The New Method / Substep / Concept Idea Thread

[brododragon]

What method?

Just looked at it, and it need A LOT of work. This is the best version:
1. EOBand
2. Sort Corners & edges
(Parity)
3. ZBLL
(Rotate)
4. ZBLL

 

[RedstoneTim]

Oh, I see. It wasn't listed on https://www.speedsolving.com/wiki/index.php/List_of_Subsets though. Is there anywhere I can find a more complete list of subsets or methods?

Generally, the "List of x" pages are outdated because they don't automatically get changed when a new subset is added. If you want to see the full lists, have a look at categories because they update when a new page is added. For 3x3x3 substeps, see Category:3x3x3 substeps. For (almost) all categories, see Category:Contents.

 

[CodingCuber]

Idk if this method already exists but I’ve been working on it for a couple months now.
It has a decently low movecount but I’m bad at explaining stuff so ill try my best.

First, on any side, you solve all 4 F2L pairs(this can be done fast and easily because no need to worry about messing up cross)
Now, bring two opposite cross pieces to the top layer. Make sure they are in a line shape and then you can do M2 to bring them down into their places,.
Once you have solved those two cross pieces, bring the next two opposite pieces to the top layer and repeat the previous step but make sure to take note of the place their centres are in.
Now the rest is just OLL+PLL.

Sorry if that made no sense. I can make an example solve if it’s easier

 

[PapaSmurf]

That is columns. Check out PCMS or just do Roux.

 

[mukerflap]

A LEOR variant that I think may be a little more efficient:

LB
EO(not stripe)
RB
M-Slice
ZBLL

Basically you go straight from EO to RB and solve DF+DB later. For me, DF+DB takes a bunch of moves after doing EO, but with RB already solved it usually only takes 5 or so moves. RB is hindered much(if at all) because using M2 allows for pairing in the M-slice easily. Example solve:

Scramble: D' L' D2 L' U2 L' B2 D2 R2 B2 F2 R D2 F' R2 D' B L' B' U2

F' E' L' F D2 F2 S M S' // LB

x2 S' U S R' U F R F' // EO

r2 U M2 U2 R' U R U' M2 U2 R U R' // RB

U' M U2 l // Stripe

y2 U R' U R2 F2 R' U2 R' U2 R2 F2 R2 U2 // ZBLL

I'm bad at blockbuilding, so LB and RB could definitely be more efficient


EDIT: Did a couple example solves on other threads, am averaging mid 40s, I imagine you could get sub-50 during a speedsolve with some practice and advanced techniques

right block step is trash because you need to preserve the eo

 

[brododragon]

right block step is trash because you need to preserve the eo

It's literally the same as LEOR RB but more freedom.

 

[AlphaCuber is awesome]

A LEOR variant that I think may be a little more efficient:

LB
EO(not stripe)
RB
M-Slice
ZBLL

Basically you go straight from EO to RB and solve DF+DB later. For me, DF+DB takes a bunch of moves after doing EO, but with RB already solved it usually only takes 5 or so moves. RB is hindered much(if at all) because using M2 allows for pairing in the M-slice easily. Example solve:

Scramble: D' L' D2 L' U2 L' B2 D2 R2 B2 F2 R D2 F' R2 D' B L' B' U2

F' E' L' F D2 F2 S M S' // LB

x2 S' U S R' U F R F' // EO

r2 U M2 U2 R' U R U' M2 U2 R U R' // RB

U' M U2 l // Stripe

y2 U R' U R2 F2 R' U2 R' U2 R2 F2 R2 U2 // ZBLL

I'm bad at blockbuilding, so LB and RB could definitely be more efficient


EDIT: Did a couple example solves on other threads, am averaging mid 40s, I imagine you could get sub-50 during a speedsolve with some practice and advanced techniques

One of the best things about leor is the speed and fluidity of RB —> zbll I feel this ruins the flow and the lookahead with very little advantage.

 

[mukerflap]

It's literally the same as LEOR RB but more freedom.

Worse lookahead, worse cases, M2 and R move switching,

 

[brododragon]

Worse lookahead, worse cases, M2 and R move switching,

Ok you just have two more places to look.

 

[ProStar]

Worse lookahead, worse cases, M2 and R move switching,

Yeah bro it's like Roux, and we all know that sucks so much


It's just a prank bro

 

[mukerflap]

Yeah bro it's like Roux, and we all know that sucks so much


It's just a prank bro

all i said was that LEOR rb was better

Ok you just have two more places to look.

therefore it has worse lookahead, pretty obvious

 

[NevEr_QeyX]

It's just a prank bro

Hey! Only I can say that.

It may or may not be just a prank bro

 

[ProStar]

therefore it has worse lookahead, pretty obvious

Just track what pieces are there, a lot of people plan LB+EO so it's just tracking those during EO. Also I think the pro to this is that it's more efficient, although it may not be faster

 

[whaffle]

New idea for 3LLSLL methods similar to COALL or VH. After making the pair, use vls algs to force specific sets of 1llls. For example: After making the pair, use one algorithm from a set of ~72 to deorient all edges and orient 2 adjacent corners. This will force OLLs 18, 19, and 20. These OLLs, especially 20, are very symmetrical, resulting in an ~90 alg 1LLL. This method only has 162 algs, therefore it has less algs than COALL and VH, and should have similar movecount. With some though, I suppose smaller sets of OLLs could be forced, resulting in less algs, while maintaining similar movecounts to COALL and VH. Is this/these methods viable, and can anyone come up with a better 3LLSLL method with this general structure?

 

[teri2769]

Is this/these methods viable, and can anyone come up with a better 3LLSLL method with this general structure?

Its very hard to make a 3llsll better than last slot > 2lll because of how last slot can be done intuitively. It doesnt make sense to make new 3llsll methods since 2llsll isnt that much.

 

[teri2769]

use one algorithm from a set of ~72 to deorient all edges

also why would you disorient all the edges when orienting all the edges is the same amount of cases?

 

[whaffle]

also why would you disorient all the edges when orienting all the edges is the same amount of cases?

better recognition imo

 

[teri2769]

better recognition imo

what?

 

[ProStar]

New idea for 3LLSLL methods similar to COALL or VH. After making the pair, use vls algs to force specific sets of 1llls. For example: After making the pair, use one algorithm from a set of ~72 to deorient all edges and orient 2 adjacent corners. This will force OLLs 18, 19, and 20. These OLLs, especially 20, are very symmetrical, resulting in an ~90 alg 1LLL. This method only has 162 algs, therefore it has less algs than COALL and VH, and should have similar movecount. With some though, I suppose smaller sets of OLLs could be forced, resulting in less algs, while maintaining similar movecounts to COALL and VH. Is this/these methods viable, and can anyone come up with a better 3LLSLL method with this general structure?

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

 

[mukerflap]

Just track what pieces are there, a lot of people plan LB+EO so it's just tracking those during EO. Also I think the pro to this is that it's more efficient, although it may not be faster

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