No announcement yet.

GPS Accuracy

  • Time
  • Show
Clear All
new posts



    • 3 Years and 10,000 Miles

      So the GPS Accuracy post has been un-stickied and hasn't been contributed to for quite some time. I suppose that's indicitive of that fact that GPS 'inaccuracy' concerns have been mitigated and it's become a non-issue. I know that for me, GPS precision and pacing performance have been outstanding since the beginning.

      Maybe I exaggerate slightly.
      In three years my mileage is at 9775.

      No exaggeration needed in describing GPS Accuracy. I've got a solid mix of road, trail and track. I've got a mix of fast and slow. Mostly slow these days! The fenix 3 has measured up against any previous Garmin model and has more than delivered on features and functionality. I'm fairly demanding when it comes to evaluating performance. I train in cycles and when I'm in a more serious period I expect to be able to hit segments precisely and accurately. The fenix 3 consistently hits marks within a few steps. Every time.

      I don't have any magic spells or secret sauces. I guess letting my watch get a solid signal acquisition is the recommened best practice behavior that I follow regularly. I advocate and use Smart data recording. I'm using UltraTrac almost exclusively in the trails these days. That's a feature that I've really come to embrace.

      Do I have suggested enhancements that I would like to see imlemented. Of course. That's the beauty of technology and the progression of the product line. I just don't have any GPS performance concerns and/or requested improvements in that regard. For me, the fenix 3 has more than met my expectations and as described - I expect a lot!

      It would be a far more interesting post if I could duplicate issues or user concerns that frequent these forums. I can not. I'm in the camp that suggests satisfied owners spend their time enjoying their devices and concentrating on training and results. Carry on.


      • Great. I guess your pleasure comes from Smart data recording. Maybe I will switch to it from Every second data recording. So far I did not feel like doing it, because I had a fear that my peak power data would be rubbish if not having every second data.

        Please educate me! If i opt for Smart data recording, but I use sensors including a power meter, how my power, cadence, HR data are recorded?

        1. At every second, while GPS data are not?

        2. At the same timestamps when GPS data, and the latter defines when to record, when not?

        3. Each data types have their own smart data algorithm, so eg at timestamp 1000 only HR was recorded, because it had changed vs that at timestamp 999 and at timestamp 1001 only those data are recorded which had changed vs timestamp 1000 etc

        You expect a lot, I hope you are a data miner, too.



        • I'm only a runner. Don't bike or swim so don't have any historical data to reference there. I don't use HR at all as all of my training plans are pace based and driven by GPS. I think you might be trying to over complicate things. If your HR changes, then a data record should be captured regardless of your GPS position. Changes in values should trigger a data write and all fields would be populated.

          If you're asking if I do a lot of detailed data analysis, then yes, I do my homework and spend a lot of time evaluating results. I wouldn't dismiss Smart recording if we're having a discussion about GPS accuracy (the topic of this thread). Can't offer any knowledge on how that responds to a power meter. My guess would be that you wouldn't be missing out on any conclusive statistics.


          • I had a funny moment during today's run GNSS completely messed up 1/3 of my run
            First 10 km were fine without any issues. The issue started on the eastest part of the map near an old pump station. Based on my heart rate my tempo could be around 4:20, but from unknown reason GNSS messed it up completely and from 500 m long path was 1 km long. I have followed the road through village Mierovo to main road 503 with slightly increased tempo. Based on map my position was completely messed up, F3 showed tempo around 6:00 (my real tempo was ~4:10). On the main road on 14th km I gave up, paused my watch and then enabled it again. Probably this disabled GNSS completely and speed of the rest of my run 'back to start was only accelerometer based.


            • I have been closely following the GPS accuracy of my watch since my last post in this thread, moreover I still occassionally use my 910XT on my other wrist, and still feel that the accuracy of Fenix 3 (HR) is inferior of the plastic triathlon watches like 310XT and 910XT (NB: i dont have the newer models like 920XT or 735XT).

              Since I use a footpood for me not the speed is the problem, but the total distance calculated automatically from the length of the tracklog by my favorite activity recording software. I improved the accuracy a bit with the trick of applying autopause at lower speeds, but it is not something which took me to the feeling of Heureka!

              On the other hand having seen the complaining users of F5 and F5+ families about the GPS accuracy of their watches, having faced the ugliness of those tracklogs, I am a bit happy that I bought a device which already knows the smart notifications function, the steps made tracking, and its GPS accuracy were just partly ruined by the developers/designers of Garmin. Let me call it as a sort of compromise.

              The accuracy of sleep tracking is a joke, optical HR (I have F3HR) is only for a feeling that I have something to be watched when I am weightlifting, I mean it is just a gadget feature, not something one can rely on.

              So again F3 family is a sort of (maybe good) compromise.

              My advice for anyone being serious with GPS and distance accuracy and who uses a Fenix of 3 or above: enjoy the smart functions, but get a pre-owned 910XT, too, for USD 100 or less, and finally blend the data of the watches in a clever way.
              Last edited by Tisztul_A_Visztula; 11-03-2018, 03:52 AM.


              • What's the latest consensus about the accuracy of this watch as a pure running\training device compared to the market standard as of November 2018 with the latest firmware? I have been doing some research and it seems that many people dog on this watch for the GPS being inaccurate (at least with a lot of the firmware that was used in 2015 and 2016), but there are probably 50x as many of these devices out there than of the lesser-known brands\models of gps\fitness trackers, so it would make sense that there are more complaints in pure volume for this device.

                I'm currently using a 7 year old Globalsat GH625-XT that has fantastic accuracy (using SIRFSTAR IV board), but the watch is pretty beat up and I can no longer get replacement wristbands for it so I'm forced to hand carry the monstrous beast until a suitable replacement is found.

                I found a good deal on the Fenix 3 HR, but am very concerned about whether or not the GPS accuracy will render this useless to me as a very serious runner who cares about the difference between a 5:40 mile pace vs the same run showing up as a 5:55 pace. I also do a lot of woods running and this might not even work for me at all if some slight tree cover is going to make it lose satellite signal or a 3D fix.

                Is anyone using the "Dozen Run" app on their Fenix 3 HR? I've read in a few reviews that this app lets you see your real-time pace in finer granularity rather than the rounded 5-second tiers. I realize that this app can't possibly account for general inaccuracy of the hardware, but it may overcome some of the limitations of Garmin's native tracking apps.



                • Originally posted by partyman66 View Post
                  I'm currently using a 7 year old Globalsat GH625-XT that has fantastic accuracy (using SIRFSTAR IV board), but the watch is pretty beat up and I can no longer get replacement wristbands for it so I'm forced to hand carry the monstrous beast until a suitable replacement is found.
                  If this is your standard, forget about GPS in a newer Garmin watch. It will not satisfy you. Combine the Garmin watch with a Stryd foot pod or buy another brand of watch.

                  The Fenix 3HR uses a MediaTek chipset, and most watches with MediaTek are worse than most watches with SirfStar. Look at this list which paints a pretty clear picture:

                  I have not used the F3HR, but I have used the F3 and now the F5X+. They cannot give you a consistent immediate pace, even if their total distance is pretty good. On my F3, pace would be all over the place as soon as I ran along a row of trees. On my F5X+ I see a more stable pace, but it slows down along those trees. And we are not talking 15 seconds per mile here - the pace on my watches will be off by 1-2 minutes/km when this happens.
                  Last edited by AllanOlesen67; 11-06-2018, 02:21 AM.


                  • Allan’s words are fair, I agree. Those users like you or me need an old watch for high accuracy and a new one for the other features like communication with phones or simple watch function


                    • Just to (re) inject some sanity:

                      1) It's not the chipset, it's the system design. Antenna design more than anything, but antenna, layout, chipset, and housing.
                      2) The Fenix 3 is not the most accurate watch on the market. It's far from the least accurate. The spread isn't big.
                      3) You won't care if you don't dig in deep looking for tiny discrepancies. You will care if you do. This is true regardless of which watch you buy.
                      4) The 310xt and 910xt are not better than newer watches; they're just as hit and miss (
                      5) If you try to evaluate whether the 310xt is more accurate than the Fenix 5X Plus Bankruptcy Edition or whatever, you'll probably find some runs that confirm your bias and conclude it is.
                      6) The 310xt and especially 910xt are EXCELLENT values, especially today, and well worth buying if they satisfy your needs.
                      7) They are, ultimately, a bit of a pain without bluetooth or wifi.

                      Ultimately, you should probably just buy whatever watch you think is going to satisfy your needs for a price you like. If the GPS accuracy doesn't seem to be up to snuff, return it, and never buy any other GPS watch, ever.

                      Edit: Oh, and I'll add - GPS performance aside, I've been increasingly impressed with Garmin's overall product offering with each new watch. The Fenix 3 treated me well, and Connect IQ was better than the competetion. The 935 is a FANTASTIC watch, and the latest versions of Connect IQ on the more capable hardware absolutely blow the competition out of the water. All the metrics provided by Firstbeat work out to be shockingly both accurate and helpful, including VO2max, LTHR, training status, and so on. Sleep tracking is meh, but that's somewhat true across the board, and I hate wearing the watch to bed anyway. I'd absolutely LOVE to get the latest Fenix 5 with wrist pulse oximetry, but I don't print money and climbing mountains isn't my job anyway.
                      Last edited by alexwhittemore; 11-07-2018, 06:43 PM.


                      • To piggyback on what Alex says, I've never bought into the argument that the Mediatek chipset is inferior to Sirfstar. Every device forum seems to blame all GPS woes on the Mediatek chipset. That seems like nonsense. I've never really seen any conclusive evidence to support that theory. The only deep digging I can do is on my own data though. I've done a lot of comparative testing between the F3 (Mediatek) and the 610 (Sirfstar). The 610 is an older generation device that has been widely heralded as having excellent GPS performance. The F3s performance is on an equal level on every account regarding GPS accuracy and pacing. It's an amazing training tool and to suggest that you need an older watch to do serious training is silly. If I didn't know that each device had a different GPS chipset I'd never be able to tell the difference by examining the results. They're consistently equal.

                        The F3 does seem to have a more sensitive antenna. It records data points at roughly a two to one level (smart) during the same activity. That might make the track seem slightly more noisy but has no influence on the actual performance. Both are outstanding.

                        I'm not a competitive athlete (well, only with myself) but I do follow a lot of structured workouts and do some semi-serious training. The F3 is always up to the task even when I'm not. I've tested nearly every Garmin in the lineup over the last 12 years and have never been disappointed. Have there been days when the F3 (and every other unit) has had a hiccup? Absolutely. The real value is the consistent results I get over time and not the occasional burp.

                        As a trail runner, I do a lot of running under heavy tree cover and I've never encountered a tree that impacted performance. Seriously.

                        Do I understand the barometric altimeter? Nope. Ha. I've tried. I've really tried. It just doesn't influence any data that I care about as a runner and I use elevation corrections anyway so it's just a curiosity thing for me.

                        I also had a test period with the F5 and it was equally remarkable. I'd upgrade in a heartbeat if my pockets supported it. While I still consider the 610 a monster the technology is dated and the current generation is hands down more functional overall. I've got pros and cons with each but again, GPS considerations are a wash since the performance is exceptional across the board.
                        Last edited by deeter26_2; 11-08-2018, 02:46 PM. Reason: 12 years, not 20!


                        • I see that there are 209 pages in this thread and honestly I don't have time to read them all. However, I have been using a 235 for years and I have recently bought a 645M and I am now using both together when I run to compare the numbers. I have been very surprised to find out that the 645M reports significantly shorter distances than the 235 for the same time interval, for example today the 235 measured 26m11sec for 5km while the 645 for the same run measured 27m07sec, with a difference of around 150m. I would have expected a small difference, but this looks excessive. Which one of the two watches is usually more precise, which one should I trust more?