@tommi: statistics you see are little bit modified. This is done to reduce huge noises that happen during recording. So in your case, point with highest altitude probably has a big altitude difference compare to points around, so it was rounded a lot to this lower value.
Ok, I understand this is not a bug but should be a feature.
Let me explain my usecase for the statistics: I plan mountain hiking tours based on Locus. Both diagram with the track profile and statistics with total distance and meters up play a big role here. These tracks are no recorded ones but drawn with the help of e.g. brouter and afterwards updated altitude with hgt 1" data from Locus/srtm folder. Even this data source has not yet the highest peaks in it. E.g. the altitude of the mountain peak in this track is 2476 m while the hgt data has 2450 m and Locus reduces the maximum again to 2419 m. So the statistics isn't too helpful.
For track recording you have defined a filter, right? This should do what you want. Why is it necessary to filter the statistics again?
And if you think you need the filter the track because it might have been recorded outside of Locus and containing a lot of noise you could offer a switch to allow filtered or unfiltered track for diagram and statistics.
And about compute of distance. It's correct. Currently all distances in Locus are computed only in flat. Anyway it's a question if compute should be done with real altitudes or flat is what you (users) expect.
Why not show both? Shouldn't cause big problems, right?
Again my usecase: I do an estimation of the tour's duration with the flat distance and the altitude up meters based on the data of two tours we made during the summer (two equations to determine the value of two variables :-))
So I use the flat distance here but to explain the actual length of the tour the real length is more interesting.
Therefore I'd like to see both.
Anyway, Locus is a big helper here :-)