Course ascent data very different between Connect and Komoot/Outdoor Active

marcossaad_ said:

In fact, there are not 2 apps to agree on total ascent…

That's normal, since the total elevation gain depends on the number of trackpoints, and the exact position of every of them. And the number and position of trackpoints are more or less random at every app (and may even differ every time you create the course). The total ascent and descent are calculated as the sum of elevation differences between adjacent trackpoints. The more trackpoints (the finer resolution), the higher the sum will be. Additionally, some apps use a smoothing algorithm to filter out smaller "bumps", others don't.

That told, it is still possible that Garmin has something wrong in their algorithm. It would require a deeper analysis - for example opening the GPX file of the course in Excel and summing the elevation of all trackpoints there.

I was referring to the same track uploaded (not created) on different apps.

None will agree on total elevation despite the fact the number of points, coordinates, altitude, speed, etc to be the same for all of them. 

marcossaad_ said:

None will agree on total elevation despite the fact the number of points, coordinates, altitude, speed, etc to be the same for all of them.

Why do you think the number of points is the same? Each app may change the number of trackpoints, and their position on import, so the result will usually differ anyway. Garmin Connect definitely does reparse imported tracks.

at least two of them, Basecamp and Plotaroute, only optimize the points if you tell them to do it. Even thought they don't agree on total ascent. 

I haven't test it but it should not be difficult to import a file to others apps export them back and open at Basecamp that give you the number of points.

I'll do it and report.

The question is whether the original track that you are importing contains also the elevation data, or only the coordinates. That's another potential source of differences. And then, as I already wrote, each service may implement own smoothing/filtering algorithm in order to avoid the "Coastline paradox" - getting too high estimates of the total elevation gain at tracks with fine resolution.