fenix 6 pro run distance accuracy ??

I have a Fenix 6 Sapphire, which is like the PRO, and as I previously wrote on another post, it’s a great watch, but it fails on one thing: it’s one of the worse GPS watch I’ve been using since 2010. The tracks are horrible, but I don’t care much about it. I care about distance measuring, because without a correct measurement all the other metrics are useless. And don’t take me wrong, I’m not looking for a perfect measurement. (Mr. moderator, I know all the problems related to GPS signals and so on. I’m an airline captain and I’ve been using GPS’s for 25 years, therefore no need to explain me how does it work) Said that, I usually run with a Polar Grit X or a Polar V800 on my right wrist and my daily driver AKA Fenix 6 on the left. Somehow on my favorite 10km run (measured with several devices to be exactly 10.150km) it looks like the Polars are always within 20 to 30 meters off the actual distance, which is a great level of accuracy for a GPS watch, while my Fenix 6 most of the times is off by 100 meters (~1%) Same story on 1/2 Marathons where my Fenix 6 is usually off by 200 meters. (Still ~1%)

Obviously most of the people would be happy with that, and the Garmin aficionados will start writing that 1% is a great level of accuracy for a watch. I’m ok with that, it is a great level of accuracy for a watch, but I would like to know why the polars have a much better accuracy then... 

One more thing... I found out that if I download the run from Garmin Connect and then I upload the very same file to Garmin Connect, the distance will change to a much more accurate value. This is telling me that the algorithm used by Fenix 6 to convert coordinates into distance is not the same that is used by Garmin Connect, which means that in theory the GPS chip of the Fenix 6 is measuring a fairly accurate distance but then the watch itself translates the data into a wrong metric. 

thanks so much - so I am not nuts? - but 200 meters over a long distance adds up - that is my issue - !! - this was off by about a mile for a 1/2 marathon I did - now, it was on trails - but fairly open trails in North East Ohio - not mountains - and I trust this event to have an accurate course - I ran with the Fenix and my 935 - I seem to trust the 935 more which bothers me because it is not marketed as the premier watch the Fenix is - and I paid much more for it - 

again, thanks for your remarks - 

Thanks for your reply and for following up this issue. I have some very fresh data to share with you. I’m just arrived from a nice run in the woods, it’s a nice 1/2 Marathon course. 
Guess what, my Fenix 6 measured 20.83km, which is less than the usual 1% off, but I just downloaded the file thru Garmin Connect and I uploaded it immediately to the same account and the distance shown is 21.16km, which is very close to the actual distance or who knows, may be it’s the real distance I run due to cornering. Anyhow, once again I think that it’s a software problem. Try to do the same, try to run a course that is familiar to you, at least 10km, then download the file and upload it. You will see it’s different and most probably the uploaded file will show an almost correct measurement. 
Let me know :-)

can you explain how you download first - unlink from the phone, then plug into computer and synch that way - ? - this process has become so automated I am not understanding....:-)

When I arrive at home, after my run, I sync the Fenix 6 with Garmin Connect Mobile, on the iPhone, in this way the system gets the data.

Then I turn on my Mac and I log on Garmin Connect, the website and since I already synced the watch with GCM my last run is right there, on my account. At this point I open the activity and I see the map and all the other metrics. On the top right corner you should see a symbol that looks like a gear. Click on the gear and a drop down menu will show Export to TCX or Export to GPX and so on...

I choose Export to GPX and I save the file on my downloads folder (in the Mac) then on the top of the screen of Garmin Connect, to the left of the user’s icon there is a symbol that looks like a cloud with an arrow pointing up. Click o that symbol and it shows “Import Data” and then if you click it shows a field where it’s written “Drop files here or browse to select” I select my download folder where I saved the GPX file and within a few seconds I see the very same run that has been duplicated, but the funny thing is that the duplicated run shows a distance that is ALWAYS very close to the actual distance.

I tried your export/import suggestion and the result is rather good. But its not a solution to our problems with the measured distance.

But hopefully Garmin can use this info in their troubleshooting.

I did the test on my 23.965 km run yesterday (distance by odometer). My Fenix 6x Pro Solar Titanium measured the distance to 23.532km which is -1.81% too short. My settings where 3D Disabled and GPS+GLONASS and I had the watch on the inner left wrist. After the export/import workaround the distance was 23.72km which is -1.0% too short.

But the question is why do we get smaller deviations when exporting and importing the activities than the original reported distance? I think that the smoothering/averaging of the data is the problem. I see deviations in the .FIT files regarding the logged m/s and the distance calculated by the GPS semicircles.

Im guessing now but i think that the import function in the Garmin Connect Web uses the GPS semicircles to calculate the distance. Without using the m/s data which is logged in the original .FIT files.

On trails, depending on how many switchbacks and turns there are, the Fenix will under-report.  I've see this on the 3hr, 5x+ and 6xPro Solar.  Stryd remedies that on foot and the Garmin speed sensor remedies it on the bike.  You can argue that you shouldn't have to invest another 150-250 to get reliable distance but it works and then the features + distance accuracy makes it better than any watch you can buy by far IMHO.

i was wondering about that - I suppose for the money saved not getting the solar or saphire Pro I could put toward the pod you suggest - thank you