It arrived in the mail today and it's a snazzy puzzle, but I'm stuck with it. I can solve every regular cube from 2x2x2 to 5x5x5, so I'm pretty sure I just need to know the steps and apply knowledge accordingly. I can already complete the two ends of the puzzle, but this leaves the middle, and when I try using standard 2x2x2 sequences to solve that part it messes up the caps, so am I just doing it wrong?

Again, I don't need a really detailed guide, just some basic steps. Help would be appreciated

What I do is match up the extensions with their base first. (what's that word You know making a 7x7x7 a 3x3x3?). After that it can be solved like a normal 2x2x2.

Reducing.

If outside in doesn't work.. try inside out .

PS Are you aware that it shape shifts? (just checking).

Thanks

Apart from the actual pieces, solving inside out and outside in are identical. Both are solving a 2x2x2 and then solving a second 2x2x2 without mixing up the first one. I would "try" solving it that way but I figured out by reducing it to a 2x2x2 it is a lot faster and super simple.

Yes. Would using the edge-pairing algorithms from the 4x4x4 help? I know a really good edge-pairing alg that should work on any size cube (as long as you adjust the deepness of the first and last turns). It swaps two pieces like so (danger: bad ASCII art ahead):
OOOO
XOOO
OOOX
OOOO
On the 4x4x4, it swaps the two Xs. Yeah, I was too lazy to make a proper image

actually, the 2x2x4 is not equivalent to solving a 2x2 then another 2x2, but solving a 2x2 then domino corners, or the top and bottom layer on a 2x2x3.
Yes, you can use the algorithms that you use to flip an edge pair or to swap 2 edges (edge parities) on a 4x4x4 or 5x5x5 cube to flip a corner pair or swap 2 corners on a 2x2x3, 2x2x4, or a domino.

Hope that helps.

Elijah

I managed to solve it using the Japanese directions (possibly the first time Japanese directions have helped?) and I also thank you guys in this thread for helping me.

So it turns out there's an algorithm to swap two of the cap-pieces without shape-shifting the puzzle, so I solved the middle like a 2x2x2 and used that algorithm to finish the caps (after I made one cap blue and the other white, of course.)

I'll make a shoddy picture and write the algorithm here for those people trawling Google and seeking solutions:

Gooday GoombaGeek,

That algo just scrabled my cube.. you might like to check it?
I would use (R2 U R2 U' R2) U' D (R2 U' R2 U R2) D' to achieve that (it swaps the RHS top).

Cheers,
Burgo.

I may have messed it up by simplifying a turn (you were told to turn the top 3 layers, I translated that to "turn the bottom layer the other way", I may have screwed up the direction of that turn. If you kept your instructions, then check the last step: that's the algorithm I used.

I have one too, but not yet scrambled it.
If you start shapeshifting the puzzle and then restore it to a tower (but not how it was before shapeshifting), can you then always still solve it with the 2x2 and domino algorithms, or are some extra algorithms needed?

Scrambling it by shapeshifting is just like scrambling a normal 2x2. You can do all the same moves and everything like normal.

I correct for the odd-parity first by turning U and then I just do a three-cycle of the three broken corners with a [3,1] commutator (needs a setup).

U, [F'2, [[U', L'2, U], R'2, [U', L'2, U], R'2], F'2]

This is [1] [1:[[1:1],1]] for 11 moves total. I haven't proven it but I'm pretty sure this is the shortest possible sequence. If there is no [3,1] commutator sequence that cycles 3 corners in the same face (which I don't think there is) then it is the shortest sequence.

I've never figured out what the square brackets in algorithms refer to

Do I just do all the moves in normal order, or do I have to repeat them in some way?

Anyway, I'm pretty sure that solving by shapeshifting will let you do the center pieces, but the caps need special treatment - either you use an algorithm or keep doing non-shapeshifting turns (R2, move the caps how you want, R2 again).

I wonder which method is better, my reduction way or solving the 2x2x2 then outside. I take about 25 twists then a 2x2x2 solve to solve it with my method. How long does it take using yours? (just solve it like normal and count the rotations)

Yeah it's notation overloading. When I write:
U, [F'2, [[U', L'2, U], R'2, [U', L'2, U], R'2], F'2]

The [] are just grouping and don't actually have any meaning.

When I write [1,1] I mean there is an X part of length 1 and a Y part of length 1 and they are put together like X Y X' Y'. So [1,1] is a 4 move commutator.

[1:1] is a conjugate so it is X Y X'.

Nesting works too:
[[1:1],1] -> [[A:B],C] -> A B A' C A B' A' C'

To me the most important thing about sharing "algorithms" is not the full sequence of moves laid out in order but rather the parts of the sequence and how they are put together to make the whole routine. I'm not a big fan of posting a 11-move sequence without saying how it was made.