The New Method / Substep / Concept Idea Thread

[PapaSmurf]

The reason ZZ solvers can't do it is because most of us can't plan FB very well. I average about 10 moves instead of 7 and even then, on the rare occasion I can do it. The best LEOR solvers aren't very good because there are probably 3 or fewer people who use LEOR.

 

[ProStar]

except no one can plan LB+EO right now, its not realistic. The best leor solver i know cant do it, zz solvers cant do it

FB+Square isn't realistic either but Kian's doing it. Limits are always beyond what's thought possible

 

[whaffle]

This isn't viable because LS->OLL->PLL. Recognition for all 3 steps is super fast, only 78ish algs. Also you wouldn't use VLS, but a seperate alg set. VLS does OLS with the pair made. Also you have to account for the split pair. Looking for 2LLSLL is useless unless it's a different 2LLL method, because CFOP is so fast already. I'd recommend looking at 2LLSLL or 2A1LLSLL

Are there any 2LLSLL methods other than ZB? Also, are there any 2A1LLSLL methods other than M-CELL?

except no one can plan LB+EO right now, its not realistic. The best leor solver i know cant do it, zz solvers cant do it

lol i just tried and i planned fb+eo in under 15 seconds. there are many people who can plan fb in under 7 seconds, so with some practice, i think planning fb+eo+dfdb under 15 seconds is possible.

 

[ProStar]

Are there any 2LLSLL methods other than ZB? Also, are there any 2A1LLSLL methods other than M-CELL?

Yes. VLS/HLS->PLL, PLS->POLL, WV/SV->CP+EO+EP. I don't know a lot about 2A1LLSLL methods, but there are a few I think. And the less versions the more likely there's a big improvement to be made

 

[whaffle]

Yes. VLS/HLS->PLL, PLS->POLL, WV/SV->CP+EO+EP. I don't know a lot about 2A1LLSLL methods, but there are a few I think. And the less versions the more likely there's a big improvement to be made

Isn't that 3LLSLL? You have to make the pair, and then do 2 more algorithms.

 

[ProStar]

Isn't that 3LLSLL? You have to make the pair, and then do 2 more algorithms.

Depending on your definition, it could be. You could always just make the pair and insert in one alg. The main point though is that no current methods are viable, so if you're looking to develop LSLL I'd reccomend that area, because 3LLSLL is already so good

 

[RedstoneTim]

lol i just tried and i planned fb+eo in under 15 seconds. there are many people who can plan fb in under 7 seconds, so with some practice, i think planning fb+eo+dfdb under 15 seconds is possible.

What is your approach to FBEO? Do you plan FB and the EO separately or try to influence FB pieces during EO/EO during FB? I'm really interested in that since it's very hard for me to consistently plan even something like square + EO.

 

[whaffle]

What is your approach to FBEO? Do you plan FB and the EO separately or try to influence FB pieces during EO/EO during FB? I'm really interested in that since it's very hard for me to consistently plan even something like square + EO.

Separately. I don't try to influence EO. In fact, I try to preserve EO as much as possible, which results in a less effecient FB. My guess is that my FBEO is ~13-17 moves. My main method is Roux, and I'm not that good at LEOR, so take my advice with a grain of salt.

 

[brododragon]

therefore it has worse lookahead, pretty obvious

So CFOP is better than Roux because you solve less pieces at a time, meaning better look-ahead.

 

[ProStar]

So CFOP is better than Roux because you solve less pieces at a time, meaning better look-ahead.

pretty obvious

It's just a prank bro

 

[Silky]

So I've been thinking about PCMS lately and believe I've come up with an interesting variant.

Overview:

Step 1: Place 2 corner edge pairs on the left face and finish the E-slice.
Step 2: SOAP method for corners (https://www.cubestuff.cf/?soap).
2a: Separate corners to their respected faces.
2b: Orient and permute last 6 corners while preserving the E-slice.
Step 3: Permute centers while placing LD and RD edges.
Step 4: Orient remain 6 edges while placing DB edge.
Step 5: Permute last 5 edges (L5EP).

Pros:

(1) Inspection: It seems realistic to plan out all of step 1 in inspection and possibly step 2a.
(2) Corners: Being that the corners are solved using a modified 2x2 method (SOAP) it should be more efficient. It also should maintain the freedom of the pairs of PCMS.
(3) Algs: There wouldn't be too many algs (SOAP has 58, L5EP has 16) and the SOAP algs are mostly 2-gen.
(4) Because this method borrows from the SOAP method it should be clean af. #CoronaCubing

Cons:

(1) SOAP isn't really a proven 2x2 method so it just might not be that fast.
(2) The potential is very reliant on what you can plan in inspection.
(3) It just may not be worth it to preform corners compared to how PCMS does.
(4) Corners may take 2 looks instead of 1 (which will probably make it not that relevant).

If this has already been preposed, sorry. If not then let me know what you guys think.

 

[brododragon]

Ok, no more Muke tyranny, let's actually be productive.

New method:
1. Wheels 2 1x2x2 at DBR and DBL.
2. EOSlice solve EO and M-Slice. Make sure crosses are solved.
3. EP Permute UL, UR, FL, FR.
4. F3C First 3 Corners - solve DFR, DFL, UFL - 162 algs.
5. L3C Last 3 Corners - 24 algs, but they are included in F3C.

162 algs. Solving CO parity, then F3C + L3C will drop down the number to 54.

 

[PetraPine]

FB+Square isn't realistic either but Kian's doing it. Limits are always beyond what's thought possible

I have a 8 move av fb and am ok at eo, ill try to learn this method to see if i can inspect fb+eo

 

[mukerflap]

FB+Square isn't realistic either but Kian's doing it. Limits are always beyond what's thought possible

when did i say that

So CFOP is better than Roux because you solve less pieces at a time, meaning better look-ahead.

Lookahead is not the only factor, i also listed R and M switching, and worse cases

 

[brododragon]

Lookahead is not the only factor, i also listed R and M switching, and worse cases

...which you also do in roux.

 

[ProStar]

when did i say that

I said that limits are beyond what people think they are

 

[whaffle]

What if you did phasing, like in ZZ-b, but with CFOP? You would put two opposite LL edges opposite of each other, not necessarily correctly oriented, but in the same orientation. This would reduce the number of LL cases from 3916 to ~600. Phasing is mostly intuitive, and only has 6 cases. Though this is technically a 3LLSLL, I think it is better then other 3LLSLL methods, such as OLL/PLL, as phasing is relatively lightweight. What do you guys think?

 

[Skewbed]

What if you did phasing, like in ZZ-b, but with CFOP? You would put two opposite LL edges opposite of each other, not necessarily correctly oriented, but in the same orientation. This would reduce the number of LL cases from 3916 to ~600. Phasing is mostly intuitive, and only has 6 cases. Though this is technically a 3LLSLL, I think it is better then other 3LLSLL methods, such as OLL/PLL, as phasing is relatively lightweight. What do you guys think?

I think the alg count would be 1327 instead of around 600.

COLL (42) * 4 + H-Perm (1) = ZZLL (169)
OLLCP (331) * 4 + Mis-oriented H-Perms (2) + H-Perm (1) = 1327

I think it could be a good method, but ZBLS + ZBLL is better for that amount of algs because it is 2 look.

 

[whaffle]

I think the alg count would be 1327 instead of around 600.

COLL (42) * 4 + H-Perm (1) = ZZLL (169)
OLLCP (331) * 4 + Mis-oriented H-Perms (2) + H-Perm (1) = 1327

I think it could be a good method, but ZBLS + ZBLL is better for that amount of algs because it is 2 look.

Since the phased edges are in the same orientation, the remaining two edges are in the same orientation as well. Thus there are 4 eo possibilities: f/b unoriented, l/r unoriented, all unoriented, and none unoriented. Thus the total number of algorithms will be exactly 4 times ZZLL, which is 169 algs. Thus this method has 4 * 169 + 6 = 682 algs, compared to ZB's 799. The f2l pair and phasing take less than 1 look, as they are simple and easy to plan through lookahead. Thus in reality, this method takes ~2.5 looks or less, if you understand what I mean. On the other hand, each of the steps of ZB have pretty complicated recognition, and will take 2 looks. Thus I think the total recognition time for my method and the ZB method will not be too different.

 

[ProStar]

Since the phased edges are in the same orientation, the remaining two edges are in the same orientation as well. Thus there are 4 eo possibilities: f/b unoriented, l/r unoriented, all unoriented, and none unoriented. Thus the total number of algorithms will be exactly 4 times ZZLL, which is 169 algs. Thus this method has 4 * 169 + 6 = 682 algs, compared to ZB's 799. The f2l pair and phasing take less than 1 look, as they are simple and easy to plan through lookahead. Thus in reality, this method takes ~2.5 looks or less, if you understand what I mean. On the other hand, each of the steps of ZB have pretty complicated recognition, and will take 2 looks. Thus I think the total recognition time for my method and the ZB method will not be too different.

ZB doesn't have complicated recognition. It's probably easier than your method, although I can't say for sure.