The New Method / Substep / Concept Idea Thread

[Shiv3r]

Okay, so y'all remember Hawaiian Kociemba? well, I think solving only 3 cross edges is actually really good. I can usually plan xcross-1 edge in inspection, and the L5EOP algs are nice(L5EOP is EO+DF, then EPLL), and the L5EP algs are pretty nice(much of the algs can be done with a conjugated EPLL, or an H-perm and then a conjugated EPLL). So here's what I use when I solve it, and though I didn't invent it (it's called 3CFCE on the wiki), it's pretty nice:
-3 Cross edges(and whatever F2L you can plan in inspection)
-F2l without the edge(often you can use the open face to make pairs really efficiently)
-CLL(I use CMLL's, they work usually)
-Edges(L5E). there are 2 ways to do this, L5EOP(L5EO->EPLL) and EO->L5EP. the alg count is the same, but L5EO is pretty intuitive for Roux solvers because you just force the Edge on D, so I suggest that strategy.

I actually got a nice scramble a while back and got a 12 sec solve with this, so it's pretty good. Any thoughts?

 

[Shiv3r]

So just a random Roux that I decided to do because it makes roux solves more fluid and I haven't heard of someone doing it anywhere. copied from a doc I am making with a bunch of LSE things I am working on.


Where BU is in the 2 ugly ULUR cases

The reason I did this is I realized that even with EOLR, there are some cases that are unavoidably ugly. In this, you can recognise BU right after EO with these cases. These are organized into 2 cases: ugly case(M’ U2 M U/U’M2) and back ugly case(M’ U2 M’ U/U’ M2).

-Ugly case: Solve with M’ U2 M (U/U’ M2 insert)
UF and R corners’ colors match: BU at DF.
UF and R colors don’t match: BU at UB.

-Less ugly case: solve with M’ U2 M’ insert
For both: BU at UF.

 

[Sue Doenim]

I actually came up with this a while back but never posted it. Could be a good CP style. What I'd probably do is something similar to my Roux CP style simply for the efficiency though. Does yours orient edges also or just bring it to a <RrUM> 2GLL? Orientation would give you more LS algs but full 2G 1LLL would have a ridiculous amount of algs and would have crappy recog. Of course ZZ/Petrus don't have this issue, but it seems like your aiming to CFOP OH.

No, I said, although not very clearly, that it was for methods that take of EO first. I think <R,U,r,M> 1LLL should be called pseudo-2GLL. It's really pretty nice. However, I'm pretty sure it would have about 688 cases.

That's more or less zz-porky, zz-rainbow or zz-orbit (I can't remember exactly which one it is)

ZZ-Orbit is very similar, but this doesn't require the pair to be made. ZZ-Orbit is to this as VHLS is to ZBLS and WV is to ZZ-c.

 

[Hazel]

There have been LSLL subsets proposed and used, but why not try to create a last 2 slots last layer subset?
My idea was to make it so you can skip OLL 100% of the time, providing you do this on every solve:
1) Solve the cross and 2 adjacent F2L pairs
2) Solve the 3rd F2L pair while orienting the 3 LL edges that would still be in the last layer (sometimes there's 4, but still only orient 3/4)
3) Solve the last F2L pair while orienting corners (this can be done with Winter Variation, and would take advantage of the orientated edges to fully solve OLL with this step)
It sounds a bit complex, but it's quite simple really.

Here's an example:
Setup: F' U2 B D' L' B' L2 U' L2 B L' D B' U2 F

3rd F2L pair + Edge orientation: U2 L' U L F U' F'
4th F2L pair + Corner orientation: y L' U' L U' L' U2 L' D' L U L' D L2 (this is just solving the F2L normally and canceling into Winter Variation)
PLL: R perm

 

[Neuro]

Hey so here's an interesting ZZ style Megaminx method if anyone is interested. It's similar to ZZ-Spike, but requires you to look at less faces during EO.

1: F2L
2: Yu Da Hyun style S2L until you have only F R U and L faces to solve.
3: Solve Balint Block on F and do full EO (very similar to ZZ-Spike recog but quite a bit easier IMO)
4: ZZ-F2L
5: LL as normal

I can make a video if anyone is confused, leave suggestions for this!

 

[Shiv3r]

Hey so here's an interesting ZZ style Megaminx method if anyone is interested. It's similar to ZZ-Spike, but requires you to look at less faces during EO.

1: F2L
2: Yu Da Hyun style S2L until you have only F R U and L faces to solve.
3: Solve Balint Block on F and do full EO (very similar to ZZ-Spike recog but quite a bit easier IMO)
4: ZZ-F2L
5: LL as normal

I can make a video if anyone is confused, leave suggestions for this!

I do Yu da Hyun(I think), I do the same thing actually except for EO. it would be easy.

 

[Sue Doenim]

Here's a couple of megaminx ideas.
1) F2L-2
2) VHLS with second to last pair
3) Build 2x2 block (edge-corner-edge-center block) on U
4) Solve last F2L pair
5) OCLL preserving block
6) Reduced PLL subset
You don't necessarily need to solve EO after F2L-2, but it makes building the block easier and has less OLLs. My next idea is probably better.
1) F2L-1
2) EO (you could do this earlier on)
3) EPLS: permute edges whilst inserting last F2L pair, very intuitive
4) OCLL preserving EP- there are algs available that do this
5) CPLL
Here's an idea I had for 3x3. It's really weird.
1) EO
2) Solve a 2x2x1 block on both the U layer and D layer
3) Solve edges
a) Solve E-slice edges
b) Solve other edges <R2,U,D> in regards to U/D layer blocks
4) L6C- solve with commutators
Might be useful for FMC. Would only be useful in speedsolving if you had the look ahead of 10 Felikses (Felicies?).

 

[Shiv3r]

Here's an idea I had for 3x3. It's really weird.
1) EO
2) Solve a 2x2x1 block on both the U layer and D layer
3) Solve edges
a) Solve E-slice edges
b) Solve other edges <R2,U,D> in regards to U/D layer blocks
4) L6C- solve with commutators
Might be useful for FMC. Would only be useful in speedsolving if you had the look ahead of 10 Felikses (Phalluses).

It would be awesome for Chess clock FMC, a game I observed at an in-n-out last saturday after a comp, which is where you and an opponent have a chess clock, and you alternate doing a move and hitting the clock. Whoever finishes first wins.

 

[Hazel]

Advanced 2x2 method idea:
1) solve corner orientation (So the U and D sides consist entirely of 2 opposite colors, like white and yellow for example)
2) advanced PBL (anybody know how many algs this would be?)

 

[Teoidus]

This sounds a lot like Guimond.

Your second step would be 7! algs for every possible case, but of course with AUFs the number would go down

 

[Sue Doenim]

Advanced 2x2 method idea:
1) solve corner orientation (So the U and D sides consist entirely of 2 opposite colors, like white and yellow for example)
2) advanced PBL (anybody know how many algs this would be?)

Note: NLL and LOLS are steps in the HD method
I'm pretty sure most people here have thought of this, but the idea gets killed by the case count. However, this does give me a new, HD-esque idea:
1) Solve 1x1x2 block
2) Orient corners
3) Permute corners
First step is once again, very easy. The second step would probably just usually be a one move setup to a LOLS case, but they could be more efficient, as only two pieces need to be preserved. The permutation would have 12 NLLs and 4 PBLs if you're willing to do 1-2 move setups and rotations.
Wait . . . . . . . . . . . . . . .
If you tweak the OBLs in Guimond you can end up with pre-NLL cube states almost all of the time . . . and for the rest you can do a 2 move setup to change this.
I think this could successfully dethrone EG as the go-to 2x2 method.
Example solves:
Scramble- F' U' F' U' R F' R2 U R
V+CO- U R' U2 F R' (5/5)
NLL- D R2 U' R2 B2 U' R2 U R2 (9/14)

Scramble- F' R2 F' R2 F R2 F2 R2 U'
V+CO- R U R' F' (4/4)
NLL- R2 F2 U' R2 U (5/9)

Scramble- F' U' R2 F' R U2 R2 U' R2 U'
z y
V+CO- U' R2 F2 U' F (5/5)
NLL- D R2 U' R2 B2 U' R2 U R2 U' (10/15)

Oh my.
As good as or better than EG in less than half the algorithms.

 

[Teoidus]

Will it hurt the 1look potential though?


ooooo, while we're here, I have an interesting idea:

CPLine
1look L6C? How many algs is this? (@xyzzy halp)

 

[Sue Doenim]

Will it hurt the 1look potential though?


ooooo, while we're here, I have an interesting idea:

CPLine
1look L6C? How many algs is this? (@xyzzy halp)

~1080, I believe. However, L5C is only 120, I think. I thought of this method quite a while ago but kind of forgot about it.

 

[Teoidus]

L6C is 1080 algs? Wow... you sure it doesn't go down when you allow AUFs/ARFs?

Edit: hm, there are 29160 cases without AUFs/ARFs. Well, maybe one could build CPLine + force the other first layer corners somewhere...

 

[Neuro]

Wait . . . . . . . . . . . . . . .
If you tweak the OBLs in Guimond you can end up with pre-NLL cube states almost all of the time . . . and for the rest you can do a 2 move setup to change this.
I think this could successfully dethrone EG as the go-to 2x2 method.
Example solves:
Scramble- F' U' F' U' R F' R2 U R
V+CO- U R' U2 F R' (5/5)
NLL- D R2 U' R2 B2 U' R2 U R2 (9/14)

Scramble- F' R2 F' R2 F R2 F2 R2 U'
V+CO- R U R' F' (4/4)
NLL- R2 F2 U' R2 U (5/9)

Scramble- F' U' R2 F' R U2 R2 U' R2 U'
z y
V+CO- U' R2 F2 U' F (5/5)
NLL- D R2 U' R2 B2 U' R2 U R2 U' (10/15)

Oh my.
As good as or better than EG in less than half the algorithms.

Nice, we could have this as the advanced form of HD! I think the biggest thing would be inspection and finding a Guimond case that makes a V. Seeing how short the algs are, one looking-while hard-wouldn't be impossible. When we start to make our final documents, wiki page, video tutorials, etc I think we should give this as a general progression assuming we keep this latest revision:

Beginner- Literally the same as the VOP method, solve a V, do a LOLS, and only one of 6 algs for NLL
Intermediate- Full HD, sounds like a little bit of a jump but its actually really quite easy
Advanced- Second method described above, although there must be a better way of describing and looking at this so it's easier for people to understand.

Let's start documenting and stuff, I have almost all of tomorrow off @Sue Doenim so I can not only finish off NLL but help in creating an official doc and explaining/finding algs for this new variation as a whole

 

[VenomCubing]

One idea, i call it ZZ-M, m standing for messy, or Mortensen, my last name. Eoline is done with any two yellow edges, same with the yellow edges in the blocks. Then coll is done, leaving the top and bottom edges to be permuted, with one algorithm. I've only found 1 alg, which is M2 U' M2 U' M2 U' M2 D M2 D M2 for two z perms. Does anyone have the full algset?

 

[Shiv3r]

Okay guys, I think we should have a MUWR list(Method Unofficial World record) list..
For example, I know that for Ribbon method the UWR is 10.5 I think.
but what is the MUWR for methods like HD?(I think it's sub-2) ZZ-4c? etc?

 

[Neuro]

Cool idea I had while thinking about 2GR

1: CP Line- Solve pair and insert LC while solving CP. No real preservation required so it's really efficient
2: F2B and Pinkie Pie setup- MUuRr and E technically
3: 2GOLL- 57 algs, super easy
4: LSE- Very short bc of Pinkie Pie setup

F2B probably takes a lot of moves but this could be a pretty good method because CP Line is actually quite easy to do as minimal tracking of corners while doing initial pair is required; and the last 2 steps are pretty much algorithmic (Only a few 4C cases are possible and BU makes it super fast/easy to recognize so it's basically like an alg from that point) Let me know your thoughts!

 

[Rubik's cubed]

Quick question: how do you generate algorithms?

 

[CeBeMind]

Quick question: how do you generate algorithms?

Cube Explorer