Here's my first activity. The only places were it wandered significantly were next to a large metal object, the Angel Of The North sculpture (a small wander) and past two tower blocks (where the track briefly crossed to the other side of the road).
Looking at the same route from my F3, the F5 looks a little better. The F3 struggled in the same places but also had issues elsewhere. Measuring this route on Google Earth, I would say the F5 has it in terms of distance accuracy.
I know a sample of one doesn't mean much, but it seems ok so far in my opinion. Not the best, but ok.
F5 definitely is better than F3. F5 is significantly better than when F3 was released.
Maybe this is also input to the fine-tuning process Garmin needs (I guess) from real life testing, to fix it in future firmwares. All new GPS watches the last 2 years seemed to need a burn-in period just after release.
Will you not see this type of anomaly with all GPS watches? I think expecting 100% accuracy is a little unrealistic. I checked my last vivoactive hr run for example and I don't remember running through walls but my tracks show very similar deviations to what you have here.
The issue is that the GPS is worse than other GPS watches.
However inaccurate GPS watches may be by nature of being a watch.... the Fenix 3, and now it seems the Fenix 5, are worse than other GPS watches when it comes to GPS accuracy.
Considering they are pretty much the most expensive consumer GPS watches on the market; I think it's fair to expect them to at leat be as accurate as an £80 Forerunner 10 for example (which literally costs 1/10 of the cost); which sadly, is not the case.
What's confusing is that they can't to a better job mitigating the multipath effect as seen on the screen caps above when they do it perfectly (given the constraints of a consumer device with a tiny "antenna") on the FR235. Without a raw/NMEA log there's no way to know what they're working with but it must be really degraded for the output to be like that.
When viewing the data on a website, you may want to make sure "Data Recording" (under "System" on the watch) is set to "every second" and not "smart" (smart is the default), as with "smart" the location isn't saved as often, and can make things look a bit "choppier".
I've done a bit more reading on this as I have never really considered it a problem on any gps watch I've had but clearly there are some that are championing for higher accuracy which I am all in favour of. There seems to be lots of differing opinions though. In this article for example, the Fenix 3 is rated among the best for GPS accuracy https://the5krunner.com/2017/02/10/garmin-fenix-3-vs-suunto-spartan-ultra-gps-test-gps-face-off/
I've done a comparison run this morning, recorded on the F5 (with internal WHR), F3 (with HRM Run) and Strava iPhone App (with Scosche Rhythm+ OHR). The setup and end distance results on devices can be seen in the pictures below and the comparison of tracks brought together at the mygpsfiles link below. Just please be aware that all satellite imagery out there on the web does not yet capture the new tram route and associated paths (in case you wonder why I'm going through fields in places or the paths seem in totally the wrong place) and I was running with my dog and did have to stop to pick up 'the obvious' on occasion (you can clearly see where I had to stop on one occasion, as the F5 track moves to the correct side of a street at that point - it's not a random jump)
In the open all 3 devices tracked fine. There are places that each of the devices struggle though. Through the trees up to the bridge near the start/finish of the run the F3 and F5 definitely tracked better than the iPhone did (first zoom in shot), but in the trees around the river area mid run the iPhone had the edge as the F3 and F5 cut the corners a little (second zoom in shot); in both these areas the trees do noticeably line/overhang the paths and are already bearing a reasonable thickness of leaves. I can't see anywhere that the tracks wondered more than 7-8 metres from where they were supposed to be; which seems reasonable and acceptable to me (may not be for some). The other key thing for me is that the distances recorded on each device were within 1% of each other. The charts on mygpsfiles also allows you to compare the HR recorded from the various sources (just make sure you select the X Axis to be time, as either distance or point always seems to end up with misalignment for comparison in my experience).
I know this is a GPS accuracy thread, but I was testing for HR performance at the same time. I'm not surprised by the HR results; the Scosche is the closest fit to the HRM Run chest strap, but still has a few wobbles if you zoom in, with the F5 loosing a couple of points where it briefly looses lock in a couple of places late in the run (the most obvious being a downhill section where the vibration on the watch is probably worst). Whilst I have low expectations for WHR performance when I get to test for HR during upper body strength training sessions (I've found even the mighty Scosche Rhythm+ is utterly rubbish for tracking HR in that scenario) with today's result I'm likely to be happy to use just the WHR on the F5 for casual training runs, but I'll still 'chest strap up' when accuracy of HR data is really important during races or for interval training.
Here's my first activity. The only places were it wandered significantly were next to a large metal object, the Angel Of The North sculpture (a small wander) and past two tower blocks (where the track briefly crossed to the other side of the road).
Looking at the same route from my F3, the F5 looks a little better. The F3 struggled in the same places but also had issues elsewhere. Measuring this route on Google Earth, I would say the F5 has it in terms of distance accuracy.
I know a sample of one doesn't mean much, but it seems ok so far in my opinion. Not the best, but ok.
You all do realize that this forum represents a VERY small percentage of actual Fenix users? My F3 has been solid and I've posted numerous tracks to prove it. Don't expect to attract happy users to this thread to post their excellent tracks.
Please, if you know you'll be unhappy with GPS traces that aren't perfect (despite total distances being within 1%), buy the F5 from a store with a good return policy. That way if you're not satisfied you can return it and move along.
Hey Olu, welcome to the next round of GPS Accuracy discussions ;)
Definitely agree with your comment; GPS Track Accuracy and GPS Distance Accrual Accuracy are not the same thing and Distance Accrual Accuracy is much more important. In my testing of the fenix 3 and now my 735XT, they both consistently deliver <1% error in distance accrual. I check this by plotting exact running course at a very detailed level compared to distance accrued on the watch and check mile markers as well as total distance.
Will probably be pulling the trigger on a fenix 5 in the next week or so, how about you Olu?
