Fenix 6 Pro FIT vs GPX Files - Distance Discrepancy

districtsoul said:

but I haven't done a scientific approach to wearing both to compare distance recorded on same runs.

I have dome this numerous times vs. a Polar V800, Forerunner 245, Coros Pace 2 and even wearing 3 Fenix 6 watches at one, one of which was paired to a calibrated Stryd.

I have never seen this shortness issue my end - actually I say 'never' but in one case the F6 came up unexpectedly short vs. a Forerunner 245 but other than that in nearly a year, I've seen it once where the shortness was way outside the expected margin for error for GPS.

I have posted loads of examples over the last few months. Including official races where the distance is known and measured.

Having said that, I am sure other people DO see this problem, they have been too many threads regarding this including compelling evidence posted by AndersB and SilentVoyager - but why it seems to impact some people is odd - I suspect environmental issues come into play. In the long running 'running distance is usually short thread' one user reported that Garmin swapped out his watch and it resolved the problem. I think if you run in thick tree cover, then this is more likely to occur.

I really wonder whether this problem might be related to not being able to run at a half-constant pace. If you have very little variations I'd guess this algorythm can deal with this a lot better. For me, having a muscle condition and reacting to every little bit of terrain, wind or street surface change I get massive differences. This is even worse when I do stride repeats of 20 sec fast, 45s slow. The watch just can't keep up.

Yes, varied running technique have impact (for me) on both distance and pace. Especially pace. If i run faster and more stable it gets better but still not good enough for me.

I think it is not only AndersB and SilentVoyager compIaining. I have already joined the forum, but I am checking the issue since June 2021. And distance reduction seems to be happening too many people, some of them are not aware of it as they only check how trace looks or they just do not have good comparison. For sure problem is applicable mostly only for runners (watch is not pointing upwards to the sky like when you ride and your movement is much slower). When you add trees with leaves the problem appears. These conditions with Fenix construction leads to weak GPS signal, which is not good enough for Garmin Algorithm calculating (smoothing distance). Even if signal is week algorithm should reduce the distance so strong as in the result it is shorter 2-10%.  I do not know you someone already check .FIT form your watch, but I can check how algorithm works with your device. I have already checked a few F6, some F5 and only one F6 x pro looked good. This came form the guy who already sent back 2 watches having the problem and the 3rd one was much better. The concept of approximation algorithm is good, but in case of weak Fenix GPS signal it unreasonable as below:

I have only posted comparisons with running - you can probably find some of my posts in the 'running distance too short' thread.

I do not see the problem on gpx export - I already posted results from a 13.1 mile half marathon compared with a Polar v800. 

This is VERY typical for me....about 70 meters difference here over 13 miles...

And 1 of these watches below was paired with a caliberated Stryd - less than 45 meters difference and most of this run was in a town with narrow streets.

Do not assume that only a few people are 'clever' enough to see the problem and those who do not are dumb-dumbs

I have results with multiple watches since early 2021  -

Ok, I thought it might play a role. I have to say that your widget is really doing a great job for me, AndersB. Here I get the results I'd expect, stride repeats look plausible, and I don't get the dreaded drop-outs.

Hmm. would you want to check out my 6s? If so let me know what you need. I can't really provide a very long run at the moment, having had a flareup recently, but I at least have some of the slowest and most varied running data you'll probably ever get your hands on

Thank you for your reply. I will check. 

5135968 said:

and most of this run was in a town with narrow streets.

so It is not our case I assume. I did not observe any issue in urban area, it works perfect. I focus only one case: slow running in the forest.

5135968 said:

Do not assume that only a few people are 'clever' enough to see the problem and those who do not are dumb-dumbs

I would not say so that we are smarter or something like this (probably we spend more time on it to dig the issue, I have learned a lot doing it), but it also happens that some people are not able to understand what we are discussing or they just do not want to (I do not how it is on thus forum, but it visible in many Garmin FB groups). It is also possible that some people do not have chance to observe like: 

• bikers
• people who run mainly with open sky areas
• running alone and cannot compare with others

If it is not common problem for this Device we just want to get know the scale of the problem. If it only a few single faulty units (it does not look so) I would like to have the comparison to make the complaint (so I ask people for .fit file to see if their devices work so good as they say). Unfortunately I have already talked with a few guys who had the watch replaced for new one and it was the same, so I would like to avoid. 

I use the following method to observe the problem with Fenix 6 distance. I take an event with a large (100+) number of runners. Then I find and open Strava activities for each participant that I can find, compare the distance and note the watch model. To find activities I use either Strava Flybys or one of segments on the race course. The segment method is pretty good actually. I look at a current year leaderboard for any segment on the race course (doesn't matter which part of the course it covers). Then I look at the entries in the leaderboard that match the race date and open each of the matching activities to see the details. I do a quick check of the GPS trace to make sure that it is complete and matches the race course, then write down the distance. This method gives me a pretty good and fairly consistent picture of distance accuracy across multiple devices.

I should note that in trail races probably 95+% of distances across all watch brands tend to be short, but some are shorter than others, so it is usually easy to spot which ones are more accurate. I looked at a number of races this way, and am starting to see a consistent pattern.

Fenix 6 tends to produce distances that are among the shortest - shorter than majority of results across the board. Other devices that tend to be as short as Fenix 6 or even worse are Fenix 5, Vivoactive 3, FR 935, and Coros Vertix 2. The majority of other Garmin watches are more accurate - 245, 645, 745, 945, 35, 45, 235, 920, etc. Coros Pace and Apex tend to be more accurate too. Sunto 9 and especially Suunto Ambit 3 are more accurate too. The most accurate devices for trail running seem to be Garmin FR 745, FR 945 LTE, Apple watch series 7, and Strava running on iPhone. In contrast, Strava running on Android is hit or miss, probably because the hardware can be so different. 

Pech13 said:

hank you for your reply. I will check. 

Not sure this was for me, but website doesn't allow me to message you.