[Help Thread] Skewb Discussion

[Ranzha]

Am generating algs that'll solve pure orientation of four/five corners while at least two corners in the U layer are oriented.

@Meep: Is there any way to restrict turns or change notations?

 

[Meep]

Am generating algs that'll solve pure orientation of four/five corners while at least two corners in the U layer are oriented.

@Meep: Is there any way to restrict turns or change notations?

There aren't any nice ways of doing it. To restrict moves you'd have to edit SkewbSearch.m to skip over certain moves. To change notation you might have to edit SkewbMove.m as well as SkewbOutput.m. SkewbMove.m is where the moves are defined; if you use moves that aren't defined in fixed corner notation then you'd have to define your own in there (and modify SkewbSearch.m to loop through those moves). SkewbOutput.m is what outputs an alg to you given an array of numbers that represent moves defined in SkewbMove.m.

However, the solver finds all possible solutions at the optimal depth. You can modify SkewbSearch.m to not stop after finding at least one solution in a given depth. Since it's already finding all possible solutions, the move restricted solutions should still be found, just at later depths; you might be better off writing an alg translator to convert these into another notation.

Also to note, some of the pure corner orientation cases are among the farthest states of the puzzle (depth 11). Here are all the optimal solutions to a pure 4 corner orientation (depth 11), and here are all depth 11 states of the puzzle.

 

[Ranzha]

There aren't any nice ways of doing it. et al.

Challenge considered.

you might be better off writing an alg translator to convert these into another notation.

I've been using this with marvellous though tedious success.

Also to note, some of the pure corner orientation cases are among the farthest states of the puzzle (depth 11). Here are all the optimal solutions to a pure 4 corner orientation (depth 11), and here are all depth 11 states of the puzzle.

 

[Ranzha]

=D

Featuring Nathaniel Knopf on Novation Launchpad.
I wanted to film an a5 with my new camera after getting UWR a12 c:
(7.40), 6.46, 6.15, (5.05), 6.39
Skewb: LanLan with Lubicle Speedy
Method: Mine, with lots of tricks.

 

[TheNextFeliks]

Does anyone know the skewb uwr?

 

[Meep]

Does anyone know the skewb uwr?

They're on the wiki.

 

[Ranzha]

Does anyone know the skewb uwr?

What Meep said. He has a5 UWR, and I have a12 and a100 UWR (each broken last night).

 

[tx789]

felt like posting this

 

[TheNextFeliks]

Gonna get a skewb. How does notation work for skewb?

 

[Meep]

Gonna get a skewb. How does notation work for skewb?

People have different notations, but here's one of the popular ones that's been around for a while.

 

[Ranzha]

Gonna get a skewb. How does notation work for skewb?

Meep's link to FCN is very useful. In addition, there's a fixed centre notation that works like this.

EDIT: Dead acubist links make me have a sad =(

 

[TheNextFeliks]

What are the common skewb methods and pros and cons of each?

 

[Ranzha]

What are the common skewb methods and pros and cons of each?

The four most common skewb methods:

The Kirjava-Meep method (aka "Kirmeep")
Pros: Fully developed; algs are useful for other methods as well. Notation is consistent with WCA notation, except "D" is used in place of "B".
Cons: Some people find FCN algs to be slower for their hands.
Meep has averaged around 6 seconds with this.

Sarah Strong's Method
Pros: Low alg count (2 algs, and they're inverses of each other! Same alg as the 1-alg method). Very easy to finger trick.
Cons: Gripping the puzzle as Sarah does can be difficult to understand for beginners.
Sarah has averaged sub-7 iirc.

Ranzha's Skewb Method
Pros: Primarily based on intuition, many algs are short and join up to create new ones, easy to "hack".
Cons: Layer stays on U, and thus recog can be difficult to get accustomed to. Many feel turning style is awkward.
Ranzha (aka me) averages around 6.

The 1-alg method (popularized by Chris Bird)
Pros: One four-move algorithm that's easy to finger trick.
Cons: Inefficient; as a beginner's method, it does the job.
Mitch Lane has averaged sub-10 with this.

I really don't advise learning the 1-alg method because it's slow, inefficient, and it doesn't take much effort to learn one of the other three methods.

Sarah and I have made video tutorials for beginners as well:
Sarah's: http://www.youtube.com/watch?v=qpykCxeMA6U
Mine: http://bit.ly/ranzhaskewb

There's more that can be said about each of these methods. If you're going for a legit speedsolving method, of these, Sarah's is most widely used. It's important to delve into each of the techniques of each method, for versatility's sake. Also, it's important to explore on your own because, who knows, you might find your own great method!

 

[TheNextFeliks]

Sarah's sounds good.

 

[TheNextFeliks]

I actually learned part of Ranzha's though I want to learn Sarah's.
I have two questions about Sarah's method:
1. What are the two algs?
2. What notation are they?

 

[Ranzha]

I actually learned part of Ranzha's though I want to learn Sarah's.
I have two questions about Sarah's method:
1. What are the two algs?
2. What notation are they?

The first alg is the 4-mover everyone knows. It's the alg from the 1-alg method. It could be called the 'sledgehammer for Skewb'.
The other alg is the inverse.
The images on Sarah's site beautifully outline how the turns are performed. If you perform (F' R F R') as it is with the notation on my site, but with a 1/2 z' rotation beforehand, you'll get Sarah's holding style.

Sarah performs it here at 3:20.

 

[TheNextFeliks]

The first alg is the 4-mover everyone knows. It's the alg from the 1-alg method. It could be called the 'sledgehammer for Skewb'.
The other alg is the inverse.
The images on Sarah's site beautifully outline how the turns are performed. If you perform (F' R F R') as it is with the notation on my site, but with a 1/2 z' rotation beforehand, you'll get Sarah's holding style.

Sarah performs it here at 3:20.

Thanks. I never thought to look if she had a video.

 

[Ranzha]

Bump.

So I have a site up and running here: http://ranzhas-skewb-method.webstarts.com/
The previously posted Webs link will redirect to the new site, so there's no problem with linking people to either URL.

L4C+CLL and Forcing R are under construction mainly because translating algs from <URLD> to <FRLB> is a pain.
I have hax for the Oa/Ob subsets as well as most of the H subset and a bit of the U set, but nothing else has been gen'd yet. I'd consider learning how to force a non-D centre to solve during Welder's. It limits the amount of algs to learn for the last look to just the 29 Oa/Ob/U/S cases, which are easy to learn and recog.

Have fun!

 

[Tim Major]

Bump.

So, 3 main methods, Kirmeep, Sarah's and Ranzha's.

I personally go layer -> R'LRL'/mirror spam until corners solved whilst trying to force at least one centre. Then solve centres with R'LRL' y2 R'LRL' and (R'LRL')*3. I average about 11-12 and my LanLan seems to pop often.

Just bumping this now that there's more reason to practise/make methods.

Also, do you guys have popping problems? Or do I just have a dodgy LanLan. I have a few QJs but they seem really slow compared to LanLan so I prefer LanLan.

 

[Ranzha]

Also, do you guys have popping problems? Or do I just have a dodgy LanLan. I have a few QJs but they seem really slow compared to LanLan so I prefer LanLan.

Popping problems result from loosening of fixed centres, bending of fixed centres at the stalks, or misalignment between the outer compartments of the centre pieces and the inner compartments. I usually salvage the pieces and buy a new skewb.

I'll have ranzha.cubing.net up as soon as I can. For a while, it'll just be skewb stuff since I have a lot documented already, but I need to generate more algs. Can anyone write a skewb algorithm translator from FCN to FCeN?

Very soon I'll make a video about WCA notation for skewb, since the syntax of the regulations may be tough to understand. In the video I'll also cover scrambling conventions such as FCeN (which is what I use), Acubist, Jaap, 2-letter, and 3-letter.

I should practice again.