Track recording currupts... again (16.17)

Already done it before. There is no chance to get a reaction from them. They told me to change my watch in shop where i bought it. But it was a present from my overseas colleagues 

Unfortunately, the 17.xx beta is also buggy. I installed it today and it overestimated the route by 1km.

These measurements are very valuable to me and it is not acceptable for them to be inaccurate in this way. I will not let my work be wasted by a few inexperienced developers

Tomorrow i will try to same settings with 16.19.

I have a suspicion that it needs to put non-standard settings for running.
Check out my activities.
Running was again inaccurate. Again it "flew off" at the very beginning. There were settings - All+multiband / recording every second

https://connect.garmin.com/modern/activity/12866384871

Later I recorded Walking in the same place. but with different settings. Everything was fine. Settings - Auto select / smart recording

https://connect.garmin.com/modern/activity/12872914283

Both without 3d speed / distance. I think I have to try alternative settings and check it out what will happen with distance.

By the way - when it happens, there is a dip in the endurance graph.

Looks like a very nice route :) to run!!!

Overall the tracking looks ok, there might probably be som minor issue around 1 hour and 9-11 minutes but it is not possible to see with without the actual gps-files. There are places where the track is slightly off compred to the map..but that is quite normal - it happens everywhere, and I can easily be related to the orginal google map recordings or the weather.

Personally I would still change the setting so it is not Multiband but GPS Only.

Secondly I would let it soak for 2-4 minutes before i press the start button (I know it turns green when ready, but let it stay green for a few mins before pressing START) =  Then you are sure that the IPO file is fully updated.  

I think I found the problem.
It's not the GPS.
The problem is with the hub sensor and the calculation.
During running, as the speed increases, GPS accuracy increases and as the speed decreases, GPS accuracy decreases.
So if you run slower, the GPS records a longer track and kilometers.
If you run fast, we record a shorter track record and kilometers.
I think part of the reason for this is the stride length setting.
So here we have por
It's your speed, your stride length, the hub sensor's perception of arm movement and the average of these.

I recorded the 3 records I put as an example before when I was running with my wife. I mean, they were slow runs and the GPS recorded almost 1KM more.
Today I ran at my own pace with the same settings and this is the result.
Same route, same settings
If anyone has a different opinion, please write

Hi

Its not that I don't believe you but that sounds very veird :) Have you tried to compare the two gps files with dcrainmakers analyzer tool??? (https://www.dcrainmaker.com/analyzer) 

I think you need to compare the two raw files just to rule out that it is not garmin connect the problems arises.

I think I should apply for a job at Garmin and investigate this problem there. :)