Interesting New Method: a bit of everything

[Smiles]

So I came up with a new method and it's not so great speed-wise, but it's quite interesting and I just feel like it's not optimized.
And I guess people will say that Roux defeats this method, and I can almost certainly agree on that.

Anyway, there are 5 main steps in the method, although you could break it down to 7 different steps.

1. Build a 1x2x3 block. This step is the same as Roux.
2. Orient all the edges (9). The concept of this step can be viewed in different ways, like doing moo or the ZZ way, which are essentially the same thing. However, the centers are free (Roux) and the whole cube is somewhat free at the same time (ZZ).
3. Place the DB edge to DB. Don't ask why.
4. Build another 1x2x3 block opposite the first one, while placing the FD edge. This can be broken down into 2 steps, building a 2x2x2 block that includes the FD edge, then inserting the CE pair at the back (example solve 2 & 3). Or, make two 1x2x2 blocks (example solve 3 redone).
5. Solve the last layer. Since the edges are oriented, this opens up some options. LS methods like WV can be used, or OCLL/PLL, COLL/EPLL, straight up ZBLL, whatever.

Step 4 is took me a while to figure out, but I did it and you can see how it works in the example solves.

Spoiler

I'M SORRY BUT I'M A RIGHT HANDED PERSON CUBING LEFTY. DEAL WITH IT.*

Example Solve 1
Scramble: R2 U2 L2 D L2 F2 D2 R2 B2 L U2 B' L2 F R2 D R U L U2

B' L U2 L2 U R U2 F U' F // 1st Block (10/10)
U L' U L l U l // EO (7/17)
U' M2 // Place DB (2/19)
U l' U2 l U' L U L' // 2nd Block + FD (8/27)
U' L' U' L U' L' U' R U' L U R' L2 U' S U2 S' U' L2 // COLL + EPLL (19/46)

= 46 STM (or 50 STM, pardon the U perm).


Example Solve 2
Scramble: F2 L B2 U2 F2 U2 L B2 L' F2 R2 D' F' U' L D B2 U B' F' L'

U' R U' R F' U2 R U R' // 1st Block (9/9)
l2 U' M' U' M2 U2 M' // EO + DB (7/16)
L' U2 L U' L' U L M' U2 l // 2x2x2 Block (10/26)
L2 U2 L U' L' U L' U2 L2 // CE Pair (9/35)
L' U2 R U' R' U2 L R U' R' // COLL (10/45)
y' M2 U' M2 U' M' U2 M2 U2 M' // EPLL (9/54)

= 54 STM. Wasn't a good solve, that F2L case really bugs me in EO methods.

So the reconstruction was structured differently from before, but only 'cause I felt like it.
Actually because that's how the solve actually feels. The first solve is just the steps.


Example Solve 3
Scramble: L2 B2 R2 D L2 F2 U' B2 D2 U' F2 R' D' L' R U2 B R2 F U

F E R' y' L' U' R U R' // 1st Block (8/8)
D l' U l D' L U' l2 // EO + DB (8/16)
L' U2 L2 U' l' U2 l // 2x2x2 Block (7/23)
U2 L2 U2 L U' L' U L' U2 L2 // CE Pair (10/33)
F' U' L' U L U' L' U2 L U L' U' L F // COLL (14/47)
L U' L U L U L U' L' U' L2 // EPLL (11/58)

= 58 STM. This solve was even worse, move count wise. I got that F2L case again.


Example Solve 3 (differently)
Scramble: L2 B2 R2 D L2 F2 U' B2 D2 U' F2 R' D' L' R U2 B R2 F U

F E R' y' L' U' R U R' // 1st Block (8/8)
l' F U F' l L U' M2 // EO + DB (8/16)
U L' U' L U2 L2 U L U' L' U2 l' U2 l // 2nd Block + FD (13/29)
R2 D' R U2 R' D R U2 R M2 U' M2 U' M' U2 M2 U2 M' U2 // COLL + EPLL (19/48)

= 48 STM. I saved 10 moves just by changing the EO, which I realized after redoing it.
Also, I see there could have been cancellations after 1st Block and 2nd Block


*In case you haven't realized, you'd have to mirror the scramble + the solve if you want to see how a righty would do this. I'm so very terribly sorry.

Pros
- Like Roux, this method doesn't require rotations (unless they exist in the LL algs but algs are algs).
- Mostly <R,r,U,M>.
- Good look ahead since EO is done and the DB edge is done. Not as good as Petrus, but better than ZZ.

Cons
- Can be hard to identify bad edges in step 2, since there are so many and they're all around the cube. However, planning the entire first block during inspection can hint as to which ones are the bad edges.
- Sometimes move restrictions (from EO and 1st block) get in the way.
- Step 4 has a lot of moves and is complicated, like tracking 2 edges + a CE pair all at once.

I had been trying to come up with a new method when I was bored, and this is the only usable idea that came to mind.

 

[Clarkeeyyy]

How would you orient the edges after building the 1x2x3? Could you do an example solve?

 

[Smiles]

it's in the spoiler, i did 4.

edit: i'll explain it here too.
if you already know how to orient in ZZ and Roux, this shouldn't be so hard.

I find all the bad edges (max 8) and, let's say i have 4 bad edges:
i can do a slice turn to get my F/B center onto the U face, so it's still actually the F/B face, not the U face that i'm looking at.
at the same time, i get all 4 bad edges onto that layer (which is F/B) and turn that layer. it changes the orientation of all the edges on that layer.
then i slice turn (or wide turn) to get that F/B face back onto F/B.

since i'm white/yellow (U/D) colour neutral for this method and roux, then what i mean by F/B face is green/blue or red/orange.

if only 2 bad edges, then i can do it how petrus/ZZ does it.
If more than 4 bad edges, i could do Roux style with M and U. it really depends on the case and where the edges are.

 

[arvind1999]

Seems a bit difficult to get used to EO, and its quite inefficient too.

 

[Smiles]

Yeah that's why I called it "interesting" and not "good" :/
it can be improved in so many ways that would just basically make it a slightly modified CFOP or Petrus.

 

[Kirjava]

We explored almost every EO early in Roux variation in othe threads and couldn't find anything good.

What happens if you have DR in DF during step 4? That case is just awful.

 

[MWilson]

You're making the whole thing too complex by orienting the LL edges. Just solve DF+DB right after the first block, then orient the remaining three edges of the second block, which can be mostly or completely done during the solving of DF+DB. Then do ZZ second block -> OLL/PLL. That may be closer to Petrus than anything else, although still not that close.