The New Method / Substep / Concept Idea Thread

[Arc]

Kind of. The alg count shouldn't be too different, based on the estimate in my original post. Well, after I produce the SF alg set we will know for sure!

I think some things that One Piece has over CR is that you don't have to worry about getting the F2L corner slot into the front face (in L3C it is easy to x' if needed) and look ahead into L3C is easier. However, I think case recognition for CR2 is easier than recognising which SF alg to use.

It's not kind of. It literally is Cardan Reduction, but without making the U pair first. It might even technically be fewer algs to solve than CR (because you now have the option of using any two of the four corners) but recognition sounds like a nightmare.

 

[Neuro]

So here's a squan method I saw proposed and have discussed somewhat. Wanted to see if there were any other opinions

1: CSP
2: FB
3: SB-Corner
4: DFDB (L2E algset)
5: TTLL

Steps 1-3 are really efficient and the last 2 steps are pure algs. It's more algs than Lin w/PLL+1 bc of L2E, but I'd argue that TTLL is easier to recognize than PLL+1 and the L2E cases would be really easy to learn. idk if it's better than advanced Lin, but I think it's at least comparable. Might be good for people unsatisfied with PLL+1 recog or who find SB+DB difficult

 

[Sue Doenim]

I'd argue that TTLL is easier to recognize than PLL+1

I would think that PLL+1 (if you're talking about +1 edge, like I think you are) would be more easy to recognise. With TTLL, you can only tell the top layer's CP case from one of the possible AUFs, but with PLL+1, you can tell your CP case instantly, not to mention that it's virtually going to be the exact same as 2 sided PLL recognition.

 

[meskelto]

It's not kind of. It literally is Cardan Reduction, but without making the U pair first. It might even technically be fewer algs to solve than CR (because you now have the option of using any two of the four corners) but recognition sounds like a nightmare.

Sorry, when I was saying "kind of" I was referring to your point about the alg count.

Recognition may not be as bad as it seems. Usually, there will only be three LL corners in LL, meaning only three possible corner pairs to consider (since SF algs are LL algs). Compare this to four single corner cases in Fish and Chips and then it doesn't seem so bad. Definitely something for me to consider though, so thanks for saying.

 

[Neuro]

So here's a squan method I saw proposed and have discussed somewhat. Wanted to see if there were any other opinions

1: CSP
2: FB
3: SB-Corner
4: DFDB (L2E algset)
5: TTLL

Steps 1-3 are really efficient and the last 2 steps are pure algs. It's more algs than Lin w/PLL+1 bc of L2E, but I'd argue that TTLL is easier to recognize than PLL+1 and the L2E cases would be really easy to learn. idk if it's better than advanced Lin, but I think it's at least comparable. Might be good for people unsatisfied with PLL+1 recog or who find SB+DB difficult

After some testing, I found that FB, Sb-C, L2E is significantly less efficient than FB, SB, insert DB in the majority of cases. Since PLL+DF and TTLL have the same amount of algs and have arguably the same quality of recognition, I recommend staying with Lin using PLL+DF (or Vandenbergh)

 

[WoowyBaby]

So this is the forum for experimental methods? I have a method I have been thinking about for a while on the 2x2. Not sure what to name it, maybe Double Block Layer (DBL).
This method is similar to CLL, but better. Here’s what’s better:
- Fewer move count
- Ridiculously simple to predict AUF (hard not to predict it’s so obvious)
- First step requires few moves and has many options each solve
- Many algs can essentially just be CLL

Method Steps:
Make a layer that has two soved blocks in it then a CLL-like alg then AUF.
Easiest way to show the block layer is an example: Scr- U’ F’ R2 U R U2 R F U2 Block Layer- z’ y’ F’ R U2 R
It needs to be oriented, so the face would be opposite colors (red and orange in this case) but the colors in front don’t have to match (in this example it is white-yellow). The blocks should be on right and left.
Making the block layer always requires same or fewer moves than a normal layer. This is unarguably better than making a normal layer.
In fact, you can predict the final move of the solve, AUF, at the start. If the front colors ar me opposite, like in the example, you’ll have no AUF. If same, you’ll have U2. If FR matches left bar color you’ll have U’, and FL matches right bar color you’ll have U.
You might be thinking that you’ll have problems with separation that is the R2 move before AUF, you won’t, because it would always just be part of the alg (btw it would never increase the length of the alg) and as long as you hold the blocks on left and right you wont have a problem.
I will put some algs with pictures on a google doc here:https://docs.google.com/document/d/122Q01aB-0oyghyvRY5zfCWTBOyKIZAwz1w0Kr97-V2c/edit?usp=sharing

This method isn’t complicated. Make two blocks in layer then CLL-like alg. That’s it. Here’s a kind of reverse solve that I just made up Scr: U R’ U R’ U R U2 R U R’
You can see you can make yellow-white block layer R U’ R2 and you can predict your final move/ AUF will be U’ (FR matches left bar) and you’ll have that specific antisune alg and it the preAUF is right so you can just go into the alg:
R U’ R2 // block layer
R U2 R’ U’ R U’ R // CLL-like alg
U’ // AUF
Now that I think, there could also be variations of this method, simplified so you orient and permute in two algs (like LBL) and a more advanced method where the left block is permuted but the right isn’t so it requires tons more algs and it EG-like.

I don’t see downsides of this method, although I might get uninterested like the HD method, but DBL is similar to CLL so I don’t think I will.
If you have any questions just PM me.

What are your thoughts on this?

 

[shadowslice e]

So this is the forum for experimental methods? I have a method I have been thinking about for a while on the 2x2. Not sure what to name it, maybe Double Block Layer (DBL).
This method is similar to CLL, but better. Here’s what’s better:
- Fewer move count
- Ridiculously simple to predict AUF (hard not to predict it’s so obvious)
- First step requires few moves and has many options each solve
- Many algs can essentially just be CLL

Method Steps:
Make a layer that has two soved blocks in it then a CLL-like alg then AUF.
Easiest way to show the block layer is an example: Scr- U’ F’ R2 U R U2 R F U2 Block Layer- z’ y’ F’ R U2 R
It needs to be oriented, so the face would be opposite colors (red and orange in this case) but the colors in front don’t have to match (in this example it is white-yellow). The blocks should be on right and left.
Making the block layer always requires same or fewer moves than a normal layer. This is unarguably better than making a normal layer.
In fact, you can predict the final move of the solve, AUF, at the start. If the front colors ar me opposite, like in the example, you’ll have no AUF. If same, you’ll have U2. If FR matches left bar color you’ll have U’, and FL matches right bar color you’ll have U.
You might be thinking that you’ll have problems with separation that is the R2 move before AUF, you won’t, because it would always just be part of the alg (btw it would never increase the length of the alg) and as long as you hold the blocks on left and right you wont have a problem.
I will make some algs with pictures in a google doc that I will put he link once I have some:

This method isn’t complicated. Make two blocks in layer then CLL-like alg. That’s it. Here’s a kind of reverse solve that I just made up Scr: U R’ U R’ U R U2 R U R’
You can see you can make yellow-white block layer R U’ R2 and you can predict your final move/ AUF will be U’ (FR matches left bar) and you’ll have that specific antisune alg and it the preAUF is right so you can just go into the alg:
R U’ R2 // block layer
R U2 R’ U’ R U’ R // CLL-like alg
U’ // AUF
Now that I think, there could also be variations of this method, simplified so you orient and permute in two algs (like LBL) and a more advanced method where the left block is permuted but the right isn’t so it requires tons more algs and it EG-like.

I don’t see downsides of this method, although I might get uninterested like the HD method, but DBL is similar to CLL so I don’t think I will.
If you have any questions just PM me.

What are your thoughts on this?

Is this not just non-matching blocks?

I use it in solves sometimes and though it's useful if you can one-look it, I really wouldn't consider it a new method

 

[Gappo]

Recently, I found that 2x2 L5C methods are always two looks. There's no 1-look method for L5C? Beacuse it'll have lots of algs?

 

[Gappo]

Could CFOP-CT be a thing? You could do EO after F2L-1. Then TSLE then TTLL. That makes it the same steps as full CFOP though. EO+TSLE+TTLL vs Insert pair + OLL + PLL. How many algs would TSLE be if edges weren't orientated?

There's a similar method called "Ribbon method"

 

[shadowslice e]

Recently, I found that 2x2 L5C methods are always two looks. There's no 1-look method for L5C? Beacuse it'll have lots of algs?

If you had one alg for each case then yes there are loads.

You could, however, use 22 (42- on 2x2) which has a relatively small number of algs though it does have a reasonable number of cases.

 

[Gappo]

If you had one alg for each case then yes there are loads.

You could, however, use 22 (42- on 2x2) which has a relatively small number of algs though it does have a reasonable number of cases.

I thought that full L5C will have 486 algs

 

[Gappo]

(My English is poor,sorry about that
L5C:
Step 1. Build a V
Step 2. Finish rest of it by 1 alg

Step 2-1.Recognize the OLL
Step 2-2.Recognize where the bottom piece goes,it'll be at BR,FR,BL,or FL
Step 2-3.Recognize the CLL (Ignore the bottom piece)

Pros:
Easier 1LLL
Building a V has less move count than building a whole first layer

Cons:
Higher alg count (486 algs)
Harder to recognize

 

[Thom S.]

I thought that full L5C will have 486 algs

It's actually 4860

From what I know there are multiple 2x2ers who are currently in the process of generating and learning L5C cases(I think UnderwaterCuber mentioned it somewhere in this thread some time ago)

Pros:
Easier 1LLL

easier than what

 

[Gappo]

It's actually 4860

From what I know there are multiple 2x2ers who are currently in the process of generating and learning L5C cases(I think UnderwaterCuber mentioned it somewhere in this thread some time ago)



easier than what

I think easier than most of the 2x2 alg sets

 

[Gappo]

It's actually 4860

From what I know there are multiple 2x2ers who are currently in the process of generating and learning L5C cases(I think UnderwaterCuber mentioned it somewhere in this thread some time ago)



easier than what

6(OLL Cases)x6(CLL Cases)x4(bottom piece)x3(orientations of FRD piece)+4(H case)x4x3+6(OLL Finished Cases)=486

 

[Thom S.]

6(OLL Cases)x6(CLL Cases)x4(bottom piece)x3(orientations of FRD piece)+4(H case)x4x3+6(OLL Finished Cases)=486

https://www.speedsolving.com/forum/threads/number-of-permutations-for-l5c.55527/
Or do I think about something different

Without auf-ing it is easy to figure out that it is:

5*4*3 = 5!/2 For permutations because 1st corner can be in any of the 5 spots 2nd in any of the remaining 4 and 3rd in the remaining 3 spots. The order of last to pieces is already known here because all edges are solved. So it becomes 5*4*3*1*1 actually.

3^4 For orientations because 4 out of 5 corners can be in any of the 3 orientations so 3*3*3*3 and once again the last corner must be in the only orientation that makes cube solvable so that's just 3*3*3*3*1.

Put those together and you have (5!/2)*3^4

 

[WoowyBaby]

I thought that full L5C will have 486 algs

Yes, exactly that many.
486 is about as much as ZBLL though, which only a few has fully learned, so learning L5C / 1LLS (one look last slot) probably
has too many algs to be useful.

 

[Gappo]

https://www.speedsolving.com/forum/threads/number-of-permutations-for-l5c.55527/
Or do I think about something different

I thought the number 486 is because I didn’t count TCLL in

 

[Duncan Bannon]

Does anybody know if the L5C algs are going to be good?

 

[Gappo]

Does anybody know if the L5C algs are going to be good?

for example R U R' U2 R' F R F' it's for one of the pi case

it will,just need some optimize