jiggy
Veteran
Well, others have already mentioned the advantages of EO, 2-gen solving and alternative Petrus last layers. However, I think it should also be of note that step 4 was not intended to be solved in a Fridrich F2L style, but instead should allow for situations where block building can offer a better solution. (In a similar way to ZZ, I suppose.)Also, I noticed that step 3 seems to be flipping edges, analogous to the first half of 2-step OLL. If there are 2 "bad edges," then one would take about 4 moves to correct this, while an OLL algorithm would take 6. While this seems to be an advantage, I realized that step 4b is greatly restricted - one usually cannot use regular F2L algorithms to solve the last slot. Is it reasonable, or perhaps even advantageous,to skip step 3, continue to step 4a, and then use F2L and OLL algorithms instead of steps 3 and 4b?
What you are essentially doing, I feel, is close to building a double X-cross and then solving with Fridrich. As Kirjava said, most things have been thought of and experimented with before. Especially when it comes to hybriding two predominant methods. I think the reason you weren't able to find much information on this in your searches is because it's a bit of a nothing idea (sorry). It's kind of an idea to which people would say "oh, well sure, yeah of course you could do that...what's your point?" If you want to solve like this, go crazy! It may turn out to be quite a fluid solving style which suits you well.
However, I do really like that you're branching out and trying different methods. Keep doing that! =)