Message #1939

From: Andrey <andreyastrelin@yahoo.com>
Subject: Re: 24cell FT
Date: Sun, 27 Nov 2011 20:19:49 -0000

No, I’ve used nothing special. My technique on all levels is the same: assume that I’ve solved some types of pieces and now work on the next type. And that the puzzle is large enough and has a couple of cells twists of which don’t intersect.
Select some cell F and some piece f of the desired type. Find a sequence of twists R that moves f on another cell G and there are no other pieces originally from layer F that are in layer G now. If you are successful, build macro (R G R’ F R G’ R’ F’) where F and G are some twists of the corresponding cells - it will be 3-loop of your type of pieces, that moves two pieces from F and one piece from somewhere else. Usually I add some setup sequence to this macro so that 3rd piece appears in some convenient position.
Then I solve puzzle face-by-face: move all pieces of the current type that have color F to face F (using 3-cycles), and then sort them inside F using half-macros (R G R’) and (R G’ R’) with twists of F between.
If you can’t find macro R in the first place, try 3-cycle from the previous stage. It moves only one piece of previous type and some pieces of the current type outside of F, so it will be more easy to find/make suitable cell G for the R macro. And if your 3-loop doesn’t move any pieces of the current type, then you’ve selected wrong order of solving: switch off pieces of previous type, find R for new type, solve it, and then solve previous type by using your macro - it will not move pieces of new type.
Something like that. It should work for all shallow-cut puzzles other than simplex, {3,3}-duoprism and some other small puzzles.

Andrey