[Video] New 3x3 method, still has no name

[qqwref]

What one COULD do:

1.) EOLine
2.) F2L Pairs
3.) COLL
4.) L6E, with edges ARREADY ORIENTED! =D

This is a cool idea. A similar one:
1) EO only
2) build roux blocks, ZZ style, M2 is allowed
3) COMLL
4) L6E (w/ edges oriented)

 

[Lucas Garron]

This is a cool idea. A similar one:
1) EO only
2) build roux blocks, ZZ style, M2 is allowed
3) COMLL
4) L6E (w/ edges oriented)

That reminds me of a trick from ZZ-TOP

How about:

1) EOLine, except the line may consist of any edges that belong either in the M slice or on U.
2) Roux blocks, no M2 needed.
3) COMLL
4) L6E with edges oriented

 

[ryo]

How about:

1) EOLine, except the line may consist of any edges that belong either in the M slice or on U.
2) Roux blocks, no M2 needed.
3) COMLL
4) L6E with edges oriented

Not bad, I was practicing that just now.
You just have to compare your EOline with your L6EO step in Roux to find out what is better.

 

[Kirjava]

You just have to compare your EOline with your L6EO step in Roux to find out what is better.

blocks are different too.

 

[Athefre]

I'm not trying to seem...uh.....arrogant? Honestly. But, it's very rare for me to see an idea that I haven't thought of. Many times a popular member will make a post or post a topic about a new method, it gets praised (even if just a little), and I feel a little upset that I haven't posted the thousands of method ideas and variations I've explored because I figured they were obvious ideas that have been mentioned and criticized before. But, I guess most people haven't turned that search into an obsession like I had, it's a good thing they haven't. And I guess it wouldn't be fair for me to claim everything.

Something new I like is Thom's CLL + 1 edge, L3E idea. That's interesting

 

[cubefan4848]

Has anyone thought of any name for this method yet.

 

[ryo]

blocks are different too.

Yes, just a little, but you're totally right. I think it's depend how you made them. I don't use Roux often, so it's just better for me because recognition is easier but for a Roux expert it may be longer (I mean for the move count).

 

[cubefan4848]

The orientation algs are up at:
https://docs.google.com/document/d/1OozsImk9rYSmvMPG935hnsVCcXGc1Zm119CyuLZtSuc/edit?hl=en

I just made a google doc because it was the quickest way to get things up

 

[irontwig]

L4C sucks for speed; as a rule of thumb for speed you want to leave edges and for FMC you want to leave corners.

 

[gogozerg]

Outline of method:
Four F2L pairs
CLL
Orient all edges
Permute all edges
OR
Four F2L pairs
Orient all edges
Permute all edges
L4C

Hum...
I see 2 rather different algorithms.
Many people thought about the first one that keeps 8 edges (and centers) unsolved for the end part. It looks interesting because edges sometimes can be efficiently managed in <U, D, M> without touching other pieces (think "last 6 edges") and symmetries are obvious, but it's hard to find a practical method, and it's difficult to locate pieces all around the cube. By the way, people proposed to do the first part through different means, for example 8 corners+4 edges or starting with E slice.

I will warn you before the example solve that the amount of moves in my solution will be higher than after I practice it and get a bit better
One colour example first
Scramble- D' F B' R2 L' U D' F U' D B L' F2 D' R2 L' B R2 F2 U' R2 D' R L' B2
x2 U' B U' R2 U' R2 U R2 U2 R U' y' M2 U' M2 R U2 R' U2 y' M' U M y' R U' R'
y U2 M' U M' U M U M
U' M' U2 M y M2 U M2 y
R' F2 U F2 U F2 U' F2 R
y' l' U R' D2 R U' R' D2 R2
57 Moves if you count M moves as one

I just got a normal 63 move solution using CFOP

Please propose an average of 12, 1 solve is not very eloquent.

z2x'R2U'x2Ur2B
r'U'R2'UrURU2RU'M2U'r
R'FRF'RU2R'U'F'U'F
M'UMU2MUMU'M2

Has anyone thought of any name for this method yet.

I think it was called "Dead End #56" ;-)

 

[oll+phase+sync]

Something new I like is Thom's CLL + 1 edge, L3E idea. That's interesting

What's the concept of CLL + 1 (didn' find the threat). Sounds like a lot of cases (42 times 8 is maybe to high as an upper bound )
Or is it anbout using MUM' conjugation?

2 opposite 1x2x2 blocks
1x2x3 block in the back or front.
CMLL EO(keeping the D layer edge correctly permuted)
L5E

"2 opposite 1x2x2 blocks" - if I were able to plan ahead this in 15 sec (even with non matching side colors ) in up to 12 moves, I would do it...

roux ending

"EO5 + L5E" - I'm doing this if I get an D edge accidentially soved during an Roux solve, others maybe too?

"1x2x3 block in the back or front" - that I'd hate because it means to refocus on a new color (not R/L color).

Does anyone think it is worth learning L4C for the last step

L3C is much better 23 cases but only 7 essetially differnet ones, also regognition is much easier, movecount much lower, and acceptance for your

methode maybe much higher.

P.S. When doing Placeing your last Slot altering between URU'R' , U²RU²R' (R'FRF') will always ensure at least on oriented corner in last layer, but

you have to try out if that might help you

Question: do You asume center pieces to be solved from the very beginning of the solve

I used to play with a method similar to this,

4 F2L pairs
CLL
Orient edges while placing D edges
U/H/Z perm

I was also trying this (did Center placement after CLL ), but never reached my friedrich times (I'm not even a regular friedrich solver)

 

[Kirjava]

What's the concept of CLL + 1 (didn' find the threat). Sounds like a lot of cases (42 times 8 is maybe to high as an upper bound )

42*8 is silly because we want any edge, not a specific one.

I've been trying to think of the best way to do it and haven't quite hit it yet - I estimate that it would be no more than ~80 algs though.

 

[Athefre]

Has anyone thought of any name for this method yet.

I think it was called "Dead End #56" ;-)

lol

 

[irontwig]

42*8 is silly because we want any edge, not a specific one.

I've been trying to think of the best way to do it and haven't quite hit it yet - I estimate that it would be no more than ~80 algs though.

Cases =/= Algs

 

[Kirjava]

42*8 is a silly estimation for either algs or cases.

 

[irontwig]

42*8 is a silly estimation for either algs or cases.

Upper Bound =/= Estimation

 

[Kirjava]

42*8 is a silly upper bound or estimation for either algs or cases

 

[Carrot]

42 is a silly upper bound or estimation for Kirjava =D

 

[cubefan4848]

Question: do You asume center pieces to be solved from the very beginning of the solve

They don't have to be but I normally do because I really suck at everything.

 

[cubefan4848]

New method ideas mean only L3C at the end easy to learn and execute

The start can now be four F2L pairs then orient
Or alternative start orient ZZ style then solve the four F2L pairs, this one means no rotation during first and second step.

Third step can be changed to intuitively make a 1x2x2 block on top. This can be made easier by different last F2L insertion and normally only takes about 3-6 moves but makes it L3C at the end after permutation is finished.

I suck at explaining things so if this is unclear hopefully I could make a "short" video