Version: 0.98-devel (kst_oldview_branch) (using KDE Devel) Installed from: Compiled sources Using KDE 3.1.0 with Qt 3.1.1, kst from CVS (kst_oldview_branch) There is an error in the statistics code. When looking at noisy, but otherwise flat data, the mean is clearly incorrect (it's larger than the max!). How to reproduce: Plot some data. Look at the scalar statistical values (ie [V1-INDEX-Mean]). The mean is incorrect (outside the min and max). Some of the other values look suspect as well (Rms?, square?). Expected behavior: The statistics scalars display correct values.
After a closer look, the statistics code is probably fine. The error is in the number of samples (for reference, we are getting our data from BLAST style dirfiles, tested with both 'slow' and 'fast' data). In both cases (fast and slow), the mean can be made correct by assuming that the real number of samples is one greater than what is given with statistics [VECTOR-NS].
*** Bug has been marked as fixed ***.