Shift in GPS track depending on wrist

As an avid heatmapper / run every street I'm particularly interested in more accurate GPS tracks.  I've been perplexed by why has appeared to be a shift of my GPS tracks to the left of where I run no matter what direction I run in.  So if I run an out and back on a road it looks like I've practically run a loop because its shifts it off the road on both sides.  In general my tracks are off the road and sometimes its more egregious than others.  I took to running on the wrong side of the road (with traffic instead of against) to get it more centered.  I then did an experiment and changed my watch to my right hand.  Low and behold my tracks started looking significantly better!  I figured there must be some sort of setting I was missing.  I found the wrist setting in my Fenix 6 and changed it from "left" (the correct setting) to "right" and it made no difference.  

Wearing the watch on my right wrist example here:  - I ran on the side of the road against traffic and its shifted to the right, which is closer to the middle of the road

Wearing the watch on my left wrist example: - notice how all the tracks are to the left where you would be running against traffic off the road

Really neither of these examples are very accurate as they don't capture where I really ran but at least with wearing it on the right hand it will actually show on the road I was running on.   I don't know if this is an intentional shift or not.  Also I'm not a fan of wearing it on my right hand (again wrist setting seem to make no difference in my tests).  Does anyone have any info on this?  

Also this isnt a one off test. All my runs with my Fenix 6 are like this and I found many with my Fenix 5 as well.  Additionally I have done several runs with it on my right wrist and they consistently produce better results every time.

Top Replies

All Replies

  • Another thing, (there was a really good presentation at the CIQ Developer Summit last year about this, but it wasn't recorded - I asked afterwards).

    Think of what happens when you run.  Your arm is moving in a circular motion, so while you are moving forward, your arm could be moving back (or just up/down) when the location is grabbed.  Add into that that the accuracy of the GPS may be a few feet (forward or backward), and the next second could overlap the one before as far as accuracy.  There's a lot that goes on here.

  • 1-second recording is pretty much the first setting I change to whenever I get a new Garmin device.  I use the Stryd footpod for pace and distance as its more accurate than GPS.  I found GPS to be notoriously bad in the trails and on the track the footpod handles those situations.

  • @ Punch - you don't seem to have the slightest idea about the number of GPS + GLONASS or Galileo satellites normally in view vs. sats needed for precise positioning so you post some bs here. Please, if nothing else, use a smartphone (ideally, also Galileo/Glonass enabled) + an app to find out.

  • you don't seem to have the slightest idea about the number of GPS + GLONASS or Galileo satellites normally in view vs. sats needed for precise positioning



    The more satellites there are above the horizon the more accurately your GPS unit can determine where you are.


    i can post more links if you want Joy 

    with your logic: why do we have all these satellites if we would do just well with a few?Joy

    We should ask to take them down honestly ..just a waste Sweat smile

  • Watch on the left wrist in both directions. Doesn't prove anything other than the shift is the same whichever way I go. 

    IMVHO you're making a mountain out of a molehill. For all the reasons that have been explained here and the limitations of a wrist-worn GPS, the tracks you have posted from January are acceptable.

    And people, stop shooting the messenger. Just because you don't understand, or don't want to understand, there's no need to be offensive. I know it's difficult, I've been there. But I like to think I've matured a bit...most of the time.

  • with your logic: why do we have all these satellites if we would do just well with a few?

    It seems that you think that they are all up there just for accuracy. Well, try to answer your own question and you might learn something useful. A few pointers, which you could google: time to first fix, coverage, redundancy.

    Every GNSS system is designed so that there are at least 4 satellites visible at any point on the earth at all times (if I remember correctly, visible in this context means at least 15 degrees over the horizon). So, if we use GPS + Glonass / + Galileo combo in Fenix, that would mean a minimum of 8 satellites visible (4 from each system) at any time anywhere, which is plenty for precise positioning. As to the typical actual number visible/usable I already advised you to use a smartphone/app to find out.

    As to your "the more, the merrier" statement, consider this: if the GNSS system is designed to provide, say +/- 5 m accuracy, can this accuracy still be improved and improved indefinitely (receiver side) just by the receiver "seeing" more and more satellites? You were willing to post links. So how about this: what is the actual accuracy increase (m or %, if you please) between using, say 6 satellites vs 8, 6 vs 10, 6 vs 12, 8 vs 16 etc.? Is this accuracy increase linear (0.5 m or 10% per each additional satellite, for example) or some kind of a curve? What is the number of satellites above which the accuracy no longer increases (if there is such a number)?

    Sorry to the thread readers for this off-topic. Back to the topic - my tracks are also almost always off-set and pretty imprecise (1 sec recording) even in wide open areas. I have not tried to switch wrists to see if that improves anything.

  • Background - I have found the same with my 645,  a persistent ~4-5m offset to the left while running.  (I have some posts and pics in the 645 thread re exactly this issue) When stopping it will 're-locate' to the real position, only to continue with the offset when continuing the run.  I have also found that if I switch wrists it 'seems' to be more correct, as the offset will now be to the other side.  (I am of the opinion it is an antenna / receiver issue) 

    However glad to say that I have not encountered this with my F6s.  

    So no solution for you unfortunately, more of a note to support your findings ito the offset and switching of wrists.

    (Running on the RHS of the road here, both tracks offset to the left.  NB :This was with my 645 and NOT with my F6s, which has very accurate tracking :) ) 

    @Dallasrpi - it perhaps possible to test this while the watch is 'mounted' or carried with face upward?  Reason for this is that my 645 tracking was very much improved (putting me on the right side of the road and much better overall tracking, even when riding at my running pace) when it was mounted on my bicycle's handlebar, or when I ran with it in my hand facing up when investigating this phenomenon.

    (PS : I though the F6 series have an 'exo GPS' antenna which will largely prevent these kind of issues? )

  • Just to be clear - you're expecting a device approximately one inch in diameter with a tiny battery and limited processing power to consistently calculate your moving position to within about 10 feet? Electronics do not work on magic; it's not as if there's a hidden control inside the watch that Garmin deliberately turns down to reduce your accuracy.

    Those tracks look absolutely fine to me. If you want more accurate recording, stop every 15 seconds and stand still.

  • There is clearly smtg wrong with hardware or software (I hope can be fixed) for some units, these are the tracks from F6 (1sec) and Forerunner 230 (smart) I made. The old one doesnt seem to ba affected by threes, houses, nothing..

  • Electronics do not work on magic

    Algorithms sometimes can do something similar than magic. Look what is happening in photocameras into smartphones, for examples in Google Pixel device. 

    Anyway, this is not the first time somebody is reporting this behaviour and I think would be better to investigate (also with the help of people than in any discussion minimize any issue) to understand in which conditions there is this offset because if it is similar everytime in distance and in conditions could be possible to adjust the algorithm with an offset to compensate the issue.