The New Method / Substep / Concept Idea Thread

[acrystal]

I recently had an idea for a method, here’s the steps:

Start out by solving the cross and the first 3 f2l pairs. Then you put the last unsolved pair in the FR slot, like this:
View attachment 51568
The solve a 1x2x2 block in the ULB corner, here’s a view of that from the back:
View attachment 51569
Then you solve the last 4 corners and the last 3 edges. I don’t know what the best way to do the last step would be, if anyone can help me with this or give me feedback it would be appreciated.

Average LS method
New 3x3 Method

 

[J-B]

Hello everyone,
I would like to propose a new method for solving the 3x3 Rubik's cube.
I'm kind of new to the topic, so I am open to suggestions and critics.
The method I invented is the BBMVA (Bawaj's Blockbuilding Method with VHLS Application), a medium-high competitive method focused on move efficiency, low rotation count (2-3 is the maximum I think) and high TPS during the last layer's solution. It should be affordable to a great part of cubers, but it also has an advanced variant for making it faster during competitions.

The method:
1) create a 2x3x1 block on the left side, facing down and in the back (Like a Roux starting block, but without the front pair.
2) Create another 2x2x1 square on the opposite side, equal to the previous one.
3) Intuitively insert the 2 remaining bottom edges
4) Making the last pair and using VHLS algorithms for orienting the top cross
5) Inserting last layer's corners with COLL algorithms
6) Permuting the edges with EPLL algorithms
[Algs count= 76]

ESTIMATED MOVECOUNT
45-55 moves

The advanced variant:
1) Steps 1 to 4 are the same
2) Completing the last layer using ZBLL
[Algs count= 525]

Thanks you for reading. I hope you'll enjoy it.

P.S.: I am also interested to study some algorithms for the third step, making it more mechanical than intuitive.

 

[Michael Wimberly]

Average LS method
New 3x3 Method

I was trying to avoid just another ls method

 

[J-B]

Is the problem the vhls?
If yes I am thinking to alternatives

 

[AK51]

The possibility of solving a 3x3 Rubik's Cube under 2 seconds - Optimizing block building

https://arxiv.org/pdf/2609.12946

Online software, no download required

3D Rubik's Cube Solver - a Hugging Face Space by AK51

Discover amazing ML apps made by the community

huggingface.co

Video:

 

[Nirab_Shreatha]

I created a new methord called psudo-FCL/psudo-FCE please review; the tutorial is in my vid:

 

[Adnylzz]

Hello guys,ive made a method tho i dont think its good ,so ill write the method here then if you guys can help optimize it then go ahead,i would love them,
Bridge method,
In this method for step one you make floating F2L then go on to insert the cross edges using M and U moves,then make the yellow cross,for last layer use any method but for the LL method for this is kinda bad,make the yellow cross ,then permute the corners then orient them then do EPLL,to orient the corenrs do the sexy spamike the LBL final step,fif you guys have any optimization,please let me know,i would want this method to be a good method

 

[J-B]

Hello guys,ive made a method tho i dont think its good ,so ill write the method here then if you guys can help optimize it then go ahead,i would love them,
Bridge method,
In this method for step one you make floating F2L then go on to insert the cross edges using M and U moves,then make the yellow cross,for last layer use any method but for the LL method for this is kinda bad,make the yellow cross ,then permute the corners then orient them then do EPLL,to orient the corenrs do the sexy spamike the LBL final step,fif you guys have any optimization,please let me know,i would want this method to be a good

Hello guys,ive made a method tho i dont think its good ,so ill write the method here then if you guys can help optimize it then go ahead,i would love them,
Bridge method,
In this method for step one you make floating F2L then go on to insert the cross edges using M and U moves,then make the yellow cross,for last layer use any method but for the LL method for this is kinda bad,make the yellow cross ,then permute the corners then orient them then do EPLL,to orient the corenrs do the sexy spamike the LBL final step,fif you guys have any optimization,please let me know,i would want this method to be a good method

Hi, i would be happy to work with you on your method. i think the starting floating f2l is a grat idea, but a few days ago i thought about something similar and i was making the whole columns, and not just the f2l pairs. after this for me you could do two different things
1) you could insert the cross pieces with M and U (the top corners will rest at their place) and then making some kind of roux style EO, always starting from the 4 top misoriented position, and then making EPLL i think. later i'll try.
2) you can just orient all the edges, and not making the cross, and then making EPLL on both sides, like in belt method. i repeat i don't know if it's a good idea cause i can't test it, because i don't have a cube with me, if you wwant to try do it, if not i'll do this and i'll tell you.
Thank you for reading

method

 

[Adnylzz]

Hi, i would be happy to work with you on your method. i think the starting floating f2l is a grat idea, but a few days ago i thought about something similar and i was making the whole columns, and not just the f2l pairs. after this for me you could do two different things
1) you could insert the cross pieces with M and U (the top corners will rest at their place) and then making some kind of roux style EO, always starting from the 4 top misoriented position, and then making EPLL i think. later i'll try.
2) you can just orient all the edges, and not making the cross, and then making EPLL on both sides, like in belt method. i repeat i don't know if it's a good idea cause i can't test it, because i don't have a cube with me, if you wwant to try do it, if not i'll do this and i'll tell you.
Thank you for reading

Thanks for the suggestion,i would love to try both but its just i dont know hwo to do EO,So maybe could your try and help do it for me?

 

[J-B]

alright, i will try to test both ways, and i'll give you two example solves, so maybe you will understand better than only with words. by the way, you don't know how to do EO in this case or just you don't know what is it? because in the second case i can explain it to you, or you can just search a tutorial on youtube. As you wish. writing to you soon. if you will like the idea we could also algorithmize the whole thing.
I also posted an experimental Method, do you have some suggestions?

 

[Adnylzz]

So,i know what EO is but i dont understand it nor know how to do it,by the way,how did it go?

 

[J-B]

i am still working on it. it takes me a bit of time, because i just came back home

 

[Adnylzz]

Hello everyone,
I would like to propose a new method for solving the 3x3 Rubik's cube.
I'm kind of new to the topic, so I am open to suggestions and critics.
The method I invented is the BBMVA (Bawaj's Blockbuilding Method with VHLS Application), a medium-high competitive method focused on move efficiency, low rotation count (2-3 is the maximum I think) and high TPS during the last layer's solution. It should be affordable to a great part of cubers, but it also has an advanced variant for making it faster during competitions.

The method:
1) create a 2x3x1 block on the left side, facing down and in the back (Like a Roux starting block, but without the front pair.
2) Create another 2x2x1 square on the opposite side, equal to the previous one.
3) Intuitively insert the 2 remaining bottom edges
4) Making the last pair and using VHLS algorithms for orienting the top cross
5) Inserting last layer's corners with COLL algorithms
6) Permuting the edges with EPLL algorithms
[Algs count= 76]

ESTIMATED MOVECOUNT
45-55 moves

The advanced variant:
1) Steps 1 to 4 are the same
2) Completing the last layer using ZBLL
[Algs count= 525]

Thanks you for reading. I hope you'll enjoy it.

P.S.: I am also interested to study some algorithms for the third step, making it more mechanical than intuitive.

Its decent enough,and i find it good

 

[J-B]

alright, i tested the two variants of the bridge method and they work. I don't know how well, because i am not familiar with floating f2l and i made a lot of moves. so give me some time to make the example solves and i'll send them to you

 

[J-B]

Its decent enough,and i find it good

talking about my method, i thought about eliminating VHLS and instead during EO at step three while solving the bottom and then making the pair and applicating WV, so i'll have an OLL skip.

 

[J-B]

Hello everyone, i have tried to develop a new LS method for 3x3 Rubik cubes. The method is a Mehta-APB variant, that has an extremely lower algorithms count than the traditional LS methods like ZB or APB and so going.
The method starts like in APB, so with a block on the left.
The steps in order are:
1)Create a 2x2x3 block on the left side, exactly like in APB
2)Insert the dBR pair and the FR edge
3)EO, using the semplified Mehta algorithms (11)
4)LXS while doing CP (The number of algorithms needed is 68, because inserting the FR edge you will need just 17 standard algorithms, and there are 4 possible corners patterns, so there are only 68 algorithms for LXS+CP)
5)Concluding the LL using 2GLL (85 algorithms)

Algorithms count: 164, that is way less than ZB, Mehta or APB method
Estimated movecount: 42-48 moves.

I am still developing the LXS+CP algorithms
Thank you for reading

 

[AJT17]

Hello everyone, i have tried to develop a new LS method for 3x3 Rubik cubes. The method is a Mehta-APB variant, that has an extremely lower algorithms count than the traditional LS methods like ZB or APB and so going.
The method starts like in APB, so with a block on the left.
The steps in order are:
1)Create a 2x2x3 block on the left side, exactly like in APB
2)Insert the dBR pair and the FR edge
3)EO, using the semplified Mehta algorithms (11)
4)LXS while doing CP (The number of algorithms needed is 68, because inserting the FR edge you will need just 17 standard algorithms, and there are 4 possible corners patterns, so there are only 68 algorithms for LXS+CP)
5)Concluding the LL using 2GLL (85 algorithms)

Algorithms count: 164, that is way less than ZB, Mehta or APB method
Estimated movecount: 42-48 moves.

I am still developing the LXS+CP algorithms
Thank you for reading

You might consider sharing this in the APB discord server. They might be able to help generate the algs, or anything else you might need.

Join the APB Discord Server!

Check out the APB community on Discord - hang out with 1287 other members and enjoy free voice and text chat.

discord.gg

But this seems like an interesting way to tackle mid-solve CP, similar to Epidote. Actually, it's pretty much Epidote, except instead of solving the DFR corner while doing EO, you solve the FR edge, and EO, this could probably be combined into one step after BRPair. Then Epidote solves CP and the DR+FR edges, where this method would solve DR+DFR.
So you might look into Epidote as well, it might give you some extra insight into how a method like this could work. I've also attached the discord invite for the Epidote server, though it's not very active, it's still good for some resources. Also, here's the megadoc for Epidote, which you can look at as well.

Epidote MEGA Doc

docs.google.com

Join the Epidote Official Server Discord Server!

Check out the Epidote Official Server community on Discord - hang out with 98 other members and enjoy free voice and text chat.

discord.gg

I know a bit about APB and Epidote, as well as CEOR, so if there's anything you need help with, I might be able to help.

 

[J-B]

really thank you, i'll give a look to the links you attached to the maiil

 

[theail03]

New method idea: Using Pinkie Pie roux variant with CFOP instead

There is a roux variant called pinkie pie:

I think this variant can be applied to CFOP.
The idea is to place the UL/UR edges while doing the CFOP cross.
Here is an example solve to show what I mean:

B R2 B' D2 F D2 F U2 B' D2 F' R D L' F L2 D2 F U F2 // Scramble

x2 // Inspection
R' D F' L R' D // Pseudo-cross
U2 R U' R' d L U' L' // 1st pair
U D' R U' R' U L' U' L u // 2nd and 3rd pair
L U L' U' L U L' // 4th pair
U' F R U R' U' R2 D R' U R D' R2 U' F' // OLLCP
U2 M2 // LSE 4b
U M U2 M // LSE 4c

I think this method is better than CFOP because finishing LSE after OLLCP is way more efficient than PLL, you have more room for creativity at the beginning of the solve and many OLLCP algs are as good as their standard OLL alg.
Also OLLCP algs are easier to recognize and look better to execute than ZBLL algs to me.
That's why I think this method can also compete with the ZB method while having less than half the amount of algs.

 

[J-B]

For B-Mehta (the name I gave to my Mehta APB variant) you can also use the standard 9 LXS cases where the FR edge is inserted and then coll and epll