The New Method / Substep / Concept Idea Thread

[shadowslice e]

Stadler I think.

It's stadler for the first bit then a K5ish (K4 on 5x5x5) LL although it does redux rather than direct solving (which is likely more efficient)

 

[crafto22]

Weird (maybe useful) Sandwich Variant

I just came up with this randomly. lol

Step 1: Cross
Desc: Same as the first step of CFOP
Moves: 6
Exec: 1-2 second

Step 2: Top layer + CO + EO
Desc: Place top colour pieces in their layer, then execute 2 algorithms (OLL followed by a special PLL that corrects parities if necessary) followed by another algorithm that solves bottom layer corner orientation and E-slice edge orientation. This is all in 2-looks as PLL won't affect the bottom layers, so a transition from PLL to CO + EO is instant
Moves: Maybe 40?
Exec: 7-9 seconds

Step 3: Permute corners + E-slice
Desc: Algorithmically permute the corners whilst placing the E-slice edges
Moves: 9-12
Exec: 2-4 seconds

Total moves: approx. 58 STM
Execution time: 10-17 seconds

Pros:

- Algorithmic approach to solving is faster than using intuition
- Algorithms for all steps are pretty great (I've checked)
- Decent movecout for an algorithm-heavy method
- Easy concepts like OLL, PLL or cross make this method easy to learn
- Execellent lookahead as one could learn how algs affect the cube

Cons:

- High-ish movecount compared to other method such as Roux or Petrus
- Intuition is occasionnaly beneficial

Okay, TELL ME WHAT YOU THINK

 

[2180161]

I just came up with this randomly. lol

Step 1: Cross
Desc: Same as the first step of CFOP
Moves: 6
Exec: 1-2 second

Step 2: Top layer + CO + EO
Desc: Place top colour pieces in their layer, then execute 2 algorithms (OLL followed by a special PLL that corrects parities if necessary) followed by another algorithm that solves bottom layer corner orientation and E-slice edge orientation. This is all in 2-looks as PLL won't affect the bottom layers, so a transition from PLL to CO + EO is instant
Moves: Maybe 40?
Exec: 7-9 seconds

Step 3: Permute corners + E-slice
Desc: Algorithmically permute the corners whilst placing the E-slice edges
Moves: 9-12
Exec: 2-4 seconds

Total moves: approx. 58 STM
Execution time: 10-17 seconds

Pros:

- Algorithmic approach to solving is faster than using intuition
- Algorithms for all steps are pretty great (I've checked)
- Decent movecout for an algorithm-heavy method
- Easy concepts like OLL, PLL or cross make this method easy to learn
- Execellent lookahead as one could learn how algs affect the cube

Cons:

- High-ish movecount compared to other method such as Roux or Petrus
- Intuition is occasionnaly beneficial

Okay, TELL ME WHAT YOU THINK

I don't understand. Is the cross on bottom?
EDIT: Never mind I understand. This is bad.

 

[shadowslice e]

I just came up with this randomly. lol

Step 1: Cross
Desc: Same as the first step of CFOP
Moves: 6
Exec: 1-2 second

Step 2: Top layer + CO + EO
Desc: Place top colour pieces in their layer, then execute 2 algorithms (OLL followed by a special PLL that corrects parities if necessary) followed by another algorithm that solves bottom layer corner orientation and E-slice edge orientation. This is all in 2-looks as PLL won't affect the bottom layers, so a transition from PLL to CO + EO is instant
Moves: Maybe 40?
Exec: 7-9 seconds

Step 3: Permute corners + E-slice
Desc: Algorithmically permute the corners whilst placing the E-slice edges
Moves: 9-12
Exec: 2-4 seconds

Total moves: approx. 58 STM
Execution time: 10-17 seconds

Pros:

- Algorithmic approach to solving is faster than using intuition
- Algorithms for all steps are pretty great (I've checked)
- Decent movecout for an algorithm-heavy method
- Easy concepts like OLL, PLL or cross make this method easy to learn
- Execellent lookahead as one could learn how algs affect the cube

Cons:

- High-ish movecount compared to other method such as Roux or Petrus
- Intuition is occasionnaly beneficial

Okay, TELL ME WHAT YOU THINK

This is essentially the Zagorec method. A lesser known but pretty fun method. It's actually not very efficient at around 70 moves.

Still a pretty fun method overall tho.

 

[crafto22]

This is essentially the Zagorec method. A lesser known but pretty fun method. It's actually not very efficient at around 70 moves.

Still a pretty fun method overall tho.

Um... No, just no. This is barely close. Zagorec has like a billion steps. This is done in 4 looks, which is WAY better. Zagorec is more of an orient-first method. This isn't really similar at all. Also, while Zagorec is super inefficient, I got a 62 move solution a couple of hours ago with this.

 

[crafto22]

I don't understand. Is the cross on bottom?
EDIT: Never mind I understand. This is bad.

Why so? I don't feel like you understand.

 

[penguinz7]

This has probably been done before.


1. EO (not eoline, just eo)
2. Columns
1. Move F2L edges off of bottom layer. If you start F2L with edges down there, you can end up with some terrible cases.. This should only take a few moves.
2. <RUL> F2L
3. COLL
3. L8E
There are several ways of doing this, inserting DL and DR and then doing L6E seems to me to be one of the better ways unless there's a really easy case.
You could also insert DL and DR right before columns, as doing them before L6E tends to be pretty awkward. That's probably better..
Pros: It's really fun!!
Cons: A lot.

 

[shadowslice e]

This has probably been done before.


1. EO (not eoline, just eo)
2. Columns
1. Move F2L edges off of bottom layer. If you start F2L with edges down there, you can end up with some terrible cases.. This should only take a few moves.
2. <RUL> F2L
3. COLL
3. L8E
There are several ways of doing this, inserting DL and DR and then doing L6E seems to me to be one of the better ways unless there's a really easy case.
You could also insert DL and DR right before columns, as doing them before L6E tends to be pretty awkward. That's probably better..
Pros: It's really fun!!
Cons: A lot.

This is sort of like the ZZroux that often pops up in this thread combined with some aspects of SSC. I think perhaps maybe an FB then SB approach may be better but there are still a lot of issues with that that cause ZZ and Roux are good for different reasons (Free blockbuilding vs restricted moveset)

 

[penguinz7]

I think perhaps maybe an FB then SB approach may be better but there are still a lot of issues with that that cause ZZ and Roux are good for different reasons (Free blockbuilding vs restricted moveset)

I really don't care about how this could be better, it was purely for fun.

 

[4Chan]

I talked with some people about this idea, and I've decided that I give up, but I'm posting it for posterity.

It's another silly ZZ variant that forces an LL skip everytime, but with better recognition than ZBLL, and has a sufficiently low alg count that makes it humanly possible.

You make an EOLine, then blocks, then with 3rd slot you make your edge/corner pair and then orient all the corners. (just 75 cases)
Now that EVERYTHING is oriented, this causes an absurdly massive reduction in cases. I estimate it to be around 500-800 as the raw number without any tricks.

Here's a crappy example solve with a crappy orientation alg because I don't feel like generating a better one:

Scramble: F R U2 L2 D' L U B2 R' L2 F U2 F' L2 B L2 U2 F L2 F'

EOLine: y x D2 L’ D R U R’ U F’ x’ y’

First Block: R’ U2 L2

Second Block: R2 U R’ U R’ U’ R2

Left Pair + Corner Orientation: U2 L’ U L + F2 L’ F2 L F2 U’ L D’ L U2 L U’ L D
(SAME RECOG AS WINTER VARIATION, 26 OF THE CASES ARE ACTUALLY WINTER VARIATION CASES)

EVERYTHING ELSE(LS + LL): U2 y’ x R2 U’ L’ U R2 U’ L U
(LOOKS LIKE PLL)


EDIT: If I were 16 again, I would totally use this as my main method just to be unique.

 

[shadowslice e]

I talked with some people about this idea, and I've decided that I give up, but I'm posting it for posterity.

It's another silly ZZ variant that forces an LL skip everytime, but with better recognition than ZBLL, and has a sufficiently low alg count that makes it humanly possible.

You make an EOLine, then blocks, then with 3rd slot you make your edge/corner pair and then orient all the corners. (just 75 cases)
Now that EVERYTHING is oriented, this causes an absurdly massive reduction in cases. I estimate it to be around 500-800 as the raw number without any tricks.

Here's a crappy example solve with a crappy orientation alg because I don't feel like generating a better one:

Scramble: F R U2 L2 D' L U B2 R' L2 F U2 F' L2 B L2 U2 F L2 F'

EOLine: y x D2 L’ D R U R’ U F’ x’ y’

First Block: R’ U2 L2

Second Block: R2 U R’ U R’ U’ R2

Left Pair + Corner Orientation: U2 L’ U L + F2 L’ F2 L F2 U’ L D’ L U2 L U’ L D
(SAME RECOG AS WINTER VARIATION, 26 OF THE CASES ARE ACTUALLY WINTER VARIATION CASES)

EVERYTHING ELSE(LS + LL): U2 y’ x R2 U’ L’ U R2 U’ L U
(LOOKS LIKE PLL)


EDIT: If I were 16 again, I would totally use this as my main method just to be unique.

Look promising. What are the algs like for the last bit? I'm guess a lot of them are pretty long...

 

[4Chan]

Last part's easy. Tons of conjugated PLLs. Anything with a block on the right side and a corner swap is just a conjugated ZBLL, so a sizable chunk are already known. Length isn't an issue really.

 

[shadowslice e]

Last part's easy. Tons of conjugated PLLs. Anything with a block on the right side and a corner swap is just a conjugated ZBLL, so a sizable chunk are already known. Length isn't an issue really.

ok then, so what's the movecount like?

And you may have a bit of work convincing people to actually use the method. That is without doing it yourself. I've git two that I know of trying any of mine and those have low alg counts.

But I think thus definately has potential. I may try to refine it as some point any have a look at some of the algs.

 

[4Chan]

I don't actually want to convince anyone to use it.

And the move count would be around 40.

 

[shadowslice e]

I don't actually want to convince anyone to use it.

And the move count would be around 40.

More power to you then I guess. I think it may be worth learning this as a supplement to OLS perhaps just on the off chance it comes up (this can be cycled to fairly easily).

 

[crafto22]

This might work for something:

EDIT: I have modified the movecounts. I thought stuff was more efficient than it actually is.

EOLine - 7 moves
Left Block - 12 moves
Corner in DBR - 2 moves
WVCP: 12 moves
Modified L6E - 12 moves

Total: 45 moves

Still really good movecount

 

[shadowslice e]

This might work for something:

EOLine - 7 moves
Left Block - 7 moves
Corner in DBR - 2 moves
WVCP: 10 moves
Modified L6E - 12 moves

Total: 38 moves

DAT MOVECOUNT DOE!

Left block will definately take more than 7 moves. It will be more like at least 10 if not more; roux FB is only ~8 fingertrickably and that's with completely free block building (X2 Y), unrestricted moves set and very good lookahead. Speedsolving it will be definitely be longer.

WVCP recog isn't very nice either tbh (I think you also forgot to factor in the pair) and I can't see L6E being very nice and I don't know where you got the movecount statistics for it either cause Roux is about 9 (discounting Eo) and that's with a nice postioning for the edges.

 

[supercavitation]

This might work for something:

EOLine - 7 moves
Left Block - 7 moves
Corner in DBR - 2 moves
WVCP: 10 moves
Modified L6E - 12 moves

Total: 38 moves

DAT MOVECOUNT DOE!

Where did you get your numbers for WVCP?

 

[shadowslice e]

Where did you get your numbers for WVCP?

https://www.speedsolving.com/forum/showthread.php?55916-Algorithms-for-WVCP&highlight=WVCP

I would guess from here but having had a quick glance it seems to be >10

 

[supercavitation]

https://www.speedsolving.com/forum/showthread.php?55916-Algorithms-for-WVCP&highlight=WVCP

I would guess from here but having had a quick glance it seems to be >10

Definitely. I'll do the math at some point, but that's >10 even without the 4 moves for making the pair.