Observing some considerable deviation

Continuing from https://github.com/graphhopper/map-matching/issues/95 (sorry, forgot about this forum):

No, I tested with 0.8 because I have this Error: Unable to access jarfile action=import problem as well (and no up-to-date graphhopper yet to build and copy those 2 jars manually as suggested there). Will test with master in a while, thanks for the update!

As for

If there are no paths in OSM how should GPS tracks be matched then?

what I meant was that users might upload GPX tracks from different sources, sometimes as recorded by an actual GPS device, sometimes from other sources. In either way, it’s likely that OSM has paths nearby, hence map-matching might make sense regardless. Seeing the effect with that particular track which didn’t come from a GPS device I was just wondering if that would/could make a difference.

Would you mind to test with master? Because it should work there already.

I’ would if I could build master (see Error: Unable to access jarfile action=import)