out of range of defined sleep times and can't be applied to the shown graph. 

  • 0 over 1 year ago in reply to global_local

    That makes absolutely no sense. I WAS sleeping during this measurement.

  • 0 over 1 year ago

    I don't think this can be defined as outside sleep time. The example isn't quite as obvious, but it's still part of the bug pattern. I've had many bug nights since I got the watch. I'm considering resetting it unless I can get out of this pattern caused by the spike one of the first nights.

  • 0 over 1 year ago

    One more bug night. Can't pinpoint 31 here, so marked 30. It's obviously wrong.