The New Method / Substep / Concept Idea Thread

[DGCubes]

So then won't anyone else steal the method and pass it of as their own?

If you made an actual method that hasn't been used before, you can name it after yourself and people will credit you for it. Misnamed methods happened more in the past (e.g. Ortega/Varasano, Fridrich/CFOP), but most new noteworthy methods are accurately named (e.g. ZZ-CT, SSC, etc.). As far as I'm aware though, you can't patent a cubing method.

That being said, what is the 3x3 method your friend came up with?

 

[tseitsei]

I have created a new 2x2 method and my friend has created a new 3x3 method. How do I confirm that it is indeed a new method and how do I patent it or something?

1. It most probably isnt new
2. Nobody really patents methods we just usually share our inventions with the community
3. If you want some feedback on the method, feel free to post your method here

 

[shadowslice e]

2. Nobody really patents methods we just usually share our inventions with the community

Well, unless your name if Chris Mowla, Tony Snyder or Unsovled and they seem really popular.

All joking aside though OP, if it really is new, you will get recognition and it won't get "stolen". And if it is good people may start to use it though they will not claim to be the creators. However, if you don't publish, there is a real possibilty that someone else may propose it and you will not get the credit.

 

[tseitsei]

Well, unless your name if Chris Mowla, Tony Snyder or Unsovled and they seem really popular.

Yeah those guys are quite amusing really

Them: "I have created a really groundbraking method for solving rubiks cube!"
Community: "thats awesome. Show us how it works please gj!"
Them: "nah I dont wanna. Unless I get money or someyhing. You'll just have to believe me"
Community:"....yeah well....."
End of story

 

[Sion]

I'm currently creating a substep for PCMS, where if you permute two edges incorrectly to save time, you can use an algorithm to still solve the cube. I might work on a CFOP variation so if you could create a cross yet some of your edges are incorrect, you can fix them later.

In this algorithm set, each case is named after the pattern it represents, going L F U B with the cube orientation with yellow on top and green in front.

O- all colors on the face are matching
A- the D edge color isn't matching the rest of the colors on the face.
α- the U edge color isn't matching the rest of the colors on the face.
H- Both the U and D edge colors don't match the rest of the colors on the face.

I call them SLE (Sion Last Edges).

Here is one example algorithm: OAHα: R2 U (R2 F2)3 U' R2

https://alg.cubing.net/?setup=z2_R2_U_(R2_F2)3_U-_R2&alg=R2_U_(R2_F2)3_U-_R2

 

[Sion]

And here is another cool tip:

some cases are actually EPLLs on both sides if structured in a certain way. I will work on the cases that can't be solved with dual EPLL first.

 

[Sion]

Here are a few extra algs for demonstration of concept:

OOOO- The cube is solved in the SLE-OOOO state.

αααα/AAAA: this is either an H or Z perm.

Oααα/OAAA: this is a U perm.

OAHα: R2 U (R2 F2)3 U’ R2

HHαα: U (R2 F2) R2 U R2 U (F2 R2) B2 D B2

OHHO: R2 U (R2 F2 R2) U2 (F2 R2 F2) U R2

HHHO: U2 R2 B2 D' F2, B2 U (F2 R2) U2

HAαO: L2 U2 (F2 R2) B2 D' L2 B2 R2 U L2

HHαH: R2 U2 (R2 F2) U B2 (R2 F2) D B2 D2 L2

HHHH: R2 L2 U2 R2 B2 U' F2 B2 D F2 L2 D2

AOHα: F2 U F2 R2 B2 D' L2 B2 R2 U2 F2

 

[gyroninja]

Here are a few extra algs for demonstration of concept:
-Snip-

Someone has already created this if I understand what you are doing (EPLL on U and D at the same time). The algs for this are usually pretty bad.

 

[Cryoo]

I recently thought about a hybrid method that involves blockbuilding so it is move-efficient, not that high algcount and very ergonomic because of no rotations and mostly roux gen <R, U, M, r, F>.

As said before, it's a hybrid method of two awesome methods that are M-CELL and ZZ-CT.

For me, it balances the pros and cons of the two methods, the algs are already gen and the movecount is lower than ZZF2L. It also uses the ZZ-CT "LL".

Here are the steps :

1. Roux FB
2. 2x2x2 pair in BDR using mainly U, R, Rw and M turns as previously said.
3. Pseudo CE pair with the last E-edge
4. TSLE/OCLL
5. EO L5E.
6. TTLL

Pros :

1. Ergonomic (mostly R, Rw, M and U turns)
2. Quickly eliminates hidden pieces
3. Low movecount compared to ZZF2L
4. No abstract concepts like the EOLine (so is direct solving)
5. Good lookahead
6. Not too high an alg count (<100- around the same as CFOP, but TTLL is mostly just conjugated PLL and really really easy algs)
7. Fast turning is definately possible with the method
8. No rotations
9. There are many options for the various ways the F2L can be constructed as it is a relatively easy shape that does not have too many pieces getting in the way of others (such as in a full F2L)
10. Algs are already generated so not much work to do
11. Can be faster than ZZ-CT because of movecount
12. L5E is easy af

Cons :

1. Nubs p) may find it difficult to blockbuild (but this is the same with many intuitive things such as cross, F2L etc).
2. 100 algs may still be more than some people may want to learn (but it's not much for a 1LLL if your lookahead gets good enough and definately less than the other 1LLL such as ZBLL
3. Lack of cubers experienced at using the method (unavoidable with any new method really, hopefully this will change)
4. Lots of steps

Example solves (Don't notice my inefficiency in blocks)

U L2 D2 R' B2 L2 U2 F2 R2 U2 R' B D F' D2 B' L R2 F' R'

z2
U l' R U2 r' U2 r' F u' M' u // Roux FB
R U M U2 R' U R M U2 M' U2 M2 // BDR 2x2x2
R U R' U' R U2 R' U' R U R' // Pseudo CE Pair
R U2 R' U' R U' R' // OCLL
U' M' U M U M' U M U2 M' U2 M // EO L5E
U2 D R' U2 R U2 R' U R U2 R2 D' r U2 r' D R2 D' U2 // TTLL

bad 72stm

alg.cubing.net

 

[Rnewms]

[U2] D R' U2 R U2 R' U R U2 R2 D' r U2 r' D R2 D' [U2] // TTLL

[U'] F2 (R U' l) D2' (l' U l') U D f2 U' D [f2]

 

[Cryoo]

[U'] F2 (R U' l) D2' (l' U l') U D f2 U' D [f2]

prefer my alg

 

[wir3sandfir3s]

I recently thought about a hybrid method that involves blockbuilding so it is move-efficient, not that high algcount and very ergonomic because of no rotations and mostly roux gen <R, U, M, r, F>.

As said before, it's a hybrid method of two awesome methods that are M-CELL and ZZ-CT.

For me, it balances the pros and cons of the two methods, the algs are already gen and the movecount is lower than ZZF2L. It also uses the ZZ-CT "LL".

Here are the steps :

1. Roux FB
2. 2x2x2 pair in BDR using mainly U, R, Rw and M turns as previously said.
3. Pseudo CE pair with the last E-edge
4. TSLE/OCLL
5. EO L5E.
6. TTLL

Pros :

1. Ergonomic (mostly R, Rw, M and U turns)
2. Quickly eliminates hidden pieces
3. Low movecount compared to ZZF2L
4. No abstract concepts like the EOLine (so is direct solving)
5. Good lookahead
6. Not too high an alg count (<100- around the same as CFOP, but TTLL is mostly just conjugated PLL and really really easy algs)
7. Fast turning is definately possible with the method
8. No rotations
9. There are many options for the various ways the F2L can be constructed as it is a relatively easy shape that does not have too many pieces getting in the way of others (such as in a full F2L)
10. Algs are already generated so not much work to do
11. Can be faster than ZZ-CT because of movecount
12. L5E is easy af

Cons :

1. Nubs p) may find it difficult to blockbuild (but this is the same with many intuitive things such as cross, F2L etc).
2. 100 algs may still be more than some people may want to learn (but it's not much for a 1LLL if your lookahead gets good enough and definately less than the other 1LLL such as ZBLL
3. Lack of cubers experienced at using the method (unavoidable with any new method really, hopefully this will change)
4. Lots of steps

Example solves (Don't notice my inefficiency in blocks)

U L2 D2 R' B2 L2 U2 F2 R2 U2 R' B D F' D2 B' L R2 F' R'

z2
U l' R U2 r' U2 r' F u' M' u // Roux FB
R U M U2 R' U R M U2 M' U2 M2 // BDR 2x2x2
R U R' U' R U2 R' U' R U R' // Pseudo CE Pair
R U2 R' U' R U' R' // OCLL
U' M' U M U M' U M U2 M' U2 M // EO L5E
U2 D R' U2 R U2 R' U R U2 R2 D' r U2 r' D R2 D' U2 // TTLL

bad 72stm

alg.cubing.net

1. F2L-2x2x1 at FDR
2. TSLE
3. 3FCE style EO (I have algs)
4. TTLL
ZZ-CT is better, step 3 just ruins it and places ZZ on top.

 

[Shiv3r]

I just realized that there are literally only 2 methods for square-1. I am not inclined to learn either, does anyone want to help me brainstorm a new square-1 method?

 

[shadowslice e]

I just realized that there are literally only 2 methods for square-1. I am not inclined to learn either, does anyone want to help me brainstorm a new square-1 method?

Actually there are more but the are very hypothetical. There's Vandenberg, Roux 'n' screw, Cubeshape last, Cubeshape parity, Cp parity, an EGish thing and a hypothetical method which is pretty similar to ZZ-CT or maybe MGLS

 

[Shiv3r]

Actually there are more but the are very hypothetical. There's Vandenberg, Roux 'n' screw, Cubeshape last, Cubeshape parity, Cp parity, an EGish thing and a hypothetical method which is pretty similar to ZZ-CT or maybe MGLS

Cubeshape last? that seems interesting... would you please describe?
I had an idea similar to petrus where you blockbuilded and then did cubeshape. so like a direct solving method would be interesting... maybe like this:
1.blockbuild D layer(this can be splie up into any number of simpler steps)
2.cubeshape last layer(like OLL)
3.square-1 PLL

 

[wir3sandfir3s]

Cubeshape last? that seems interesting... would you please describe?
I had an idea similar to petrus where you blockbuilded and then did cubeshape. so like a direct solving method would be interesting... maybe like this:
1.blockbuild D layer(this can be splie up into any number of simpler steps)
2.cubeshape last layer(like OLL)
3.square-1 PLL

And, of course, do parity. I know next to nothing about Squan so don't quote me on anything I'm about to say.
Seems great. I'd use it. Algs seem like they'd be alright. *thumbs up*
Idk what else to say tbh. Last step recog may not be too good.

 

[Shiv3r]

And, of course, do parity. I know next to nothing about Squan so don't quote me on anything I'm about to say.
Seems great. I'd use it. Algs seem like they'd be alright. *thumbs up*
Idk what else to say tbh. Last step recog may not be too good.

It would be just as good as PLL recognition. cubeshape last layer's recognition would be even more insanely easy to recognize than OLL.
I need someone to generate the algs though, because I have no computers that run windows and can therefore run jaapsch's square-1 solver.

 

[Teoidus]

It would be just as good as PLL recognition. cubeshape last layer's recognition would be even more insanely easy to recognize than OLL.
I need someone to generate the algs though, because I have no computers that run windows and can therefore run jaapsch's square-1 solver.

Just use WINE to gen the algs

 

[Shiv3r]

I dont know of any wine for a chromebook, but Ill look into it.

 

[shadowslice e]

Cubeshape last? that seems interesting... would you please describe?

This is the thread that was posted a couple of months ago

I had an idea similar to petrus where you blockbuilded and then did cubeshape. so like a direct solving method would be interesting... maybe like this:
1.blockbuild D layer(this can be splie up into any number of simpler steps)
2.cubeshape last layer(like OLL)
3.square-1 PLL

What does the "OLL" look like? I don't think the algs will be very nice...