Same tour, same time, different values?

From looking at the two activities the HR tracks are basically pretty much the same ave 118 or 120 BPM - one has a spike of 175 and it looks like that HRM is not making good contact as it has large and constant fluctuations. Also have you looked closely at the tracks on the map to see how closely they match with the paths or trails you were running on. The 910 seems to track a smoother course whereas the Fenix looks like a very wobbly line - does the path reflect that ? - my guess is that the 910 is closer but you are familiar with the terrain not me. Also around about Kaiserwald 2 there seems to be a spot where you are running around at random in a small area - this happens on both watches but I would imagine could account for a lot of the difference in distance between the two watches as the meandering seems to be more pronounced on the Fenix than the 910. Does this area have difficult satellite reception ? You only have 10% difference in distance between the two watches and this sort of meandering could easily account for that. I'm not familiar with the Fenix 2 or 910 can you see the satellite accuracy once they have found the satellites? I find with my 610 that to get an accurate track it is important to let the accuracy settle down to below 20ft before I start my run - if I do that then the accuracy of the whole run is generally good. If I set off before it has settled down then it never regains the same accuracy - so I would suggest that you don't set off immediately you have the satellite lock but wait about another minute. I would be interested to hear what you think.

I have imported both activities into Garmin Basecamp. There it said that one activity was 17.7 km, the other one 17.8 km. Then I copied the data from Basecamp into Excel and did my own calculation, coming to the result that one activity is 16673 m and the other one 16667 m.

Now I am happy that these values coincide that much within each application, but still I wonder that they differ between applications.

So basically I have now four different values to choose from, between 15.7 and 17.8 km.

Maybe different applications apply different levels of data filtering, that even depends on the device that the data was recorded with?

Overlaying both tracks in Basecamp it looks like the fenix has a shorter time interval between two GPS measurements, therefore doing less implicit smoothing and therefore measuring a longer distance (see https://en.wikipedia.org/wiki/How_Long_Is_the_Coast_of_Britain%3F_Statistical_Self-Similarity_and_Fractional_Dimension). Or the fenix just adds more measurement noise due to the higher measurement frequency. The "meandering" is probably accumulated GPS inaccuracy during a break.

Judging from the map both tracks are reasonably close to the pathes I went, at least regarding the usual GPS accuracy. There are no large deviations. But the distance between two measurements from both devices at the same time differs up to 25 m.

I did not try Strava because I would have to register there first.

Anyways, it looks like the data reported by Garmin Connect (or any other application in this respect) is not more than an educated guess.