From cube-lovers-errors@mc.lcs.mit.edu Thu Nov 11 18:02:33 1999 Return-Path: Received: from sun28.aic.nrl.navy.mil (sun28.aic.nrl.navy.mil [132.250.84.38]) by mc.lcs.mit.edu (8.9.1a/8.9.1-mod) with SMTP id SAA06302 for ; Thu, 11 Nov 1999 18:02:32 -0500 (EST) Precedence: bulk Errors-To: cube-lovers-errors@mc.lcs.mit.edu Date: Thu, 28 Oct 1999 15:38:18 -0400 (EDT) From: Daniel B Knights To: Norman Richards Cc: Cube-Lovers@ai.mit.edu Subject: 3-Cube in 1 Look Message-Id: >> There was discussion of blindfolded solving during March 98 on this >> list. At that time no one was able to do it. You may be the first. > I'd love some more elaboration of the specific sequences the >original poster used. (or any suggestions others have to make) I >think this would be quite a trick. I'd be happy just being able to do >a 2x2x2 cube like this or maybe one face of a 3x3x3... Solving one face of a 3x3x3 in one look does not require any special system for planning. If it can be done in 20-25 moves, then I've found it is feasible (although cumbersome) to simply plan out all 20 moves while keeping track of each move's effect on the relevant pieces. If you are interested in solving the whole cube in 1-look, (or a 2x2x2 cube) I suggest you try doing 2 looks first: 1 to solve corners (without changing edges), 1 to solve edges. Once you can do this, it should be clear how to do it in one look. If 2-looks for the whole cube is still too difficult at first, then try solving just the corners with 2-looks, 1 to position them and 1 to orient them correctly. (But don't mess up the edges while you do it!) You don't really have to "update" your memory as you go, because you can basically plan out the entire solution before you start, as follows: 1. it's only 3-5 corner permutations to get all corners in the correct location, and you just have to keep track of each corner's orientation (i.e., needs to be rotated clockwise or anti-clockwise in place). For these permutation moves, I most often use a single-layer 3-corner interchange that preserves the corner orientation relative to that layer. The move [Ri F Ri B2 R Fi Ri B2 R2] accomplishes this in the top layer. 2. 1-3 more sequences to orient all the corners. I use simple moves like [(R U Ri Ui)^2 D (U R Ui Ri)^2 Di] to re-orient two corners in the bottom layer. (**Now you've SOLVED a 2x2x2 Cube!**) 3. Then usually 5-7 sequences to put all edges in place, but again keeping track of which ones are flipped. For these permutations, I will again often use a 3-edge swap like the following: [R2 Ui Fs R2 Bs Ui R2] 4. then a few more edge-flip sequences. I use the (very) inefficient two-edge flip maneuver: [Ls Fi Ls Di Ls B2 Rs Di Rs Fi Rs U2] Then you're done! (about 200 moves later.) This is not an easy "trick" - I still find it quite challenging to correctly plan out the entire corners solution and the entire edges solution, and to then implement them correctly from memory. The real "trick" for me is that I don't memorize the locations of the pieces, just the sequence of permutations that I planned out to solve them. This way, if you can plan out the 5 corner permutations in advance, then you only need to remember those 5 items to solve the corners, which don't change throughout the solution. If you instead memorize the locations of the pieces, you have to keep memorizing new locations throughout the solution. (which is impossible for me.) I first plan out the entire solution with my eyes open, and memorize it. (planning it out correctly may be the most difficult part.) Then you only have 15 permutations to remember and execute correctly, without any new memorization after you close your eyes. Well, that's enough about that. I just want to make this approach clear because I think blindfolded cubing is well within the bounds of "normal human memory capabilities." Good Luck!