The New Method / Substep / Concept Idea Thread

[PetraPine]

I think you're both wrong xD cause there's a time where doing edge control doing F2L is more efficient and doing ZBLS is more efficient.

True

 

[TheSlykrCubr]

Not in most situations it literally isn't, last slot is restrictive making edge orientation worse depending on the case.

Depending on the case being the key words

 

[TheSlykrCubr]

is CPEOLL/2GLL worse than OLL/PLL

 

[Username: Username:]

In 2H, generally, yeah.

 

[TheSlykrCubr]

Anyway, do you think the wvcp/ell thing bad?

 

[Username: Username:]

Anyway, do you think the wvcp/ell thing bad?

Yeah, it's bad.

 

[TheSlykrCubr]

Yeah, it's bad.

same with the LEOR thing?

 

[Athefre]

Transformation Based Method

There are a lot of ways in which transformation can be applied. Corners, edges, groups of pieces, and more. All of the uses won't described here. It can be seen on the transformation wiki page. This method combines corner and edge transformation in a simple way to increase efficiency. There is also a lot of freedom for users to choose how in-depth they want to go with the application. Users can start with the easier forms of blockbuilding and, with experience, add to the skill set.

Step 1: Solve a 1x2x3 on the left.

Step 2: Solve any M-slice edge and any right side 1x2x2 and place on the lower layers.

Step 3: Solve any 1x2x2 and align above the pieces from the previous step. Perform an r or r' move to complete the transformed first two layers.

Step 4: LL then adjust.

Spoiler: Example Solves

Example 1
Scramble: D B L2 D' L2 D' B2 D B2 U B2 F2 D' F L' D R2 F2 R' D2
1x2x3: z2 F U2 r' D' U' r B
2x2x2: R U2 R' U' R' U2 R2 M' U2 M'
Transformed F2L: U2 M' U M U' r U r'
LL (CLL+1)

• CLL+1: U2 R' U' R U' R' U R' F R F' U R
• L3E: U M2 U M' U2 M U M2 U - r U'

Example 2
Scramble: U D2 R' D2 L2 B2 R B2 D2 L U2 R' U2 B' U R' F R2 D L D'
1x2x3: y2 x M2 D R2 D r' F
2x2x2: R U R' U R2 U' R' U2 r' U' M' U R'
Transformed F2L: U' r U' r' U r
LL (OLL+PLL):

• OLL: F U R U' R' F' U' F R U R' U' F'
• PLL: M2 U M2 U M' U2 M2 U2 M' U2 - r' U2

Example 3 (Advanced)
Scramble: U B2 D R2 B2 D L2 F2 L2 U R U2 F2 D U2 L' F L R
1x2x3: R U R' U' x U' M F
2x2x2: U2 R U' R' U' R U2 R2 U M2
Transformed F2L: U2 R U R' U' R U M'
LL (OLL+PLL):

• OLL: U2 M U R U R' U' M' R' F R F'
• PLL: R' U L' U2 R U' R' U2 R L U' - r' U2 r'

Example 4 (Advanced)
Scramble: L D2 R2 D' L2 B2 U2 B2 D F2 L2 U2 L' B' U2 F2 R' U' B' L
1x2x3: z2 x L D F2 r' B
Pseudo 2x2x2: R' U' M U r2
Transformed F2L: U2 M' U M U r
LL (CLL+1):

• CLL+1: U R U2 R2 U' R2 U' R2 U2 R
• L3E: U R' U' R U M U' R' U r U' - R' U2 R r2

Example 5 (Advanced)
Scramble: B L2 D2 B2 L2 F' L2 B' D2 B U' F2 R' D2 F D2 F U' B2
1x2x3: y2 x l' D B2 U' r' F'
Pseudo 2x2x2: U' r' U' M2 U R' M' U2 M
Transformed F2L: U2 R' U' M
LL (NMLL):

• Separation: U M2 U M' U2 M U M2
• Permutation: F R2 U' L' U R2 U' L U F' - M' U R U2 r' U2 R

Positives:

• Move-count
• Same good ergonomics as methods with similar shapes
• Lots of possible variations and options from which to choose

Negatives:

• LL recognition is more difficult

Notes:

• Edges can be oriented along the way if the solver desires to use a certain LL method. NMLL is particularly good because recognition is the easiest and the move-count is low. Orienting edges would also make step 3 easier.
• The approach presented here doesn't have to be the only way. ZZ, Petrus, EOMR, Two 1x2x3s then DF+DB, and similar methods can be used to reach the same point. The move-count for some of these would likely be lower than the primary strategy presented in the example solves.
• Sometimes the 1x2x2 that is built in step 3 will match the 2x2x2 in step 2. This means there will be solves where the F2L isn't transformed. It's all about building the easiest or most efficient solution.
• If A3 is used in combination with this method, it is a perfect fit. The move-count then becomes even better.
• It can be taken to even more extremes such as misoriented corners or edges or UL or UR edges in the transformed F2L. Whether this is beneficial or not depends on how it's done. Recognition is of course different. Some transformed F2L types will lead to shorter solutions; others will require long adjustments at the end.

 

[Redcrosscfop]

New method proposal
1) EO223 : In BD, this can be done it a variety of ways. Like solving EOarrow or EO222 and then finish the block.
2) Slice : Finish E-Slice + 2-3 corners
Steps 3 and 4 can be swapped
3) L5EP : Permute the last 5 edges
4) LC : Last 4 or 3 corners using commutators

This method is just for fun, some variations of this could be faster, don't have an algorithm generating program rn, I also don't know how how to find the number of cases for each step, nor the avg number of moves. Does it have potential? And what can i fix about it. Let the hate begin lol.

 

[PetraPine]

New method proposal
1) EO223 : In BD, this can be done it a variety of ways. Like solving EOarrow or EO222 and then finish the block.
2) Slice : Finish E-Slice + 2-3 corners
Steps 3 and 4 can be swapped
3) L5EP : Permute the last 5 edges
4) LC : Last 4 or 3 corners using commutators

This method is just for fun, some variations of this could be faster, don't have an algorithm generating program rn, I also don't know how how to find the number of cases for each step, nor the avg number of moves. Does it have potential? And what can i fix about it. Let the hate begin lol.

this is just HK petrus/HK ZZ with Roux EO but worse because you do L5E before corners

 

[PetraPine]

Worse because
if corners were to be algorithimized they would be bad because you would have to keep edge orientation and edge permutaion

 

[Username: Username:]

New method proposal
1) EO223 : In BD, this can be done it a variety of ways. Like solving EOarrow or EO222 and then finish the block.
2) Slice : Finish E-Slice + 2-3 corners
Steps 3 and 4 can be swapped
3) L5EP : Permute the last 5 edges
4) LC : Last 4 or 3 corners using commutators

This method is just for fun, some variations of this could be faster, don't have an algorithm generating program rn, I also don't know how how to find the number of cases for each step, nor the avg number of moves. Does it have potential? And what can i fix about it. Let the hate begin lol.

Last X Corners type methods are usually bad, the corners are restricted cause the edges are already solved and typically, it's bad when it's this way.

 

[Redcrosscfop]

this is just HK petrus/HK ZZ with Roux EO

I completely forgot HK exists oops.

but worse because you do L5E before corners

You can do last few corners before L5E and it wouldnt mess anything up.

Worse because
if corners were to be algorithimized they would be bad because you would have to keep edge orientation and edge permutaion

Algs might be slightly better if you do corners before edges. I say in my post that steps 3 and 4 can be swapped.

Last X Corners type methods are usually bad

Nice to know.

I completely for got HK exists. So pretend like i didnt come up with that method and move on.
Edit: Maybe the LC algs can ignore EO?

 

[PetraPine]

Just do this but Corners before edge permutation its way better, more efficient
and there are already optimized algs.
Basically, EO
Petrus Block in Back
HK pairs
CLL
Roux style LSE.
good try tho,
if hk didnt exist this would be pretty cool

 

[Redcrosscfop]

if hk didnt exist this would be pretty cool

Thx i guess lol.

My idea was to make a method that started with EODRDL and I came up with this, except I remembered EOarrow being a thing, and EO222. So i just used that. I guess i'll come up with smth else. Thx for the feedback

 

[TheSlykrCubr]

Aight so basically this method is like the ZB method but closer to CFCE

1) Cross

2) F2L-1

3) ZBLS variant that orients corners (call it ROLS for now)

4) 1LLL subset for oriented corner cases (ROLL temp name)

 

[Username: Username:]

Aight so basically this method is like the ZB method but closer to CFCE

1) Cross

2) F2L-1

3) ZBLS variant that orients corners (call it ROLS for now)

4) 1LLL subset for oriented corner cases (ROLL temp name)

You aren't clear enough in step 3, does it solve only the last pair and the corner orientation? if so, it's defs not a ZBLS variant, if it orients the LL edges and LL corners while solving the pair, then it's OLS.

Also, if you're setting up to 1LLL subsets that have all the corners oriented, I think I've heard of this before.

 

[TheSlykrCubr]

You aren't clear enough in step 3, does it solve only the last pair and the corner orientation? if so, it's defs not a ZBLS variant, if it orients the LL edges and LL corners while solving the pair, then it's OLS.

Also, if you're setting up to 1LLL subsets that have all the corners oriented, I think I've heard of this before.

Yea I meant more ZBLS-esque.

Can't find anything like this on the wiki.

@Username: Username: If it doesn't exist anywhere, do you think it has potential?

 

[PapaSmurf]

This has been proposed before (not on the wiki, but on this thread). 503 algs for if the F2L edge is orientated in last slot, more if not. The final step has multiple names, the most common being COALL, but also ELLCP/CPELL.

 

[TheSlykrCubr]

This has been proposed before (not on the wiki, but on this thread). 503 algs for if the F2L edge is orientated in last slot, more if not. The final step has multiple names, the most common being COALL, but also ELLCP/CPELL.

Oof. Does it count as a new method is I say do F2L-1 however you want?