The New Method / Substep / Concept Idea Thread

[Shiv3r]

you could blockbuild the M slices maybe???

 

[TDM]

hey what if we did something like COL/ZZ-line on left and did FB on d face and SB on U face, then did CMLL on side then you can use E and R moves to solve LSE???
LOL nope this would suck pretty bad

Gilles suggests this as an option for F2B but then rotating for CMLL, which could work. However you have to rotate and then regrip before CMLL, and you also may not be able to recognise CMLL immediately since you can't see what's on the D face until after rotating.

 

[Shiv3r]

you guys are taking this sooo seriously. I was joking people.

 

[Kubzy]

This question goes out mostly to people who know old pocchman or Yau.
I was recently thinking of 3x3 methods and I though of Yau. I'm not knowledgeable at 4x4 but I know kind of how it works. This method would use like 3 steps and someone good at yau/4x4 could probably make this into a speed method.
The steps:
1: Do the centers(troll step 'cause there are no centers to be made on 3x3)
2: Do the edges, simple beginner method algs would probs work
U R U' R' U' F' U F
U' L' U L U F U' F'
3: Do the corners :/ Here is my problem, I don't solve 4x4 so I don't know how you put the corners in. I know in old pocchman you use like the Y and J perm, so I was thinking of doing something like that but with actually looking. If anyone knows if this would be possible then please tell me the algorithms.

SECOND METHOD FOR 2X2 - Pattern Method

This is a very simple method based on some LBL algs from 2x2 (That's why I'm making a 2x2 method). There are 4 simple steps to this method and I've tested it out.
1: Make the first layer, but instead of making all the colors match make it like a checkerboard, orange and red next to each other and blue and green next to each other while the white is done at the bottom.
2: Do your OLL on the opposite layer.
3: Do your PLL. In my method there are 2 cases (with 3 algs in total)
The horizontal case,
R U R' U' R' F R2 U' R' U' R U R' F'.
If you get the horizontal case, you'll have to perform the diagonal algorithm on the bottom to solve the cube.
The diagonal case,
F R U' R' U' R U R' F' R U R' U' R' F R F'
THAT ALGORITHM FOR DIAGONAL USE SHOULD ONLY USE IN THIS METHOD IF YOUR PLL CASE ON TOP WAS THE HORIZONTAL CASE.
If you get the diagonal case on your pll, match up the pieces together. If your pieces match, you can simply do R2 B2 R2 to solve your cube.



I sincerely hope someone can find out the solution to Yau on 3x3, and also that someone can give me feedback on the 'pattern' method on 2x2. Remember, this is my first post on this website so please don't go too harsh on me if I do something wrong or if you disagree.

 

[Rcuber123]

For the 3x3 method, what does that have anything to do with Yau and why do you think that method is good? For the 2x2 method, or you could just use ortega

 

[Kubzy]

M

For the 3x3 method, what does that have anything to do with Yau and why do you think that method is good? For the 2x2 method, or you could just use ortega

y 3x3 method is pertty similar to Yau, it's just that the execution is different because there's one less layer on each axis. I never said this method is good, I just wanted some feedback. For the 2x2, Veressano is a good method, and I know that, but maybe some person could master this method and get a WR/UWR, you never know. This time give me some constructive
criticism. Thank's for your reply?

 

[TDM]

the first one is just EF, and the second is beginner's anti-CLL (anti-LBL?)

 

[Teoidus]

For your edges first yau3x3-method, corners can be solved by using a bunch of triggers like triple sexy or triple sledge

 

[Sion]

Here's a substep idea that I've been toying with. Its called galaxy cross, and it's a new first layer formation to aide in solving.

https://alg.cubing.net/?setup=RUR-U-_L_U_L-U-_MUM2U-M2_U_M-_U

It consit's of two Center Edge and Corner Edge pairs. Because of that, a slice will be open for edge pairing and even for some other sort of cool ideas. Thats why I also call it 'Blockbuilder's Cross'

Here are a few ideas of how Galaxy Cross can be used as a first step in solving a cube:
GFOP: Same thing as Cfop, but with Galaxy cross instead of standard cross. This can be faster, since slice moves are open.

Galaxy Roux: Use Galaxy cross to start off F2B. Since there is less to permute, it's quicker to solve.

Crooked Roux: Solve the 1x2x3 blocks vertically, leaving two empty 2x2x3 spots.

Quasar: Create 1x2x3 blocks in the two open slots, and solve the last four cubies on the last layer in a final step.

 

[shadowslice e]

This question goes out mostly to people who know old pocchman or Yau.
I was recently thinking of 3x3 methods and I though of Yau. I'm not knowledgeable at 4x4 but I know kind of how it works. This method would use like 3 steps and someone good at yau/4x4 could probably make this into a speed method.
The steps:
1: Do the centers(troll step 'cause there are no centers to be made on 3x3)
2: Do the edges, simple beginner method algs would probs work
U R U' R' U' F' U F
U' L' U L U F U' F'
3: Do the corners :/ Here is my problem, I don't solve 4x4 so I don't know how you put the corners in. I know in old pocchman you use like the Y and J perm, so I was thinking of doing something like that but with actually looking. If anyone knows if this would be possible then please tell me the algorithms.

SECOND METHOD FOR 2X2 - Pattern Method

This is a very simple method based on some LBL algs from 2x2 (That's why I'm making a 2x2 method). There are 4 simple steps to this method and I've tested it out.
1: Make the first layer, but instead of making all the colors match make it like a checkerboard, orange and red next to each other and blue and green next to each other while the white is done at the bottom.
2: Do your OLL on the opposite layer.
3: Do your PLL. In my method there are 2 cases (with 3 algs in total)
The horizontal case,
R U R' U' R' F R2 U' R' U' R U R' F'.
If you get the horizontal case, you'll have to perform the diagonal algorithm on the bottom to solve the cube.
The diagonal case,
F R U' R' U' R U R' F' R U R' U' R' F R F'
THAT ALGORITHM FOR DIAGONAL USE SHOULD ONLY USE IN THIS METHOD IF YOUR PLL CASE ON TOP WAS THE HORIZONTAL CASE.
If you get the diagonal case on your pll, match up the pieces together. If your pieces match, you can simply do R2 B2 R2 to solve your cube.



I sincerely hope someone can find out the solution to Yau on 3x3, and also that someone can give me feedback on the 'pattern' method on 2x2. Remember, this is my first post on this website so please don't go too harsh on me if I do something wrong or if you disagree.

For the 3x3 method, it would simply be very inefficient without a large number of algs as you are essentially doing edges first which has been toyed around with before and is generally considered to not be great for speedsolving compared to popular current methods.

The only way you could do the corners really efficiently(ish) is by essentially doing 3-style which mean learning quite a few algs and acquiring an understanding if commutators which may not be worth the time as the method is less efficient even with all of these. Basically, it comes down to having to destroy and restore in order to solve more pieces. This is why SSC and Roux are more efficient than CFOP; you don't have to destroy and restore what you have built so much sp those extra moves are negated.

With regards to the 2x2 method, this is a part of a method called varasano/ortega (the name depends on who you ask) though it is more restrictive as you don't get the freedom to build a face which is one of the advantages of those methods.

Anti-CLL and EG could also be considered extentions af these concepts and those are what seethe best solvers (or at least the fastest solver) in the world use.

...
.

This is definately an option that is sometimes used in Roux, however, bar a lucky case, this is slower than the FB->SB as you have to track more pieces and it is less efficient.

Also, for CFOP, it would be more efficient almost definately but more efficient does not necessarily mean faster as you need to factor in lookahead

 

[Sion]

Although it may have slightly harder lookahead, wouldn't it be more beneficial in the long run once you get used to Galaxy Cross?

 

[Teoidus]

It's just double-square to start, which can be good but doesn't ensure efficiency. You can get some pretty ugly cases that take up to 8-9 moves because pairs get stuck in BDR/BDL

 

[shadowslice e]

Although it may have slightly harder lookahead, wouldn't it be more beneficial in the long run once you get used to Galaxy Cross?

Possibly though I am not entirely convinced it would be better overall. Pure FreeFOP is probably more advantageous as you aren't restricted in your blocks and doing FB->SB->M-slice is better lookahead and more efficient though it would be worth doing if you have a lucky case.

 

[Metallic Silver]

I use my own method, Lazy CFOP, but its only recommended for beginner cubers only. Its basically all methods combined while solving it in a CFOP state. So for example for us beginner cubers, if the cross it too hard to make, we either use roux, columns first, freeFOP, Salvia, Petrus, and etc. It all depends on the scramble. We dont stick with one method in every scramble tho, were basically turning a hard scramble into an easy scramble. So like after i do Roux which is making the 2 1x2x3 blocks, i could do CFCE after that instead of CMLL&LSE. Lazy CFOP only takes the easy parts of all methods and use it as Lazy CFOP tricks. There are a lot of examples and explanantion into this method. Another example is when there is a easy cross, we do the first method CFOP, then instead of F2L, we can rather do keyhole method or edges first, so if i cant do F2L since i cant find a pair quick enough (Lazy CFOP doesnt required lookahead), i can solve the equator slice first, then solve all corners easily, if the corners are in the correct place but not oriented correctly, we can do MGLS, 8355, or just solved the F2L pair by VLS, WV, or normal. Then at last layer we could do OLL, PLL, ELL, COLL, or whatever thats comfortable for the cubers to use. There is also SLL which means Special Last Layer, which is basically easy ZBLL, or easy 1LLL. Thats how lazy CFOP, we use all methods to help us get faster on every each step. We only use easy algorithms, but u can use hard algs if youre comfortable with it.

 

[AlphaSheep]

This question goes out mostly to people who know old pocchman or Yau.
I was recently thinking of 3x3 methods and I though of Yau. I'm not knowledgeable at 4x4 but I know kind of how it works. This method would use like 3 steps and someone good at yau/4x4 could probably make this into a speed method.

Firstly, welcome!

Secondly, some of the criticism you receive in this thread may seem harsh but you have to understand that almost every new cuber goes through a stage where they want to invent a new method, and every one stumbles on the same handful of ideas. Don't take it personally. A good place to start is to spend some more time understanding existing methods and what makes them good or bad.

Thirdly, keeping in mind my point above, you obviously have no clue how Yau works on a 4x4. Adapted to a 3x3, it would work like this.
1. Cross colour centre and opposite centre (automatically skipped)
2. Three cross edges.
3. Remaining 4 centres (automatically skipped)
4. Fourth cross edge
5. Edge pairing (automatically skipped)
6. CFOP finish with cross skipped.

So if we leave out the automatically skipped steps, it reduces to CFOP which is the most popular 3x3 method anyway. (In fact that's the whole idea of Yau)

The edge pairing step on 4x4 doesn't refer to actually solving the edges. There are two edges for each combination of colours on a 4x4 and the edge pairing step simply involves connecting the two, not necessarily placing them in their final position.

For your 2x2 method, what you propose is actually an improvement over the standard layer by layer method. You may want to look into the Ortega method, which goes one step further in that in your first layer, the side stickers can be any colour, not just opposite. It is one of the most popular 2x2 methods and is very easy to get fast with.

 

[SimonV]

Here's a substep idea that I've been toying with. Its called galaxy cross, and it's a new first layer formation to aide in solving.

https://alg.cubing.net/?setup=RUR-U-_L_U_L-U-_MUM2U-M2_U_M-_U

It consit's of two Center Edge and Corner Edge pairs. Because of that, a slice will be open for edge pairing and even for some other sort of cool ideas. Thats why I also call it 'Blockbuilder's Cross'

Here are a few ideas of how Galaxy Cross can be used as a first step in solving a cube:
GFOP: Same thing as Cfop, but with Galaxy cross instead of standard cross. This can be faster, since slice moves are open.

Galaxy Roux: Use Galaxy cross to start off F2B. Since there is less to permute, it's quicker to solve.

Crooked Roux: Solve the 1x2x3 blocks vertically, leaving two empty 2x2x3 spots.

Quasar: Create 1x2x3 blocks in the two open slots, and solve the last four cubies on the last layer in a final step.

I decided to make a walktrough solve with the cross solved already.
https://alg.cubing.net/?setup=RUR-U...easy_OLL U_F-_U-_L-_U_L_F_//_OLL_and_PLL_skip

 

[xyzzy]

I came up with this 4×4×4 direct solving method a few months ago. I don't think it's faster than sandwich (which it's based on), but it's certainly fun to play with.

Step 1: L and R centres. (Let's say white is on L and yellow on R.)
Step 2: Create a white edge pair and extend it to a 2×2 block, then attach it to the L centre. Do this for two more edge pairs.
Step 3: F2L-on-left except for the missing edge block we didn't solve in step 2.
Step 4: CLL-on-right. (Or z' CLL z.)
Step 5: R-layer edges (as in sandwich).
Step 6: Last two L-layer edges.
Step 7: r-slice edges.
Step 8: L4C. (Niklas spam!)

Sample solve.

Other than CLL, this pretty much needs only four algs, though using alternative algs for centre control helps, of course. Step 6 can be done with the slice-Sune as in the sample solve and step 8 can be done with Niklas. Step 7 can be done with just these easy-to-remember (but long) algs:

UBr-DBr swap: (U' R2 U 2R)2 (U' R2 U) 2R2 (U' R2 U)
UFr-DBr swap: (U' R2 U 2R)3 (U' R2 U) 2R' (U' R2 U) 2R2 (U' R2 U)

L4C gets pretty fast since around half of the centres are already solved (or even more if we use centre control in steps 6/7), though this comes at the expense of step 5 being quite heavily restricted.

 

[wir3sandfir3s]

So, if you guys are at all familiar with me by now you will know that I often propose strange and abstract methods/concepts here. This one is no exception, and I think it's by far my best. I kinda wanna keep this post short because I'm learning Algs currently.
1. FB, 2x2x1 in the back.
2. SB, 2x2x1 in the back on the other side.
3. Solve FR and FL
4. Roux style EO and make sure M slice is offset by an M2 for next step
5. LSSLL: Left Side Semi LL, solve the rest of the LL face and the 2 B M slice edges (which are now in front for better recog), while also resetting the M slice for next step, all in 1 alg.
6. RSSLL: Right Side Semi Last Layer, solve the rest of the R slice and the 2 F M slice edges. Cube solved.
The main strong point in the method, I believe, is recog because the first 2 steps solve most of the back part of the cube pushing everything you need to see in front. Also, no rotations (except maybe in Algs) .
Suggestions/criticism? This method is still a WIP which is why it's not too great yet.

 

[wir3sandfir3s]

So I don't know if this is already made or not and I don't know if this works or not because I don't have a cube nearby. Here it is:

1) you solve first layer by making edge-corner pairs and inserting.

2) you solve the E layer using E moves with the freedom of the unsolved last layer.

3) last layer using any alg you want.

This can also be done differently. This came to my mind first:

1) you leave one corner open in first layer

2) you leave one edge open in E layer

3) LSLL or something like that. It should solve the open slot and any substep of LL.

So does this exist already?

Yes, it's called LBL/Beginners method with a different, maybe less efficient way of doing edges.

 

[Teoidus]

So, if you guys are at all familiar with me by now you will know that I often propose strange and abstract methods/concepts here. This one is no exception, and I think it's by far my best. I kinda wanna keep this post short because I'm learning Algs currently.
1. FB, 2x2x1 in the back.
2. SB, 2x2x1 in the back on the other side.
3. Solve FR and FL
4. Roux style EO and make sure M slice is offset by an M2 for next step
5. LSSLL: Left Side Semi LL, solve the rest of the LL face and the 2 B M slice edges (which are now in front for better recog), while also resetting the M slice for next step, all in 1 alg.
6. RSSLL: Right Side Semi Last Layer, solve the rest of the R slice and the 2 F M slice edges. Cube solved.
The main strong point in the method, I believe, is recog because the first 2 steps solve most of the back part of the cube pushing everything you need to see in front. Also, no rotations (except maybe in Algs) .
Suggestions/criticism? This method is still a WIP which is why it's not too great yet.

Don't you have to solve the corners at some point?