Siamese cube methods

Kenneth

Not Alot
Joined
Aug 10, 2007
Messages
1,693
Location
Kauparve, Hejde, Gotland, Sweden
WCA ID
2005GUST01
YouTube
Visit Channel
First, is it [wiki]siamese cube[/wiki] or siamese cubes?

I use singular...

Talking about the real thing = siamse 3. (http://www.twistypuzzles.com/cgi-bin/pdb-search.cgi click that link and use 'siamese' for keyword and you will see images of diffrent types)

What is your method?

I use about Roux. I made myself a new siamese the other day that have better cubes than my old one and had a non lucky 53 seconds as best so far, the average is around a minute.

Anyone who uses Roux will easily solve a simaese but you may use Fridrich or even Petrus. I will later edit this post with examples of how to solve the problem cases for some methods (the cases where you cannot use the normal algs).

But first, any method will have the same start, solve the R/L centres and LF + LB edges, the first block in the Roux method. You may do this later but it will only complicate things, better to have it done:

First block:

Pair up L edges and L centre on either F/B or best, the R side and then place the whole triple using E (or u). If you pair in F or B you place the pices in the middle layer. Often you will have to orient the block before you place it and then you have to move it to the R side using E and do R2. This is why it is better to pair in R, then you put the pieces in the S slice, orient first edge in RD, second in RU, put the centre using E and then R or R' + E2 to place. But far from all cases are easy to do in R, so you have to use all variations to save moves.

FB example 1: r2 U r' R2 U2 u2 r2 R2 u2 R r' u' r' u U' r' U2 u2 r u R2 r2 u R2 u (Using MEGA-SCRAMBLER)

Make the block in B:

Orient one edge: Rw'
Move it to the side: U
Insert centre: M
Move the other edge to position: E'
Block up: Rw
Put in position: u'

Example 2: U' r R U R2 U r2 R U' u2 r' R2 u R u R r' u2 R U u' R' r2 U2 R'

Block up in B again: R' M2 ... orient and place block : u R2 u2

Easy =)

Example 3: U r' U R2 r' U' R2 r U2 R u2 R u2 U R2 U' u2 r U' R2 U' r2 R' U2 u

Make block in R: U Rw' R' U M' E R u2

Use your inspection time to find the moves for the first block in both halfs, start from the hardest and leave the one that is easy to remember for the seconf half.

Second block:

Here I will put examples of how to solve any piece into the block if only one is missing. but also a few basic block building moves.

Missing RD: Put the edge in the M-slice and move it to D using M or M', then do R2 U/U' M/M'/M2 U/U' R2.

RD in position but flipped: R2 U M2 U2 M U R2

RF: R U M/M'/M2 U' R' and the same style for RB (works the same as the RD algs, just a diffrent R move at first)

Corners: the usual 2-gen F2L moves.

CMLL:

Is the same as 2-gen CO, nothing new...

L6E:

Same, same... but some algs that uses face moves (F2 usally) may not work.

Fridrich example:

This is continued from the first example for the left side block (same scramble and same soution for the block).

Cross: U' R U' Rw U Rw U2 Rw'

RB slot: U' R U' R' U Rw' U' R U M' .. the solution to the tricky insert.
RF slot: U' R U2 R' U2 R U' R'

OLL

Corners: R U2 R' U' R U' (R2) U2 R U R' U R .. Anti-Anti but the second from the backside to save one move (two R' makes one R2).
Edges : R U R' U' M' U R U' Rw' ... 3-cycle ELL

U-PLL: M2 U' M U2 M' U' M2

AUF : U'

62 moves in total. The problems are pairing up some pairs, inserting a wrongly oriented pair, long corner orientations (but that goes for any method) and the fact that most OLL does not work so you have to learn sune-OLL or use 2-look as I did here.

Petrus:

Start from the first block and add M-centres and the BD edge (almost a full 2x2x3), you may do a 2x2x2 first but it is easier if you do the first block and add the M-slice parts after that.

You leave the FD edge at fist because you need a keyhole for the EO.

EO :

Move three oriented edges to the R-side by doing U R U R... If there are no oriented edges you use the one in the FU position, do M' to orient and then you can store it using U R, then move the next edge to orient to UB and do M and you have the slice back and two edges oriented.

When you got all three R-side edges oriented you use MU moves to orient the rest and then also place the FD edge.

Now you are ready for 2-gen!, compleate the R-side using RU and then go for [wiki]2GLL[/wiki] :P

If you do EO like that you will solve the siamese easier than you do using Fridrich, or you will at least use lesser moves on average.

Example:

Using the same scramble and start as the first block example.

M-slice BD : R U' M'

EO:

R-side edges : R U2 R
U-layer edges : M' U' M U2 M' U' M

Compleate 2x2x3 : U M' U2 M

Now the rest is 2-gen:

F2L : R U2 R' U R U R' ... R' U R U' R' U2 R U' R' U R
LL : U R U2 R' U' R U' R' (Antisune) ... U' R U' R U R U R U' R' U' R2 (U-PLL, Use MU to save moves)

61 moves (STM), 1 better than Fridrich but a shorter U-PLL is possible. Petrus is usally a bit better because the problems with Fridrich is the fact that the edges are having the wrong orientation for pairing up and also placing the pairs in 50% of the cases, doing the EO before that helps. Not to mention the possibility to force a EP skip making it a 1LLL.
 
Last edited:
OCLL+EO in one step?
7*21 = 147?

just <r,R,U>

But far too many algs for me for siamese cubing :p

Nope, it is the same number as OLL, corners are always permuted so you will have 8 cases of EO for each CO, that are 8 including solved 8*8=64 minus some duplicates makes 57.

You can use Sune-OLL for this...

Another advanced method is to use EO and then ZBLL. ZB is 8 CO times 12 EP wich makes 96 cases but again there are a few duplicates (85 cases, see [wiki]2GLL[/wiki]).

If you solve F2B and L6EO + place centres and FD/BD edges you will have 1-look LL using this. (remember it is not a waste learning all that just for the siamese, you can use the same algs when you get the cases speedsolving the 3x3x3, and it is fast RUM moves)

EDIT : EG mabye? :P

pi : R U' Rw' U2 R' U Rw U2 R U' R'
corners oriented: R U' Rw' U R' U2 R U' Rw U R'

I just found it is possible, there are 8 EG1 cases but no EG2 (because two FL corners are fixed in position). Don't know if there is any good use for it but it is fun =)

Editmore: hmm, mabye shorter algs on average than the EG0/CLL/CO, I will test for optimal algs I think :)

Na, longer...
 
Last edited:
I use Fridrich, but have to use M slices and turns such as r, when solving the entire F2L. Then for LL, I use two gen algorithms.
 
Back
Top