Old Pochmann or Classic Pochmann is a blindfold cubing method for 3x3x3 created by Stefan Pochmann [WCA], [web] some years ago. At first it was namned only Pochmann but later he also created M2 and R2 that are more advanced and this later name came in use. After some discussion, to not conflict with 3OP (or only OP) that is an other BLD method the last name 'classic' came up and is the one that is mostly used the later times.
The idea is to use 2-cycles to swap two pieces, one sitting in the buffer and one in the shooting position, the piece in the buffer is the one that is solved and the one in the shooting position goes to the buffer and that's it, you solve piece by piece in a chain. Poblem is that it is not possible to swap only two pieces on a puzzle that has not got 'real paritys', like the 3x3 or the Megaminx. On the 3x3 it is possible to get 'pseudo paritys' that swaps two corners and two edges (like T-PLL), that because all quarter moves are double 4-cycles and such may be transformed to 2 corners and 2 edges using a 3+3 cycle. But on the Mega all moves are double 5-cycles so that is not possible, we have to stick to algs that swaps 2+2 edges or 2+2 corners and that complicates things a little, but not so much...
Good thing is; there are no paritys! :p
Using a edge/edge 2+2 (or corner), we only need one of the swaps, the other must be a slave swap that alters the pieces in the slave buffers in even/odd numbers of swaps.
Setup moves, moving goal positions to the shooting position, that changes positions for the slave buffers... the slaves must always be restored before the shooting, this makes things a bit harder than it would be if we had a corner/edge 2+2 cycle, aspecially for cornes that may be moved using two diffrent U-adjacent sides (R, F, L..), edges are only affected by the side it is on so there the problem is smaller.
This is a block swap alg (see Intuitive PLL), the second part restores the corners that was swapped at first but because it is the mirror swap from the first the blocks will have diffrent edges, this makes 2+2 edge swaps. The corner alg works the same but there the corners are diffrent form the first swap, not the edges.
This is the 6 move 3x3 OLL and it's mirror inverse in combiantion.
This is the same alg as for edges but with an extra U move in the middle and one in the end.
This is the usual 9-move 3x3 A-PLL, B at the R-side in the CCW alg.
In the images green arrows are the shootings, the buffers are at UF/UFR and the goals at UB and ULB (B at the L-side). Red arrows are the slave buffer swaps at UR<->UL for edges and URB<->ULF for corners. If you already know an alg for any of these cases that does the same but with a little diffrent turns, then use that one instead!
Solving pieces to the slave buffers:
You can solve almost all pieces using only the X algs. The only exeption is when the piece to solve belongs to one of the slaves. But it is possible to solve these and also simulate we swapped the slaves at the same time using a 3-cycle PLL. From start keep count of the number of swaps you did, or rather keep track of even/uneven number of swaps (swaps MOD 2). This is important because we need to do it diffrently depending on that.
If you have done 0, 2, 4, 6.. (even) swaps, then look inte first column and if you have swapped 1, 3, 5, 7.. times (odd), then look in the second column. Colours for arrows are like in the X-swaps, green are the buffer<->goal shootings and reds are the slave swaps.
To memo the Megaminx I think it is a good idea to divide it into smaller blocks. If you pick 4 corners and 6 edges for each block you will have 5 of them and if you choose the blocks as in this picture:
The blocks will fit into each other and make one for each side of the last layer. You can then think of the five blocks as diffrent rooms in a house or something like that. Then let the pieces in the blocks represent diffrent thing within the room or anything. This is just a suggestion to ease up things a bit, I will not work on a full method for memo because most blindfold cubers are having their variation of a variation of the variation...
- Is this very hard?
- No, it is not, the number of pieces are exactly the same as if you do 2.5 cubes multiple blindfold, 20 corners, 30 edges and no centres, it should be a bit easier than the 4x4x4 to memo.
To make the setup turns a bit easier it is possible to use varations of the X-PLL's that orients the pieces in the buffer and the goal while swapping them:
Edges with flips:
Bah! but it was the shortest that was found using the Lars V solver [link]. As a ELL on 3x3 this is one of the harder cases (H with opposite flips). You can also use the normal X-PLL alg and setup using Bl' Br' (y), it also makes 15 turns if you continue the Br' to merge with the first R' (so for real you will do something like (y) Br' R2'..).
For this case the inverse is the same as the mirror so it is only one cycle. See also 3ELL for more edge cycles.
All possible last layer corner 3-cycles are at the 3CLL page