Square One

[Scott]

Can anyone here solve Square One blindfolded? I am thinking about trying it.

 

[Kirjava]

No-one has done it yet. I worked out a couple of BLD algs for it though, but the biggest problem is getting past the fact that it changes shape. For that i had to use another notation. (turn x number of pieces) etc.
They're on my website but it's down atm. I'll post them when it's back.

~Thom

 

[grsbmd]

I don't think you'd be able to do it without paper.
You'd just have to calculate the position of each piece after you restore it to a cube.

 

[Kirjava]

Originally posted by grsbmd@Jun 27 2006, 06:45 PM
I don't think you'd be able to do it without paper.
You'd just have to calculate the position of each piece after you restore it to a cube.

no, you wouldn't.

 

[grsbmd]

explain?

 

[pjk]

I can see how it's possible.

 

[grsbmd]

Well, if you didn't calculate it before, wouldn't you have to keep track of it in your head?

 

[GuillaumeMeunier]

Hello,

Stefan Pochmann has solved the Square 1 blindfold !!!This man is crazy !!!
Very nice Stefann !!

Guillaume

 

[Kirjava]

Indeed. Congrats stefan, you crazy man

I assume to did it 'speed-blindfold' style?

Megaminx Blindsolve;
Notation: http://www.snkenjoi.com/archive/cube/minx.jpg
Edge perm (123):
RUR'UR'U'R2U'R'UR'URU2'
Edge orient:
(CA'UC'AR'CA'U2'C'A) B (A'CU2AC'RA'CU'AC') B'
X = (A' C U2' L R U2' A C') and then do (X B X' B')
Corner perm:
(345) L U2 F' D' F U' F' D F U' L'
(234) CU'A'UC'U'AUB'U'BU'B'U2BRUR'UR'U'R2U'R'UR'URU2'
Corner orient:
X = (RD'R'DRD'R'D) do: (X U X U X U2') or: or: (X' U X U')

Orientation: 'leader' stickers

Square-1 Blindsolve;
Corner-Edge Swap:
-2,0/-1,-1/1,-2/-3,-3/-3,-4/3,0/-6,-5/3,-2/-1,0/-6,-4/5,-5
Inverse:
-5,5/6,4/1,0/-3,2/6,5/-3,0/3,4/3,3/-1,2/1,1/2,0

There are my abstract algorithms. For getting past a barrier with the quare one, I wanted to make it that with an algorithm, you read it as "Turn x number of pieces clockwise" instead of "turn x number of degrees". You can see that you'd need a new notation for this. So far I havn't come up with anything decent though. Back to research . I think my biggest problem is that I want to solve it 3x3x3 style.

I'll remember to work on this

~Thom

 

[Stefan]

Originally posted by Kirjava@Jun 29 2006, 10:43 AM
Indeed. Congrats stefan, you crazy man
I assume to did it 'speed-blindfold' style?

Depends on what you mean with that. Turned out to be a nice scramble, I saw how to get to cube shape easily. Then I analyzed what the cycle structure would be after those moves to cube shape. As soon as I'm in cube shape, I only use a single alg repeatedly with setup moves. It's a 2x2 alg: L2 D L2 D' L2 D L2 D' L2. I comment the D turns with "right, left, right, left". I always stay in double square shape except during that algorithm but it's still easy to execute since the DBR corner always gets aligned (to the middle layer).

That alg swaps two corner/edge pairs on U. Then adjust U (for example by 1 to get an edge3cycle or 2 to get a corner3cycle), apply the alg again and unadjust U. With this method (which I've been using for PLL in speedsolving already) you can easily do anything except fix the parity problem. Which fortunately didn't occur because it would've taken me quite some time to fix since I just can't remember an alg for that (I hate this puzzle) and always reinvent a basic fixer every time.

Let me emphasize I am *not* able to blindsolve the Square-1 in the general case. Or to be more precise, I might be able to, but on average it would take a lot longer because the average distance to cube shape is larger and because getting to cube shape in general is still a bit trial-and-error for me even in speedsolving. I'm too lazy to learn the shape graph (I hate this puzzle). And since it would take much longer and I'm lazy, I might be able but not willing to blindsolve Square-1 in the general case. Not yet at least. I have an idea for a better method but I don't know when I'll get back to that. When I saw this thread I was just reminded and got angry that I still can't do it so I just tried it and was a little lucky (not "lucky" in terms of the rules, though). And maybe now we can tackle this more seriously.

 

[Kirjava]

Nice algorithm. I've been messing around with something similar but R2/U instead. Didn't think it'd apply well.

The approach reminds me of your 3x3x3 method but I don't think this way would be competition friendly, like you said.

What I've been looking for is an algorithm that'd work no matter what the shape of the puzzle. My idea is to find something like that they tells you to move x number of pieces/cuts/etc. and use to to swap pieces.

I had no idea currently however else you'd do it.

 

[pjk]

Congrats Pochmann, you're awesome! As least when you BLD a Square-one, you can feel the shape and at least get something from that, unlike on a cube. Good work.

 

[Me]

OH NOES!! :blink: looks like someone beat your record! (grats to Lucas Garron)

 

[CraigBouchard]

I tried this with that scramble you used Stefan, didn't turn out so hot...not sure what happened...maybe if I knew my algs better

Craig