The way I understand the transitions to work is as follows (although I could be wrong):
The ONLY place that moves with the transition is at a junction where a material starts or stops.
So for instance:
You can have a wall that has three transition junctions - one where concrete changes to siding on the outer portion of the wall at 2'-0" high - One where a frost wall stops on an inner portion of the wall at 3'-0" high - and one where siding changes to shakes on the outer portion of the wall at 6'-0".
You set the default transition default height to 2'-0" because typically that's where you want to step the concrete as the grade moves. if you want to raise the concrete to 4'-0" high, you edit the wall and change the transition to 4'-0" on the transition tab. The concrete/siding junction moves to 4'-0" and should be the only part that moves.
Say you then want to move the siding/shake transition to 7'-0", You edit the wall definition locally to the drawing to move the transition height to 6'-0" high (the current transition height in the definition for siding/shakes). Then you edit the wall and on the transition tab change the transition to 7'-0". The siding/shake junction moves to 7'-0" and should again be the only part that moves.
As long as you don't have multiple possible transitions at the same point in the original wall definition, the ONLY part that should move is the junction at the CURRENTLY defined transition height.