The backtracking orientation calculation is based on the geometry of a pair of trackers , i.e. the distance p between trackers, and the width w of each tracker.
These 2 variables are supposed to be equivalent from tracker to tracker. If it is not the case PVsyst will calculate the backtracking for the less favourable pair of trackers, and this calculation may not be optimal for more spaced pairs of trackers.
In the present time, PVsyst doesn't calculate the backtracking for trackers at different altitudes. However even when we will implement this, the altitude differences will be necessarily the same for all trackers (i.e. the tracker array will be in a same E-W inclined plane).
Moreover, remember that in PVsyst, when using the tracking option the orientation of all trackers is identical at a given time.
On a Hill
Now on a hill, when the altitudes differences are different from tracker to tracker, it is physically impossible to define a "pure" backtracking strategy. If you calculate the backtracking Phi angle between the trackers A and B, it will be different than for the trackers B and C, so that the backtracking cannot be optimal for both pairs simultaneously: either you will have residual shades, or the "back"tracking will begin before it is necessary.
Some people propose a situation where all trackers take a different position: this leads to extremely complex calculations (the optimization of the Phi angle of each tracker should be performed simultaneously on all trackers for each sun's position) without ensuring a perfect solution. This optimization involves machine-learning techniques for a given installation.
We don't envisage such a development in PVsyst in a previsible future.
For an approached calculation in PVsyst, please remember that basically the mutual shadings and "true" backtracking give very close irradiance results, as you intercept the same light tube. What you loose as shading losses in the first case will be lost as mis-orientation and IAM losses in the second case. Only the electrical mismatch shading losses give an advantage to the backtracking.
The mutual shadings are quite correctly calculated by PVsyst with trackers on a hill.
However a not perfect backtracking will loose either by unexpected mutual shades, and by "early-back"tracking (leaving some sun rays unintercepted between trackers). So we don't have an accurate way to calculate it.