I made the system up after looking at the various expansions on Fridrich's last layer method, where the idea is to do stuff to the last layer as the last corner-edge pair is solved. Most of what I saw added a severe number of algorithms to an already complex system.
Lars Petrus already has a method which gets stuff done on the last layer once the first 2 layers are done--the edges are oriented. That being said, it still requires 3 steps to finish the last layer (permute corners, twist corners, permute edges), or 2 steps if you combine the corner steps.
So I came up with my variation to see if I could blend the advantages of Petrus with the ideas being explored by the Fridrich method expansions as described above.
Because the edges are already taken care of in the twist the edges portion of a petrus solve, and because of restricting moves to the U and R layers, the number of cases that you need to know is reduced. This is also why the FR corner edge pair needs to be solved last, and why it needs to be assembled and parked. There would be way too many algorithms and patterns to learn if the corner edge pair wasn't assembled, or if either the FR or BR CE pair was the last to be inserted.
