The New Method / Substep / Concept Idea Thread

[Berkmann18]

So, I'm sure this has been looked into before, but I can't seem to find evidence of it anywhere: what prevents one from inventing a 2step method for 2x2x2 where you orient everything, then permute everything?

Because it's a massive amount of algs to generate and no one really use Guimond enough to get that far into it.

 

[Teoidus]

How massive are we talking, like ZBLL massive or like, actually thousands and thousands of algs? Could it be doable?

 

[shadowslice e]

How massive are we talking, like ZBLL massive or like, actually thousands and thousands of algs? Could it be doable?

About 8!/16/2=1260 algs
or 630 if you include inverses.

 

[Attila]

How massive are we talking, like ZBLL massive or like, actually thousands and thousands of algs? Could it be doable?

In my opinion, enough to know the usual Guimond orientation algs,
and 2x14 algs for permutation.
Just to do "V pattern" on both opposite faces during orientation,
which will reduce the permutation cases.
A few useful algs here: https://www.speedsolving.com/wiki/index.php/VOP_Method

 

[DELToS]

Tree(3) 3x3 method

Okay so I made a 3x3 method a long time ago (actually it was in November 2015) and just finished it now. First of all the name, I named it Tree3 but remembered that Tree(3) was a number so I renamed it to that for fun.
So, the steps.

1) Solve F2L minus the cross

2) CELL (Corners of the Last Layer without regards to edges on the top and bottom layers)

3) Bottom layer edges

4) ELL

So now an example solve:

R2 F2 D B2 U' R2 F2 D R B' F U L' F' D2 L2 B2 U' L2 B'

Inspection: z2

1 - D2 U R' D F2 U Rw2 D2 R'

2 - Rw U R' U' Rw' F R F'

3 - M' U' M y' M' U2 M y' M' U M

4 - M2 U' M U M' U M' U' M' U' M' U M

Pros -
low move count (if you pretend M is 1 move)
fairly intuitive (for the first half)
1st step can be done very quickly and efficiently if you practice it a lot
very few rotations

Cons -
59 total algs, although you may already know some of them
it takes a lot of practice to get the 1st step really fast

 

[sqAree]

I think you don't necessarily have to do steps 3 and 4 as bottom layer edges, then ELL.
Just solve LEE Roux style similar to SSC but with EO first?

 

[supercavitation]

Okay so I made a 3x3 method a long time ago (actually it was in November 2015) and just finished it now. First of all the name, I named it Tree3 but remembered that Tree(3) was a number so I renamed it to that for fun.
So, the steps.

1) Solve F2L minus the cross

2) CELL (Corners of the Last Layer without regards to edges on the top and bottom layers)

3) Bottom layer edges

4) ELL

So now an example solve:

R2 F2 D B2 U' R2 F2 D R B' F U L' F' D2 L2 B2 U' L2 B'

Inspection: z2

1 - D2 U R' D F2 U Rw2 D2 R'

2 - Rw U R' U' Rw' F R F'

3 - M' U' M y' M' U2 M y' M' U M

4 - M2 U' M U M' U M' U' M' U' M' U M

Pros -
low move count (if you pretend M is 1 move)
fairly intuitive (for the first half)
1st step can be done very quickly and efficiently if you practice it a lot
very few rotations

Cons -
59 total algs, although you may already know some of them
it takes a lot of practice to get the 1st step really fast

This is a slightly less efficient PCMS.

 

[Shiv3r]

new method brainstorm

Speaking of alternative methods... anyone do petrus 2x2x2 the the cross to make Xcross really easy?
Anyone have a Hybrid method they use or think would be usable to make a method really streamlined? just shout it out here, Im all ears(or eyes, or... computer screens? Idk anymore)

 

[Teoidus]

Pika****** is a sort of Roux/Petrus hybrid, if that's the kind of thing you're looking for.

 

[Shiv3r]

hybrid method idea--Prius Method

Okay, so after learning Fridrich and Roux, I think I have come up with an interesting hybrid method. one that may even use fewer moves than Either one, and be faster. It is a combo of a roux-style LSE(cause Fridrich LL isn't too move-efficient) and a Fridrich F2L. If you have a Name for it(for now its called the Prius Method-- cause its a Hybrid)
steps:
Step 1:SemiCross during inspection, visualize how to solve a cross on the bottom layer(or left or top, etc.). Well, not exactly. First get two white solved pieces opposite each other(really easy to visualize during inspection). and then instead of solving the other two cross pieces, orient and place any yellow or white edge(the easiest one is usually best). This is nice because it often requires less time and less moves than a normal cross

Step 2:Intuitive Fridrich F2L Solve the F2L any way you want(or can)

Step 3: Orient last layer and permute corners Remember your 2-look OLL's and your Corner-switching PLLs(2, maybe 3 cases--easy to remember)

Step 4: Roux-style LSE Hold your cube as if you had just done Roux.Since you have oriented all the edges on the cube(That's why we did that 2-look OLL earlier, and that SemiCross-- so we skip the most annoying ministep of LSE) If you need help with this, go look at Gilles Roux's website Replacing LL with LSE does away with the fincky fridrich method last layer and also shortens the beginning steps, because two of the edge's position doesn't matter, just their orientation.

Variation: If you aren't bothered by the top corners not being permuted during LSE, you can to that corner permutation after you have done LSE. It might be faster, who knows?

 

[Aiminer357]

12POw! Method - Dumbest idea I thought of

Hello! I came up with another method to solve the cube and it is quite ridiculous.

Ask any non-cuber how do they think the cube is solved. They will most likely answer side by side, not layer by layer. I thought of this and I want to put it to the test. I tried solving the cube face by face and came up with this ridiculous method. It is called 12PO. 12POw! just makes it sound cooler.

12PO stands for:

1st face: Solve one face like solving First layer on LBL.
2nd face: Solve second face. Slightly harder but possible with a few basic algs.
Permutation: Permute the remaining five edges and two corners.
Orientation : Orient the remaining edges and corners.

The first and second step can be combined. You could do F2L but only insert two F2L pairs and insert the other corners faster. OR you could make 2x2x3 block and then build the two faces at the same time.

Permutation and Orientation can be swapped to make it easier (or not). I did permutation first because orienting pieces on different faces is hard(for ZZ users is a lot easier).

And finally the parity. To solve it you need to stand on a chair and punch the air while saying 1! 2! POW!

So my method works but not amazingly. Unless somebody wants to try it out? Maybe generate the algs for it? Improve on it?

I may work on it a bit more and do some practice runs. Who knows? It might be great.

 

[shadowslice e]

Okay, so after learning Fridrich and Roux, I think I have come up with an interesting hybrid method. one that may even use fewer moves than Either one, and be faster. It is a combo of a roux-style LSE(cause Fridrich LL isn't too move-efficient) and a Fridrich F2L. If you have a Name for it(for now its called the Prius Method-- cause its a Hybrid)
steps:
Step 1:SemiCross during inspection, visualize how to solve a cross on the bottom layer(or left or top, etc.). Well, not exactly. First get two white solved pieces opposite each other(really easy to visualize during inspection). and then instead of solving the other two cross pieces, orient and place any yellow or white edge(the easiest one is usually best). This is nice because it often requires less time and less moves than a normal cross

Step 2:Intuitive Fridrich F2L Solve the F2L any way you want(or can)

Step 3: Orient last layer and permute corners Remember your 2-look OLL's and your Corner-switching PLLs(2, maybe 3 cases--easy to remember)

Step 4: Roux-style LSE Hold your cube as if you had just done Roux.Since you have oriented all the edges on the cube(That's why we did that 2-look OLL earlier, and that SemiCross-- so we skip the most annoying ministep of LSE) If you need help with this, go look at Gilles Roux's website Replacing LL with LSE does away with the fincky fridrich method last layer and also shortens the beginning steps, because two of the edge's position doesn't matter, just their orientation.

Variation: If you aren't bothered by the top corners not being permuted during LSE, you can to that corner permutation after you have done LSE. It might be faster, who knows?

This is a method which is proposed by virtually every single new cuber at some point. Its less efficient than roux (due to lack of blockbuilding) with less fingertrickability and the OLL in the middle means more pauses for recognition than either roux or CFOP meaning you would need even higher tps as well as learning more algs than either. Also, the EO step in LSE is actually quite nice as it means you can do a lot of UL/UR shortening in it with far less moves than a standard OLL (and it's 2-gen. And lookahead is much nicer.)

Lastly, if you want efficiency, just build the Roux block. The "semi-cross" you descibe is what is done for second block anyway because it make lookahead slightly easier but it ideas efficient. In addition, non permuted corners would make it far less efficient than the method already is as that mean you would need to finish on a PLL so you would still have all the pauses of a 3-LLL CFOP solve as well as the algs. The only way you could really rectify that is if you used OLLCP and that is also a lot of algs (300+ IIRC).

Hello! I came up with another method to solve the cube and it is quite ridiculous.

Ask any non-cuber how do they think the cube is solved. They will most likely answer side by side, not layer by layer. I thought of this and I want to put it to the test. I tried solving the cube face by face and came up with this ridiculous method. It is called 12PO. 12POw! just makes it sound cooler.

12PO stands for:

1st face: Solve one face like solving First layer on LBL.
2nd face: Solve second face. Slightly harder but possible with a few basic algs.
Permutation: Permute the remaining five edges and two corners.
Orientation : Orient the remaining edges and corners.

The first and second step can be combined. You could do F2L but only insert two F2L pairs and insert the other corners faster. OR you could make 2x2x3 block and then build the two faces at the same time.

Permutation and Orientation can be swapped to make it easier (or not). I did permutation first because orienting pieces on different faces is hard(for ZZ users is a lot easier).

And finally the parity. To solve it you need to stand on a chair and punch the air while saying 1! 2! POW!

So my method works but not amazingly. Unless somebody wants to try it out? Maybe generate the algs for it? Improve on it?

I may work on it a bit more and do some practice runs. Who knows? It might be great.

I experimented with something similar a while ago and, while being a fun method with a potentially low movecount, the algs needed foe the last two steps are unfortunately pretty terrible and solving the the last 1x1x3 can be pretty hard to do intuitively. But cool idea nonetheless.

 

[Berd]

Square one method:

Cubeshape
Bottom layer
Cp Parity or PLL
Epll (1/2 the time)

 

[shadowslice e]

Square one method:

Cubeshape
Bottom layer
Cp Parity or PLL
Epll (1/2 the time)

Isn't this basically Roux 'n' screw?

Also, if you do PLL why would you do EPLL? Do you mean CPLL+parity?

 

[Berd]

Isn't this basically Roux 'n' screw?

Also, if you do PLL why would you do EPLL? Do you mean CPLL+parity?

I'll look into detail on Roux n' Screw, This was the intermediate version proposed, with full LL at around 44 LL algs with the beginner being around 8.

 

[Aiminer357]

Here is a link for a detailed explanation on this method. I will continue to update it when I have the time

https://docs.google.com/document/d/1buKjGjfWu4KF65IQJCwwsZn5Q6Nm2w231sryMMIBVd0/edit?usp=sharing

 

[willtri4]

Isn't this basically Roux 'n' screw?

Also, if you do PLL why would you do EPLL? Do you mean CPLL+parity?

This is more like an intermediate Baum-Harris LBL than Roux n Screw. And EPLL is for if you don't have parity so you don't have to learn all the parity plls

Also:

 

[Mohammadahmadi]

hi. imagine we can always predict a 14 move solution (made of two steps first step is 5move and second is 9move) for 2x2 cube.
what will be our average in this way?

 

[PurpleBanana]

hi. imagine we can always predict a 14 move solution (made of two steps first step is 5move and second is 9move) for 2x2 cube.
what will be our average in this way?

It depends how fast you turn.

 

[IQubic]

hi. imagine we can always predict a 14 move solution (made of two steps first step is 5move and second is 9move) for 2x2 cube.
what will be our average in this way?

This is called one-looking. Many advanced 2x2 solves already do this. Chris Olsen has a tutorial on it: https://www.youtube.com/watch?v=pfPqeBWqQT8