Fenix 6 PRO GPS accuracy

You use the same settings as me. What exactly are you complaining about? From the measured distance, pace or from the map image? Which activities are you unhappy with running, walking or otherwise? Have you tried to run on a completely open terrain with a nice GPS soak and then check the data?

Only Garmin can answer your first question but remember that Garmin is all about GPS. They do it 24/7 for outdoor, marine, aviation and they do it for both consumers (like we), business to business and military. If it would be as simple as a software "fix" I'm quite sure they would have released it. I remember the exact same discussions when Fenix 3 was released. Some things got a bit better the first year just as it did with F6 but after that no drastic changes.

I use the same settings as you, test GPS+Galileo from time to time. For me it is ok except for instant pace. Bought a Stryd to get more accurate pace and found out that my F6 distance is almost spot on compared to a calibrated Stryd. In open areas the F6 distance is a bit to long and on trails a bit to short but still very close.

Why do the accuracy differ from one to another?

First of all GPS, GLONASS and Galileo has it's glitches, good days and bad days. It also differs a lot depending on where on the earth you are.
Secondly watches differ du to assembling issues, just like any other piece of hardware. Unfortunately that impacts the accuracy.
Finally, make sure you use a good map source. I've posted examples of this before and here is an example of yesterday's lunch walk:

According to Strava (Mapbox) I walked through the house but according to Garmin Connect (Google) it was pretty much spot on. Is Google always better? No but at this specific spot it looks better...

Well, I have been through this too. The thing is that Fenix6 is known to have a bad accuracy. I am on my second Fenix 6 and there is no difference.  You just have to suck it up or buy another watch. 

The only REAL thing you can do to improve accuracy is to turn your arm to that the clock face is always facing the sky. This will improve the GPS reception.

Hi,

My experience is about the same.

Accuracy I get with Forerunner 230 or 235 in the right wrist for the same run is much better. Also my wife with 245 improves by far the accuracy of Fenix 6 PRO when we run/trail together.

I would not say Fenix 6  is a "disaster", because the total distance is basically ok, but the lack of accuracy along the way makes that the instant pace (or 30 sec rolling pace) is completely inaccurate and useless. Specially (but not only) when running in the city even in open big streets where 230/235/245 tracks quite well with minor deviations.

No matter the configuration (GPS/Glonass/Galileo) or other settings. Always wrong side of the street, and inaccuracies of 30-50 meters. Good enough to see in a map your track, but useless for specific training workouts (200,400,800,1000m bouts)

Most of the times, besides some inacuraccies, when I run along a certain path, and come back by the same path, I get 2 "parallel" lines 30 meters or so from each other ... I do not know if this is common for any of you. 

I have tried something that I read here is this forum.

I put the watch on the narrow side of the wrist thus the watch is always pointed upwards to the sky. The accuracy has increased tremendously.

Good point. Yes, it actually did in my test yesterday too, reducing/cancelling  the offset. (I also ran segments in the usual position to confirm the offset came back again)

If it is about the antenna angle vs sky, can it be corrected by software or should we live with it for ever?

One issue that a very few people bring is discrepancy between the distance you see on the watch and the distance along the GPS track logged by the watch. These two distances aren't the same. In my experience, when running on trails, the reported distance is generally 5% shorter than the distance on the track. I'd be quite satisfied with the latter because in most cases that is much closer to the actual distance (although still a bit short when compared to Suunto Ambit3 Peak for example).

I understand that Garmin seems to be doing some sort of filtering / smoothing of distance to counteract wobbling and avoid overshooting the distance when running on roads. However trails are often naturally quite wobbly / wavy, and the excessive filtering always leads to under-reporting the distance. One would think, that being an outdoor watch, Fenix should be optimized for the trails, at least when recording a trail running activity. 

silentvoyager said:

One issue that a very few people bring is discrepancy between the distance you see on the watch and the distance along the GPS track logged by the watch. These two distances aren't the same. In my experience, when running on trails, the reported distance is generally 5% shorter than the distance on the track.

A Stryd foot pod is considered one of the most accurate out there and last summer I decided to buy one. After calibration I agree that distance on trails are generally to short but still close (not 5%). For road and jogging tracks I tend to get shorter distances using Stryd than GPS. 
I personally favor precision over accuracy since that will make runs on the same route comparable and I had better precision without Stryd. The main reason why I still use the Stryd is accurate pace.

It depends on trails. For example, many trails that I ran in California are more straight than most trails in Washington where trails tend to be quite twisty. 

Compared to official wheel measured race distances the difference here is typically more than 5%, although I should mention that pretty much all GPS watches, even the most accurate, tend to be short on trails in this area. For example an official 1/4 marathon course in a local watershed preserve was measured by F6X as 6.12 miles. That is 6.5% short, and those trails are not the most twisty by local standards.

5% is the typical difference between the reported distance and the distance on the logged track (the track logged by the same watch). If I do distance correction in Strava the distance typically increases by about 5%. For example, today when running in local mountains my F6X reported 16.86 miles and after distance correction on Strava it became 17.65 miles. That is 0.79 mile difference or just under 5%. Again both results were produced by the same watch.

I have tried a couple of days with the watch in my right wrist,  and results consistently improve, reducing or even cancelling the offset, getting a much better accuracy. I change wrist on the go to compare, so conditions (weather, satellites, buildings,...) are identical, but results are clearly different.

Where can be the logic behind? In both cases the "12" is pointing downward. Any idea?