GPS Accuracy

Between buildings, all GPS watches will struggle. I wouldn't use this to compare to your other run. And more generally, the most recent watches are not better for GPS tracking. Looking at forums for Garmin watches or Suunto, i see that the recent watches are often doing less good than the good old ones. They won a great battery life with the Sony chip, but accuracy is often a bit less good. I saw that when I switched from my Ambit 3 to my Fenix 5. If I was running on a track often, it would be a problem for interval training. For my other runs, is that a real problem ? No. The distance is often quite fine (even if Garmin is often compensating too much for GPS errors, while other brands can give distances longer than the actual run). Now I have maps, great battery life, notifications and more. The compromise seems very acceptable in my case. I see many complaining a lot about it, when actually they are not doing interval training. How serious of an issue is it if the trace is some meters away from the real one? It's a 80gr device that does so many other things. I think it's important to keep that in mind. We often get disappointed for something that actually doesn't change anything in our training, or results. 

Sorry for the long answer, I'm ranting a bit

I am just concerned that this grave inaccuracy means the numbers for my run are also totally useless. I'd sacrifice a stack of the features for accuracy. 

Thanks for that. I've seen similar in short laps in urban environments. That I think is simply a factor of GPS. It's always worth trying the different combinations GPS, GPS+GLONASS, GPS+galileo on the same route to see if there is a difference. For me there is not. FWIW it's worth I did this run this morning with an F6 on one wrist and a Marq on the other. Can't remember which is which now but one of them has me trotting through houses in places. My point is that it all depends on what's around at the time. They're not too far apart though for any concerns overall.

Thanks, I am just miffed as I was doing a training workout and started pouring over the numbers... then spotted how bad the gps map was :( 

I think you are right, it is a matter of living with it. I think I am just noticing it more on repeated loops and roads. On a single lap or route it won't be as obvious. That said the lines through the buildings are disappointing especially for per second GPS setting. 

Hi, David, yes you must live with it. A month ago i had 20-days e-mail communication with garmin support about the issue. I had the same problem and comparing with my previous fenix chronos, the old one was much better in GPS tracking (also a distance measurement). I made many tests. On some group competitions, my watch accuracy was terrible comparing to other competitors tracks. The good think was that after my complain, garmin changed my watch. The bad think was... it didnt fix the issue. I can give you some advices, that can lower the impact of your problem (even if they sounds really stupid). First. Every time before you start tracking wait at least 5 minutes even if your watch is with green gps ready. Second try to track with the watch on your right hand (looking at your track im pretty sure you are wearing it on your left hand. In my case the problem is that the track is 15-20 meters to the left, when i wear my watch on the left hand). Because of this running counterclockwise is making your tracks shorter, and running clockwise is making your tracks longer. My watch is much more accurate on track, when i wear it on my right hand. (Strange, but true). I have a feeling, that the gps accuracy is affected by my own body. Now im living with it and i enjoy the other extras (Design, paying, music, durability)

I think that Garmin thinks that this a special feature than poor quality.

Yes, I have the same behaviour with the fenix 6x sapphire. I had/have the same bahaviour with the fenix 5x.

There are multiple incredible behaviours. The devices always record a track which is inside the actually run path. Always. It seems that Garmin records some seconds later and interpolate something. I tried different methods to record (GPS alone, GPS+GLONAS, GPS+Galileo, 1s, intelligent, ...)

Evaluating gpx-tracks lead to a different result than the distance shown at the end! More determined out of the gpx-track! But getting a result at a quarter marathon race of 10,26km instead of 10,55km is rediculous. That means 2,8% less! That's in fact impossible. calculating the Distance by using the recorded gpx-track ist would be

typical problem: Garmin seems to clairvoyant abilities. The following blue line in the picture was the path within the fences at the race! The red line was recorded by Garmin fenix 6x sapphire! Garmin always cuts the corners. With my Polar V800 it mostly overestimates at changes of direction, but corrects very soon. It is impossible that the Garmin-device knows where I am going to run before. That means Garmin interpolates because of bad accuracy after determining the locations and records some seconds later.

Yes, I agree that improvemnet has to be done by Garmin.

Example: (& if you like to determin the track-length out of the gps-file by yourself the according vba-code)

dist = 6378.388 * acos(sin(lat1) * sin(lat2) + cos(lat1) * cos(lat2) * cos(lon2 - lon1)) with lat, lon in radians

In Excel you can add the user-funtion (in vba) and determin the differential distances (in km) as follows:
Function Dist(lat1, lat2, lon1, lon2)
' distance:  in km
' lat, lon in degrees (not radians!)
lat = (lat1 + lat2) / 2 * 0.01745329 '0.17.. = pi/180!
dx = 111.32387 * Cos(lat) * (lon1 - lon2) ' 111.3 = 6378,388/180*Pi; degrees!
dy = 111.32387 * (lat1 - lat2)
'lat1, lat2, lon1, lon2: in degrees 
Dist = Sqr(dx * dx + dy * dy)
End Function



that means for miles the value of 111.32387 has to be divided by 1,609: 69.18823 

I have the same problem. I run along the same route with another garmin much much cheaper and with my phone S10. When running with my Fenix 6 Sapphire  the difference is about 20- 40m in every kilometer. That's awful. What can i do except "wait at least 5 minutes", because it doesn't help much.

There is a site for planning GNSS measurements, www.gnssplanningonline.com
1) go to the site, mark your location, height, date, do not forget the time zone and set 40 degrees of sky visibility - this is the most for urban building, after that apply the settings.
2) go to the charts and look at two charts: the number of satellites and DOPs (the lower the DOPs the better, we need a yellow-horizontal for comparison, the red field is understandable, bad constellation), and on the left side click on checkbox, choose the GPS only, GPS and GALILEO, GPS and GLONASS, and compare values of DOPs, obtained with such combinations of GNSS system.

I hope, that the Garmin add GPS+Beidou in the future, becourse for my rejion it will be optimal

This is not the solution. The cheaper Germin has better measurement than F6.