I think I might know what's going on here.
The selection info angles (while rel to grid) are measured against the imaginary 2D plane of the grid, when the shock starts out in some odd orientation and you gather the sel info in that state you actually measuring it as if the shock was in abs/grid orientation.
If you then change the grid to match the shock in order to do the rotation you will get the wrong result as the angles had the wrong reference.
In short you need to set the grid orientation to the same plane as the segment it most be rotated over. The is usually done by setting the grid to the shocks orientation before getting the angles.
Hope this is somewhat clear.
The selection info angles (while rel to grid) are measured against the imaginary 2D plane of the grid, when the shock starts out in some odd orientation and you gather the sel info in that state you actually measuring it as if the shock was in abs/grid orientation.
If you then change the grid to match the shock in order to do the rotation you will get the wrong result as the angles had the wrong reference.
In short you need to set the grid orientation to the same plane as the segment it most be rotated over. The is usually done by setting the grid to the shocks orientation before getting the angles.
Hope this is somewhat clear.