Hi Kevin,

Yes

There are two cases,
1) the corner belong to an edge, or to more edges
2) the corner belongs to no edge, all adjacent colors are wrong

solution of case 2 = transform it to case 1)
case 2)
Bring the corner with a 90 degree turn away. Exchange the 2x2x2 blocks in the way that edge has same color as the 2x2x2 block. This causes minimal destruction of builded blocks.
Then turn the 2x2x2 to left or right. After that you can turn it back and the corner belongs to one edge.
Now you can repair the other blocks.

case 1)

The finger (2x1x1 block) must lie horizontal in the upper layer. Look to the sticker with different color adjacent to the finger. Permute the other 2x2x2 cubes in that way that the color of destination position matches. You can exchange 2x2x1 inner blocks and 2x2x1 outer horizontal blocks and turn this blocks to make a match with the moving "fingers".
If two "fingers" have the same colors on the heads then I name it "candidate". You can move "candidates" to its color and turn this position. In this way you transform your "fingers" to 2x2x1 blocks.

To solve the 2x2x1 blocks, its not a problem, I think.

Perhaps my English is too much poor. I must make a video, I think.

Cheers,
Andrea

Better and faster exchanging two 2x2x2 blocks.

I developed this sequence in past ( perhaps 15 years ago). The sequence, like an idea of ultimate solution for big-cubes can exchange corners, edges or groups of edges.
It exchanges the right top corners and the corners in down layer.
The sequence is self-inverse!

sequence= F D' R D2 R' D F'

try it, its very fast !

To make a clean outer 2x2x1 exchange make
U sequence U sequence U
To make a clean inner 2x2x1 exchange make
uU' sequence uU' sequence uU'

(uU' in multislice notation is u in wca notation)

perhaps this makes it a little bit smoother to solve the ai 444.
advantage is, that no cube-rotation is included.

Cheers,
Andrea

Hi Andrea,

Thanks for sharing all your sequences. I like your patterns more than mine. I didn't use a sequence to generate them this time, in this case I found it easier just to solve to that state.

I got my AI cube (white ) two days ago.
Andrea, thank you for sharing your 2x2x1 swap!

Here is my move sequence for swapping 2x2x1 blocks:
I'll use WCA notation:
e.g. u = inner slice below U; Uw = (U u) (Uw spoken as U wide is a synonym to 2U in WCA as well). The SiGN notation Andrea is referring to is quite different (e.g. 2U is the inner layer below U and u are two layers together). I think it is less confusing if we stick with one standard wthin one thread (instead of three, WCA, Randlshofer, SiGN ).

I use on almost any cuboids a pure corner 3-cycle:
F2 (R2 U' L2 U)x2 F2 (10 moves) which is the cuboid version of the Corner Piece Series (CPS) R U' L' U R' U' L U.
It can be used on cuboids with edges too achieving a pure corner 3-cycle.

As I want to 3-cycle 2x2x2 blocks, it becomes Fw2 (Rw2 Uw' Lw2 Uw)x2 Fw2.
Now I add two outer layer moves to get a 2x2x1 swap: U2 Fw2 (Rw2 Uw' Lw2 Uw)x2 Fw2 U (12 turns).
This swaps the 2x2x1 blocks originally located at UFL and UFR relative to the u layer blocks.


I use front-view / back-view pictures as in Gelatinbrain.
Situation 1 shows the setup, situation 2 the result of the 12 moves sequence U2 Fw2 (Rw2 Uw' Lw2 Uw)x2 Fw2 U

Andrea's move sequence has the advantage that the reduced blocks remain at their original position.
It is a bit longer, if you do not care about the final location.
My sequence has the advantage that the basic sequence will look familiar to most solvers and it can be used on any cuboids as well.

My sequence for swapping two 2x2x1 blocks is I think quite similar to Andrea's. It's very intuitive.

Hold the two needing swapping at UFL and UFR.

Starting with these (showing the UFL and UFR needing swapping)


and then carrying out

U2 y Ll Dd (Ll)' (Rr)' (Dd)' Rr Ll Dd (Ll)' U' y'

give us




I worked this out when I was solving the cubic 336, and realised it'd apply to this as well. Of course, i could leave out the y and y' and alter the corner swap in the middle, but this is what works for me.

Phew!

I have just spent the best part of a week working my way through the AI cube. When I say a week, i mean that my dayjob got in the way, so probably "a number of hours" is more accurate. I take longer than some others around here to "grok" things. And I decided I needed to get my head around it. I've appreciated Burgo's original example video, and i'm also very grateful to him for helping me get my head around the main concept of the colour matching, via PM. That was the thing that was missing from my understanding.

Once I "saw" what needed to be done, it made figuring out the right approach for each 4th block case a lot easier. The hardest one was the twisted corner, but I got there today. I really feel like getting my head around this cube has improved my solving skills, which is strange, since there's only one sequence being used. It's more in the sense of thinking through why to do what when.

As usual, I've made some videos showing the whole process from start to finish, including 6 cases of a 4th block scenario. And as usual, the videos are geared towards those who need things explained (hopefully) simply and clearly. I have no doubt that some of the methods may not be the most efficient, but I think they're ok.

The videos with written tutorial are here.

The youtube playlist is here.

Here's the video for the single twisted corner.

And finally, could prof_grizzlebizzle consider changing the title of this thread?

Hi rline,

your video

Great work. Respect. Good description.

Pattern:
4 Zig Zag and 2 checkerboard in checkerboard

Hi rline and all other

perhaps this is one of the hardest cases:

Hi Andrea,
nice pattern!
Thinking in two swaps of edges makes your case pretty easy.

Here is an applet visualisation of this pattern:

Zig-Zag 4 sides (15 btm).

A bit optimised for good measure

Per

This has rapidly become one of my current favourite cubes!
It doesn't look like much because of the big blocks but giving a scrambled one to a non-puzzler very rapidly they realise they cannot do anything at all with the small cubies!

Initial help from Konrad (via PM) has allowed me to solve it with some considerable effort! That man is a genius!!! I just do not know how he does it!! A 9 page pdf file of helpful hints, beautifully formatted with pictures! This provided me with some much needed techniques!

Then, Rline adds in his website instructions and I can combine the 2! I can now solve this cube from any position! I now don't even need to refer to notes for algorithms!!
Thanks very very much guys!!

Twisty Puzzles is an awesome resource, the boards and the people! I have sent several of my blog corespondents here to whet their appetites for twisty puzzles or get detailed information!!

Hi. I'm trying to avoid using other people's algorithms or solution videos for this one, so I'm really just looking for a hint. I can get 3 complete 2x2x2 blocks easily enough. The only thing useful I've found beyond this is an intuitive way to swap the top 2x2x1 section of two opposite 2x2x2 blocks. I'm wondering if this alone is sufficient to solve or if I should start looking for some other sequences.

You'll need a little more than just that unfortunately. What you need to be able to do is make the 2x1x1 fingers from the disparate pieces everywhere. Generally, you want to colour match edges with centers from a different block. Sometimes the corner is twisted so you need to make sure that ends up colour matching correctly with an edge as well. I don't think anyone could post sequences here for reducing it (Konrad has some sequences for 3-cycles above I think) because the turns you need to do depend on the colours you're trying to match. But for reducing it first to the 2x1x1 fingers, and then to the 2x2x1 blocks, you're generally trying to match colour without breaking much of what you've done already. (The only sequences you need are things like (F'RFR')x2 and variations, along with U turns.)

I'm aware this is still vague, but you only wanted hints. As you know, there are now a number of videos to look at if you need to.

Thanks. I kind of understand what you're saying. I'll keep working on it. I really like this puzzle and want to try to solve it on my own, but it's wearing me down.

UPDATE: I managed to finally solve it! I don't know if it was luck or not. I decided the key was to break a third 2x2x2 block [I had previously tried to keep two 2x2x2 blocks together all the time]. The first time I tried it, it worked! This is possibly the most complicated puzzle I have solved! Yay!

Hi Konrad,



Why the ?

Is the white color awful ?
My camuflage and badaged set arrived today on the custom office. One blue 1x1 is missing.

Cheers,
Andrea

Hi Andrea,

I saw your post in the "not so rare puzzles" thread about the missing blue tile.

I had ordered a black AI and the order was confirmed, but I got an email afterwards that the last black was broken. White is not bad, but I usually prefer black.

BTW, it seems that I'm the only one who thinks that solving the last two blocks of the AI via 3-cycles or (2-2 swaps) of edges and centres is very simple. I admit, though, that I did not really look at other methods talking about 2x1 fingers.

Hi friends,



Yes, I agree.

I watched the video from rline.
rline: Very good videos your cases 1 .. 6 !

The first three 2x2x2 blocks I solve like Burgo's method, 2x1x1 squares first.
The AI 4x4x4 is one of my favourite cubes.

Cheers,
Andrea

Andrea

Thanks very much for these words. You are a superb solver so this means a lot!

I agree with you that having worked out the AI, it is now definitely a favourite. It's a puzzle that is not too easy but requires a fair bit of strategy and thought. Well worth the money (if there was one I could actually buy!)