The New Method / Substep / Concept Idea Thread

[Sion]

I've been paying attention to my method for solving, which is Columns first. I've started to realize my solving habits are different than the method I proclaim to be using, which is PCMS. I've been able to write down the steps I use to solve, including steps I want to add to fit my solving style:

1: Solve Corner-edge pairs. When permuting one pair, I want to start working at simultaneously orienting the next. PCMS works at finding good and bad pairs, and reorienting bad pairs after finishing the pair, I work on solving one pair, but set up the next after. This helps keep my movecount lower while getting rid of all the pesky cube rotations in PCMS.

2: permute corners. this shouldnt be too much of a step, but I'm going to learn a cll set soon.

3: orient DB, DR, AND DL edges. they need to have the Color of the D layer center facing correctly, but they don't need to be solved necessarily.

4:L5EO

5: PUDLE (Permute U and D layer edges, using a single alg)

 

[Teoidus]

Not too bad but still not great efficiency wise:

Pairs 18 STM + CLL 9 STM + EO-1 8 STM + L5EO 8 STM + PUDLE 12 STM = ~55 STM

 

[Sion]

Sorry for the double post, but I was playing with my 4x4, which is a puzzle I have yet to learn how to completely solve. However, I have been playing around with some intuitive blockbuilding and I think I created a new method. It's really intuitive for f3l, and the LL is something I need go work on more, but I'm hoping go collab.

1: Solve a 3x3x4 block on the back. To me, this was the longest step, but probably more because I am not too familliar with how Cubies move around it.

2: at front, solve two 1x1x3 pillars. This should leave f3l in a state of the FD edges unsolved and it's center.

3: Solve the remaining centers and FD edges.

4: 3 look last layer
4a: DLL, Dedges last layer. You set up the Last Layer dedges so the last steps are like LL on 3x3.
4b: OLL
4c: PLL
I was able to do everything except solving the dedges, which I doubt there is an algorithm set for, since most method's set the 4x4 into a 3x3 set before finally solving.

 

[shadowslice e]

Sorry for the double post, but I was playing with my 4x4, which is a puzzle I have yet to learn how to completely solve. However, I have been playing around with some intuitive blockbuilding and I think I created a new method. It's really intuitive for f3l, and the LL is something I need go work on more, but I'm hoping go collab.

1: Solve a 3x3x4 block on the back. To me, this was the longest step, but probably more because I am not too familliar with how Cubies move around it.

2: at front, solve two 1x1x3 pillars. This should leave f3l in a state of the FD edges unsolved and it's center.

3: Solve the remaining centers and FD edges.

4: 3 look last layer
4a: DLL, Dedges last layer. You set up the Last Layer dedges so the last steps are like LL on 3x3.
4b: OLL
4c: PLL
I was able to do everything except solving the dedges, which I doubt there is an algorithm set for, since most method's set the 4x4 into a 3x3 set before finally solving.

This is somewhere between K4, OBLBL and Hoya. It's interesting but i think you would have a fair bit of work for the comms used by K4 for LL

 

[Sion]

This is somewhere between K4, OBLBL and Hoya. It's interesting but i think you would have a fair bit of work for the comms used by K4 for LL

Do you think it's efficient at least?

 

[shadowslice e]

Do you think it's efficient at least?

Not more than the methods unmixed.

 

[Sion]

Can someone do some example solves comparing this to other methods? I can't fully do last layer yet.

 

[Shiv3r]

I need some algs for DLL, then yes.
and instead of OLL/PLL, I would suggest doing some EO magic and then doing COLL->EPLL+parity because you skip PLL Parity

 

[Shiv3r]

And Anyone have any thoughts about ZZ-4? after we polish it, I'm considering making a video tutorial...

 

[AlphaSheep]

Can someone do some example solves comparing this to other methods? I can't fully do last layer yet.

The nice thing about example solves is that you can take your time on them and think carefully about each step. As long as you think about factors that could play a role in a speedsolve (for example, if a case would be difficult to see, or whether a sequence of moves needs a lot of regrips or is hard to execute fast), then you can get a reasonable example solve without being that fast at speed solving. It's also OK to look up algorithms for steps you don't know in an example solve.

Algorithms exist for every step in your method. The tricky part is 4a. I'd solve this step with commutators to pair the edges two at a time. Have a look a the algorithms for the K4 method and find some that are suitable. You'd only need a handful of them.

 

[Sion]

The nice thing about example solves is that you can take your time on them and think carefully about each step. As long as you think about factors that could play a role in a speedsolve (for example, if a case would be difficult to see, or whether a sequence of moves needs a lot of regrips or is hard to execute fast), then you can get a reasonable example solve without being that fast at speed solving. It's also OK to look up algorithms for steps you don't know in an example solve.

Algorithms exist for every step in your method. The tricky part is 4a. I'd solve this step with commutators to pair the edges two at a time. Have a look a the algorithms for the K4 method and find some that are suitable. You'd only need a handful of them.

Yeah. I'm probably calling my method BDOP (Blocks, DLL, OLL, PLL.)

DLL is just an alg to move the individual edges to properly pair them up to lower the alg count in OLL t9 only 3x3 cases.

 

[Shiv3r]

A quite possible Idea is to pair them up early, like do all the block except like one edge...
See, All you need for M-slice edge pairing is pretty much the pieces on the U layer...
The algs I would need to do this: A last 2 edges alg that keeps the blocks solved, and maybe a last 4 edges(pairing last 2 edges and flipping either UL/UR edge) case like EOpairing, then after I get those two/three algs I would be good. Maybe even solving edges and orienting them is a good Idea... so we can just dp COLL/EPLL+parity(9 algs total)

 

[Sion]

A quite possible Idea is to pair them up early, like do all the block except like one edge...
See, All you need for M-slice edge pairing is pretty much the pieces on the U layer...
The algs I would need to do this: A last 2 edges alg that keeps the blocks solved, and maybe a last 4 edges(pairing last 2 edges and flipping either UL/UR edge) case like EOpairing, then after I get those two/three algs I would be good. Maybe even solving edges and orienting them is a good Idea... so we can just dp COLL/EPLL+parity(9 algs total)

While this would be more efficient, there wold probably be an extremely high algorithm count, including being (probably) long. I mean, maybe we could try an L5DO to throw the puzzle into a 3X3 state prior, but once again, it would be a bit excessive on the algorithm count.

 

[Shiv3r]

b

While this would be more efficient, there wold probably be an extremely high algorithm count, including being (probably) long. I mean, maybe we could try an L5DO to throw the puzzle into a 3X3 state prior, but once again, it would be a bit excessive on the algorithm count.

but for the second Idea, I just need those two algs. I could do the rest intuitively.
While I do lile this Idea, I am still rooting for ZZ-4,(Im calling it Lewis Method) because that is pretty much Yau/Meyer for ZZ, which really needs a bigcube application.

 

[Sion]

b

but for the second Idea, I just need those two algs. I could do the rest intuitively.
While I do lile this Idea, I am still rooting for ZZ-4,(Im calling it Lewis Method) because that is pretty much Yau/Meyer for ZZ, which really needs a bigcube application.

ZZ-4/Lewis and BDOP/Sion methods are very different.

BDOP is focused on directly solving blocks for the f3l, strait on In the pattern of 3x3 extended to 3x3x4, and two 1x3x3 pillars to filling in the FD edges and last centers. The whole purpose is to avoid reduction pairing by solving blocks instead to directly contribute to finishing instead of reduction to a 3x3 state and then solving.

BDOP doesn't really focus on how to solve the centers, although I say adjacent are better since it gives more wigglw room to solve. So if you were building your major block based off of the Red, white, and blue corner, you would solve those centers first. Then, you would focus on adding a 1x3x3 extension to that, including the fourth center. Then, after solving the 1x1x3 pillars, you would focus on the FD edges and solving the last two centers, which in my experience is insanely easy and intuitive because they are adjacent.

Then DLL, OLL, PLL, which is the only reduction part of the method.

 

[Shiv3r]

ZZ-4/Lewis and BDOP/Sion methods are very different.

BDOP is focused on directly solving blocks for the f3l, strait on In the pattern of 3x3 extended to 3x3x4, and two 1x3x3 pillars to filling in the FD edges and last centers. The whole purpose is to avoid reduction pairing by solving blocks instead to directly contribute to finishing instead of reduction to a 3x3 state and then solving.

BDOP doesn't really focus on how to solve the centers, although I say adjacent are better since it gives more wigglw room to solve. So if you were building your major block based off of the Red, white, and blue corner, you would solve those centers first. Then, you would focus on adding a 1x3x3 extension to that, including the fourth center. Then, after solving the 1x1x3 pillars, you would focus on the FD edges and solving the last two centers, which in my experience is insanely easy and intuitive because they are adjacent.

Then DLL, OLL, PLL, which is the only reduction part of the method.

I know, but I have gotten Lewis mostly polished, and I'm advocating for it. Not that I'm against this method, however, I am helping to develop it because I like developing methods.

 

[Teoidus]

what are average movecounts for ZZ-4?

 

[Sion]

I know, but I have gotten Lewis mostly polished, and I'm advocating for it. Not that I'm against this method, however, I am helping to develop it because I like developing methods.

I got it!

DOLL (Dedge orient last layer, so just simply flipping the edge instead of moving them around) COLL, EEPLL (Extended Edge Permutation last layer)

So BDCE

 

[Shiv3r]

I got it!

DOLL (Dedge orient last layer, so just simply flipping the edge instead of moving them around) COLL, EEPLL (Extended Edge Permutation last layer)

So BDCE

I think that Dedge pairing is simpler, and then EO+ COLL-EPLL+parity is better, mostly because EPLL+parity is so amazing.

what are average movecounts for ZZ-4?

except for the 3x3 stage, slightly more than Yau/Meyer because of the EOpairing. The 3x3 stage, however, is fairly low movecount since it is ZZ and not CFOP.

 

[Sion]

I think that Dedge pairing is simpler, and then EO+ COLL-EPLL+parity is better, mostly because EPLL+parity is so amazing.

except for the 3x3 stage, slightly more than Yau/Meyer because of the EOpairing. The 3x3 stage, however, is fairly low movecount since it is ZZ and not CFOP.

Actually, it's harder, since they would need to both flip and commute, and the algs would be far more confusing.

Edit: not to mention the EPLLs are more roux-like (most likely) and MU algorithms are much more easy to remember and are also quick to execute.