Superflip? How about SuperTwist?

[blgentry]

I was playing around and decided to create a condition on the cube where all edges were solved, but corners were only permuted properly, but not oriented. Using 8355 you can "solve" the cube this way, though it takes some effort to make sure all corners are un-oriented.

It's easier if you start with a solved cube though. I do it by doing a corners only OLL on the top to unorient all corners, then fix the edges that got moved with a U perm. Then I turn the cube over and do the same thing to the other side.

So, superflip has a number of algorithms that produce it, including some with great symmetry like: ((M' U)*4 x'z)*3

Do you think SuperTwist has a similar algorithm? How about it's "difficulty"? I know superflip is considered one of the more difficult cases for optimal solvers to solve. I wonder how SuperTwist rates?

Brian.

 

[Ethan Rosen]

R2 B2 D B2 F2 L2 R2 D2 U' F2 R B2 F2 D2 U2 L'

 

[Johannes91]

Do you think SuperTwist has a similar algorithm?

Do you have some specific position in mind? I count 7 unique positions with edges and CP solved and all corners twisted.

Superflip is special because it's so symmetric.

Here's one supertwist alg that's similar to the superflip alg you posted:
([B R' D2 R B', U2] z)4

 

[qqwref]

I converted Ethan's alg to R2 B2 [D S2 M2 U'] D2 B2 [L S2 E2 R'] but I still don't understand how it works. However the bracketed sequences are very similar so maybe that is the key.

 

[blgentry]

Do you have some specific position in mind? There are 6 unique positions with edges and CP solved and all corners twisted.

I actually didn't realize that until I ran the previous poster's alg and saw that it was different than the one that I made "by hand". Then doing rotations between steps I realized there was a "family" of supertwist positions. I didn't know it was 7, but that sounds about right.

Superflip is special because it's so symmetric.

Right; only one of them and it's completely symmetric.

Here's one supertwist alg that's similar to the superflip alg you posted:
([B R' D2 R B', U2] z)4

I'm either reading the alg wrong, or the alg doesn't work. Is the comma significant? I tried the alg 4 times and got a scrambled cube each time. All of the corners are twisted for sure, but they are not permuted properly and the edges are flipped and/or permuted as well.

Brian.

 

[fanwuq]

Here's one supertwist alg that's similar to the superflip alg you posted:
([B R' D2 R B', U2] z)4

I'm either reading the alg wrong, or the alg doesn't work. Is the comma significant? I tried the alg 4 times and got a scrambled cube each time. All of the corners are twisted for sure, but they are not permuted properly and the edges are flipped and/or permuted as well.

Brian.

It's a commutator.

Becomes B R' D2 R B' U2 B R' D2 R B' U2
It twists 2 corners in U.

Try (R U R' U R U' R' U R U2 R' L' U' L U' L' U L U' L' U2 L z2)*2.
Double-sune from left and right.

 

[qqwref]

Here are two I just found by hand:

F B U F' B' L' R' U L R
M2 U M2 U M2 U' D
L' R' U' L R F B U' F' B'
D' U M2 U' M2 U' M2 (31 stm)

R U2 R2 U' R2 U' R2 U2 R U2
F2 M B2 M'
U2 R' U2 R2 U R2 U R2 U2 R'
M B2 M' F2 (28 stm)

 

[Lucas Garron]

I'm either reading the alg wrong, or the alg doesn't work. Is the comma significant?

Well, yes...
http://www.speedsolving.com/wiki/index.php/Commutators_and_Conjugates

([B R' D2 R B', U2] z)4

 

[Swordsman Kirby]

Here are two I just found by hand:

F B U F' B' L' R' U L R
M2 U M2 U M2 U' D
L' R' U' L R F B U' F' B'
D' U M2 U' M2 U' M2 (31 stm)

R U2 R2 U' R2 U' R2 U2 R U2
F2 M B2 M'
U2 R' U2 R2 U R2 U R2 U2 R'
M B2 M' F2 (28 stm)

Both algs are great comms, but I like the second one better because of the OLL alg.

 

[byu]

Do you have some specific position in mind? There are 6 unique positions with edges and CP solved and all corners twisted.

I actually didn't realize that until I ran the previous poster's alg and saw that it was different than the one that I made "by hand". Then doing rotations between steps I realized there was a "family" of supertwist positions. I didn't know it was 7, but that sounds about right.

Superflip is special because it's so symmetric.

Right; only one of them and it's completely symmetric.

Here's one supertwist alg that's similar to the superflip alg you posted:
([B R' D2 R B', U2] z)4

I'm either reading the alg wrong, or the alg doesn't work. Is the comma significant? I tried the alg 4 times and got a scrambled cube each time. All of the corners are twisted for sure, but they are not permuted properly and the edges are flipped and/or permuted as well.

Brian.

[A, B] = A B A' B'

 

[TMOY]

Another one: (L U L' U L U' L' U L U2 L' M2 S2 x2)^2

 

[blgentry]

@fanwuq, Lucas, BYU: Thanks for the notation help. I didn't know that (obviously).

@fanwuq: That double Sune from both sides is exactly the kind of symmetry thing I was hoping for. Very nice!

Brian.

 

[Stefan]

Just 14 moves: U F2 L S2 U2 D2 R2 L' B2 D2 U' S2 R2 L2

Quite a few moves but easy to remember: (RU)35 z2 (RU)35

Four right+left Sunes: (R U R' U R U2 R' L' U' L U' L' U2 L z)4

And very simple: (R'Fz)140

 

[blgentry]

Stephan, I don't understand how you figure out things like this. ...and seriously 140 steps?!? Wow.

Brian.

 

[brunson]

Stephan, I don't understand how you figure out things like this. ...and seriously 140 steps?!? Wow.

Brian.

That'd be 280 QTM. Got time on your hand?

 

[d4m4s74]

how bout (((U R U' R)2 D (R U R' U')2 D)2 Y2)2

 

[Stefan]

Stephan, I don't understand how you figure out things like this.

I'm genius, that's all.

how bout (((U R U' R)2 D (R U R' U')2 D)2 Y2)2

Even if you change the wrong R to R' so it doesn't scramble the cube anymore, this only twists four corners.

Why don't people just check their goddamn algs before they post them? Try these:
http://thearufam.brinkster.net/cube/wrapplet.asp
http://alg.garron.us/

And blgentry, these tools are the real way I figure these things out. I do *not* really repeat random moves 140 times myself. I let the program do it in an instant. Using that, it's mostly trial and error with a little bit of thinking.

 

[cmhardw]

Do you have some specific position in mind? I count 7 unique positions with edges and CP solved and all corners twisted.

Johannes, I count more than 7.

If you ignore symmetric cases then actually I count 86 supertwist cases. I posted a long time ago about combining the superflip and the supertwist into the "super-superflip" (posted on Macky's glossary). There are 86 possible super-superflips where every piece is correctly permuted but none correctly oriented. Since the only difference between these and the supertwist is that the edges are all correctly permuted but flipped, then you should get the same number for supertwist cases ignoring symmetry.

If you do count symmetry I count the following supertwist cases:
1) 1 corner rotated clockwise, 7 rotated counter-clockwise
all these cases are symmetric to eachother (8 cases)

2) 1 corner rotated counter-clockwise, 7 rotated clockwise
all these cases are symmetric to eachother (8 cases)

All the next cases are ones with 4 corners rotated clockwise and 4 counter-clockwise. They are split into the following sub-cases based on symmetry. I list one example of each class.

1) UBR, UBL, UFL, DFR rotated clockwise, and all others rotated counter-clockwise (24 cases)

2) UBR, UBL UFL, DLF rotated clockwise, and all others rotated counter-clockwise (12 cases)

3) UBR, UBL UFL, DBR rotated clockwise, and all others rotated counter-clockwise (12 cases)

4) UBR, UFR, UBL, DBR rotated clockwise, and all others rotated counter-clockwise (8 cases)

5) UBR, UFR, DBL, DFL rotated clockwise, and all others rotated counter-clockwise (6 cases)

6) All U layer corners rotated clockwise and all D layer corners rotated counter-clockwise (6 cases)

7) UBR, UFL, DFR, DBL rotated clockwise, and all others rotated counter-clockwise (2 cases)

Giving 9 unique ways to supertwist corners, with the number of actual cases on the cube marked next to the example for each class.

 

[qqwref]

I also counted 7 cases for 4 edges flipped each way, although I neglected the two cases where only one edge is flipped in one direction and 7 are flipped in the other. Good catch there.

 

[Johannes91]

The 7cw and 7ccw cases are inverses of each other and 2) and 3) are mirrors/inverses. I should've been more careful and explained what I meant by "unique", but the point was just that there's more than one.