The New Method / Substep / Concept Idea Thread

[Shiv3r]

Okay guys, I have already converted some Freeslice Users to this big cube method I call 3dux (Threedux) .
Basically it's a variation of Yaudux(I used to call "Gay redux" turns out it has already been named), and I suggest using it on 6x6+ however you can do it on 5x5.(I main it on 6x6 but not 5x5).
so here are the steps:
1.Centers
2.with freeslice, pair up 3 cross edges and place them correctly
3.freeslice 5 more edges, and place the first edge you solve on the D layer with the intuitive 5-move Shiv3r's variation algs I posted earlier
4. A modified M-slice Freeslice L4E, where when you need to "flip" two pieces with a U2, adjust with a big wide move so you don't disturb the 3 cross dedges and do a U2. besides that it's basically normal Freeslice L4E
5. Solve 3x3, by first solving the 4th cross edge. You can usually track the cross dedge's position while finishing L4E.

any Questions/Critique? Me and Pedro Seca are the only people who I know who main this method at the moment.

 

[crafto22]

Hm, I'm not sure if corner separation in 3 moves is possible while preserving the E slice.

Also, having generated random CP algs in the past I'm tempted to say CP + DL will be more around 10-11 moves. That's just a guess though.

Even whilst preserving the E-slice it's 4 moves at most, and that's including AUF. The idea was bad though, so it doesn't really matter.

 

[crafto22]

New LSLL idea:
1. Orient three U-layer corners, intuitive (4/4)
2. Finish F2L and solve CP, 36 cases with good recog, same as TTLL for the corners (8/12)
3. ELL, 29 cases (9/21)

Compared to regular LSLL:
1. Finish F2L (7/7)
2. OLL (9/16)
3. PLL (12/28)

Saves 7 moves and only requires 65 algs compared to CFOP's 78

 

[Sue Doenim]

". . . my first proposals were absolutely awful (ZZ with UL/UR in the line to do LSE . . ."

Why is this bad? It's just ZZ-L6E. I found out about it a little while ago and have used it over ZZ-VH since (I switch methods incessantly). Are you just talking about in terms of compared to ZZ-CT and such? Also, I'm planning on learning WVCP (and probably SVCP) for the R U' R' (and R U R') case(s) to increase COLL skips from 1/162 to 1/27 (I use way too many parentheses).

 

[crafto22]

Why is this bad? It's just ZZ-L6E. I found out about it a little while ago and have used it over ZZ-VH since (I switch methods incessantly). Are you just talking about in terms of compared to ZZ-CT and such? Also, I'm planning on learning WVCP (and probably SVCP) for the R U' R' (and R U R') case(s) to increase COLL skips from 1/162 to 1/27 (I use way too many parentheses).

Yeah honestly ZZ-L6E seems pretty good. L6EP is like 9 moves, only a bit higher than EPLL which is like 7.5, so why not?

 

[Thermex]

New LSLL idea:
1. Orient three U-layer corners, intuitive (4/4)
2. Finish F2L and solve CP, 36 cases with good recog, same as TTLL for the corners (8/12)
3. ELL, 29 cases (9/21)

Compared to regular LSLL:
1. Finish F2L (7/7)
2. OLL (9/16)
3. PLL (12/28)

Saves 7 moves and only requires 65 algs compared to CFOP's 78

I like this idea, but are you sure about the alg count and recogniton for step 2? It seems a bit off to me (I might be wrong). You might want to check that out.

 

[crafto22]

Okay so I was looking into L5C and stuff and basically here's what I came up with:

1) F2L-1. Anywhere from 18 to 30 moves depending on the method.

1.5) If edges are not already oriented, orient them whilst inserting the FR edge. 3 moves at most for Heise, ZZ, Petrus, etc. Around 6 moves for non-EO first methods.

2) Orient three U-layer corners. 7 moves, mostly intuitive.

3) Solve the corners. 9 moves, 16 cases.

4. EPLL. About 7 moves, 4 easy cases.

This would give a 26 move LSLL (versus 23 for pair + COLL + EPLL) for EO first methods, or 29 (versus 28 for pair + OLL + PLL) for non-EO first methods. This is done with only 20 algs and produces solutions that are only very slightly less efficient than regular LSLL.

 

[crafto22]

I like this idea, but are you sure about the alg count and recogniton for step 2? It seems a bit off to me (I might be wrong). You might want to check that out.

No worries, I have all the cases generated already, only 36 surprisingly. And recog is the same as TTLL for corners since the only one that is twisted on the U-layer is the DFR corner (not necessary for recog) and identifying the location of the edge is instant.

 

[Shiv3r]

so I will post again, however no one seems interested very much in what I am working on right now.

 

[Thermex]

No worries, I have all the cases generated already, only 36 surprisingly. And recog is the same as TTLL for corners since the only one that is twisted on the U-layer is the DFR corner (not necessary for recog) and identifying the location of the edge is instant.

Oh you never specified the DFR corner was the one twisted, so I thought it would be way more algs. Could you show us the algs? (Im pretty interested).

 

[crafto22]

Oh you never specified the DFR corner was the one twisted, so I thought it would be way more algs. Could you show us the algs? (Im pretty interested).

Yep, I'm actually leaving for class right now but I'll share the document when I get back!

 

[Neuro]

@crafto22 you may be interested in the VOP method (2x2) Here's a link to the wiki page and here's a solve

 

[crafto22]

Yea so VOP is sorta like the equivalent of what I'm doing except for 2x2. Also, I'm making better algs, I'll post them in a separate thread once I'm done.

 

[2180161]

Okay so I was looking into L5C and stuff and basically here's what I came up with:

1) F2L-1. Anywhere from 18 to 30 moves depending on the method.

1.5) If edges are not already oriented, orient them whilst inserting the FR edge. 3 moves at most for Heise, ZZ, Petrus, etc. Around 6 moves for non-EO first methods.

2) Orient three U-layer corners. 7 moves, mostly intuitive.

3) Solve the corners. 9 moves, 16 cases.

4. EPLL. About 7 moves, 4 easy cases.

This would give a 26 move LSLL (versus 23 for pair + COLL + EPLL) for EO first methods, or 29 (versus 28 for pair + OLL + PLL) for non-EO first methods. This is done with only 20 algs and produces solutions that are only very slightly less efficient than regular LSLL.

I had a similar idea to this, it went:

F2l-1
Insert FR corner while orienting edges
COLL
L5E
I think it could work though.

 

[Sue Doenim]

@crafto22 you may be interested in the VOP method (2x2) Here's a link to the wiki page and here's a solve

If you think about it, VOP is ZZ-CT for 2x2.
What if you mixed this with an EG-style V? You would only need 26 algs, I think. Making the V is really easy, I'm not sure it can even be over 2 moves. With ~4 moves for orientation, all that's left is permutation.
Method
1) Match 3 stickers on a face, ~1.5 moves (can someone explain what the movecount things in brackets are, and how to get them?)
2) Orientation ~4 moves
3) Permutation -I don't know, I'll generate some algs

 

[crafto22]

I had a similar idea to this, it went:

F2l-1
Insert FR corner while orienting edges
COLL
L5E
I think it could work though.

Your movecount would be roughly:
DFR + EO: 5
COLL: 10
L5EP: ~11 (according to this)
Total is 26.

This method would have 54 algs. Although you might as well use VH:
VHLS: 9
COLL: 10
EPLL: 7
Total is 26 too, but only 46 algs.

My method's (estimated) stats:
Orient 3 U-layer corners: 4
CP + FR: 9
ELL: 9
Total is 22, 65 algs.

 

[Sue Doenim]

Example solve, EG VOP
Scramble: R2 F R2 F R U' F2 R2 F'
x y'
V: R'
Orientation: U R F R F' R'
I guess these will be longer than I thought because the V has to be preserved.
Permutation: (U D) R2 U R2 U'
13 moves total. EG's average is 12.35 moves.

 

[crafto22]

You mean what these mean? --> (x/y)
x is the number of moves required for the step.
y is the total number of moves so far after this step.

That 2x2 method looks phenomenal!! I strongly recommend you develop it! I would definitely use it.

 

[shadowslice e]

If you think about it, VOP is ZZ-CT for 2x2

No ZZ-CT is VOP for 3x3.

 

[Neuro]

No ZZ-CT is VOP for 3x3.

Agreed. VOP was made before ZZ-CT (or at minimum completed before)