The New Method / Substep / Concept Idea Thread

[Rcjas]

This method name is ALE. A-Avinash. L-Last. E-Edge

 

[xyzzy]

This is a zz method where we have to place edge to skip oll.
You have to make f2l-1 edge and place that edge face you and apply th alg for that specific case, so there are 26 algs.

Done before, and it's called Brooks Last Edge (BLE).

 

[Spencer131]

After the L/R step ( step 4b), I do a U2 of what normal cubers do. This places the white bar on yellow and vice versa. In L4E there will be no chance of the 'Dots' permutation and if there is a 'single dot' then there will be a L4E skip/ much lower movecount. This results in better cases and more skips.

Can you provide an example? I'm confused.

 

[Sue Doenim]

After the L/R step ( step 4b), I do a U2 of what normal cubers do. This places the white bar on yellow and vice versa. In L4E there will be no chance of the 'Dots' permutation and if there is a 'single dot' then there will be a L4E skip/ much lower movecount. This results in better cases and more skips.
This thread may be moved to the Concepts thread, fine by me.

The thing is, 4C (3rd step of L6E) starts as soon as you put in the two edges. It is usually then AUFed to solve L and R because those cases closer to solved. You're just arbitrarily adding in a U2. You're using the same dots algorithm, you're just doing the <M,U> version, cancelled into it. See the CFTOP method by @Tao Yu .

 

[Rcjas]

Done before, and it's called Brooks Last Edge (BLE).

I dont know about how many methods are there because i am new to cubing so i dont know about BLE.

 

[BECubed]

Is there any method for 4x4 that is kind of like CSP? I think it would be awesome to fix OLL/PLL parity errors during centers/edges, although, I'm not sure how possible that would be.

 

[Thom S.]

I dont know about how many methods are there because i am new to cubing so i dont know about BLE.

One of the rules of this thread is to do research, if your method has been proposed before.

Is there any method for 4x4 that is kind of like CSP? I think it would be awesome to fix OLL/PLL parity errors during centers/edges, although, I'm not sure how possible that would be.

Centers last

 

[xyzzy]

Is there any method for 4x4 that is kind of like CSP? I think it would be awesome to fix OLL/PLL parity errors during centers/edges, although, I'm not sure how possible that would be.

Determine wing parity during inspection, and solve the first three centres with the corresponding amount of quarter slice/wide moves. (Ideally, fully plan the first three centres during inspection as well, but this is almost impossible because most of the inspection time is already spent on tracing wings.) This avoids OLL parity.

Combining this with redux or Hoya is easy (since those start with doing the first three centres anyway) but using it with Yau is considerably harder because you have to keep track of parity throughout first three centres and the three cross dedges. (And it gets even worse for Yau with half-centres, because you'd need to keep track of parity throughout F2C + 3 cross dedges + four centre bars.)

(Technical detail: actually, you don't need three centres; tracking parity up to solving two adjacent centres also works. The reason is that once you have two adjacent centres solved, as long as you don't "break up" those two centres, the permutation parity of the wings won't change.)

It's basically impossible to avoid PLL parity during inspection because that depends on how edge pairing is done. You can leave pairing the last two edges to the last layer, then determine permutation parity on the fly and do an appropriate alg to finish edge pairing, but this is slower than just doing a PLL parity alg if/when it comes up (unless you have crazy fast ZBLL/1LLL recognition skills).

 

[BECubed]

One of the rules of this thread is to do research, if your method has been proposed before.



Centers last

...

 

[Thom S.]

...

I can't seem to find it currently, but I remember reading on a Speedsolving Wiki Page that Centers Last Methods, such as Cage, avoid Parity, as by solving the Wings and Edges first, the interchangeable Centers are always in an even permutation

 

[xyzzy]

I can't seem to find it currently, but I remember reading on a Speedsolving Wiki Page that Centers Last Methods, such as Cage, avoid Parity, as by solving the Wings and Edges first, the interchangeable Centers are always in an even permutation

This explanation is kind of wrong. The real reason is that neither "OLL parity" nor "PLL parity" are relevant to cage. There's a different type of "parity" where the last few wings to solve (in order to finish the cage) are in a 2-cycle or 4-cycle, but those have shorter and nicer algs (1-12 moves) than OLL parity (14-16 moves).

 

[Hazel]

I know this isn't the right place to post this, but I don't know where else to.
So I'm switching methods again, I was using M-CELL for a while but I dislike blockbuilding. But I don't know what method I want to use, so some suggestions would be nice. I'd like a method that requires little to no blockbuilding (definitely less than Roux or Petrus) but also requires not too many algorithms (like much less than CFOP), and I'd much prefer one that doesn't end with LL like in ZZ, because I'm a bit sick of LL. Any ideas?

 

[sqAree]

I know this isn't the right place to post this, but I don't know where else to.
So I'm switching methods again, I was using M-CELL for a while but I dislike blockbuilding. But I don't know what method I want to use, so some suggestions would be nice. I'd like a method that requires little to no blockbuilding (definitely less than Roux or Petrus) but also requires not too many algorithms (like much less than CFOP), and I'd much prefer one that doesn't end with LL like in ZZ, because I'm a bit sick of LL. Any ideas?

This sounds like SSC would be perfect!

 

[Abram Lookadoo]

Any ideas?

columns first method
human thistlethwaite algorithm
human kociemba algorithm
corners first methods

 

[Hazel]

This sounds like SSC would be perfect!

I looked over Shadowslice's proposal and I like it! I think if I switch to SSC I'll go for the original proposal version rather than the newest one, it makes much more sense to me and requires a bit less thought haha

 

[sqAree]

I looked over Shadowslice's proposal and I like it! I think if I switch to SSC I'll go for the original proposal version rather than the newest one, it makes much more sense to me and requires a bit less thought haha

The problem of the original proposal is that you are supposed to plan eo + 3 edges in inspection (which is a lot already) and later one of those edges gets unsolved anyway. But both variants use a similar number of algs (whereas wv/sls algs are a bit harder than ol5c ones iirc).

 

[Hazel]

After solving a bit with SSC I got bored, so I think I'll probably just go back to CFOP and start doing an event other than 3x3

 

[efattah]

I know this isn't the right place to post this, but I don't know where else to.
So I'm switching methods again, I was using M-CELL for a while but I dislike blockbuilding. But I don't know what method I want to use, so some suggestions would be nice. I'd like a method that requires little to no blockbuilding (definitely less than Roux or Petrus) but also requires not too many algorithms (like much less than CFOP), and I'd much prefer one that doesn't end with LL like in ZZ, because I'm a bit sick of LL. Any ideas?

LMCF beginner's method variant fits the bill. Only 26 algorithms and you can get sub-10 singles. No block-building, no LL.

 

[BECubed]

This explanation is kind of wrong. The real reason is that neither "OLL parity" nor "PLL parity" are relevant to cage. There's a different type of "parity" where the last few wings to solve (in order to finish the cage) are in a 2-cycle or 4-cycle, but those have shorter and nicer algs (1-12 moves) than OLL parity (14-16 moves).

Interesting.

 

[Abram Lookadoo]

new method (B3pL)

1) build a 2x2x3 block
a) build a 1x1x3 block
b) place the 4 edges and 2 centers

2) solve F2l and a U layer pair
a) forge 2 pairs made only of F2l pieces, and one pair of only U layer pieces
b) place the 2 F2l pairs and the 1 unused F2l edge while orienting the U layer pair

3) solve last layer (1-look <= 972 algs (<= 486 not including mirrors))
or
a) solve the last layer corners while maintaining the U layer pair (min 53 algs, max 106 algs)
b) solve the last 3 edges (12 algs)

screamble:U B D U' D' B F' B F' U' L B' D' U F' D' L' F' B D F D' B F' U L
0)x' y
1a) F' D F2
1b) E M' U' r2 - u' R u' R U R u2 y
2a) DU' R U R' D' - F2 U2
2b) F U' L F' L' - L' U2 L y U' L' U2 L

3-1lll) U' L' U2 L2 U2 L2 U' L2 U' L2 U L U'

3a) U R U R' U R U2 R'
3b) U' L' U L' U' L' U' L' U L U L2 U'