Thread: "The physical 2x2x2x2 as seen by a 'regular' cuber."

--------------210D1070D7D9B00D4C9AC76C
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: quoted-printable

Hello Brian,

This is very cool that you solved the puzzle, and especially nice to see wh=
at it looks like from someone approaching it purely as a 3D puzzle. Your de=
scription makes a wonderful introduction to our group!

Of course this stuff is difficult to describe in text and we're still feeli=
ng around for the best notations. I really appreciate your gyro video clips=
, and I hope you can be convinced to upload a video of showing a full solve=
with descriptions of what you're doing at each step. Don't worry about len=
gth if that's a burden. I'm sure that lots of us would love to watch it clo=
sely regardless of length.

Regarding layers, we've tended to call them "slices" in MagicCube4d. Twisti=
ng them is done by holding down number keys while twisting left or right. T=
he number '2' affects the second slice, '3' the third, etc. You can hold do=
wn any combination to compose a kind of chord or "slices mask" as it's call=
ed in the code. Those puzzles can be solved layer-by-layer or by piece type=
. Neither approach really applies to the physical 2^4 though.

Welcome and happy puzzling!
-Melinda

On 2/21/2018 6:06 PM, pentaquark394@yahoo.ca [4D_Cubing] wrote:
>
>
> Skimming through the video above, it seems the mains difference between o=
ut methods is that after what I call pre-orientation, you solve one L/R cel=
l directly, while I immediately separate and solve the two cells almost as =
two regular 2x2x2s. That means I also have to deal with the case where I ha=
ve to transfer corner twists between the L/R cells.
>
> Also, I tend not to use any moves on=C2=A0 I/O cells except the slice typ=
e moves. Here's how I approach pre-orientation.
>
> -Do L/R moves & a gyro-rotate so a favourable # of pieces(you actually wa=
nt a few) have the target colour "pointing out".
> -Get as many target colour stickers facing U/D, and move the pointing out=
stickers to one layer of one cell.
> -Twist non-sticker stickers so a gyro-rotate gives you a 2x2x2 oll case.
>
> A few more questions I thought of:
> What is RKT?
> In 3d cubing, a distincti on can be made between a layer and a face. How =
do you make this distinction in 4d?


--------------210D1070D7D9B00D4C9AC76C
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable



">


Hello Brian,



This is very cool that you solved the puzzle, and especially nice to
see what it looks like from someone approaching it purely as a 3D
puzzle. Your description makes a wonderful introduction to our
group!



Of course this stuff is difficult to describe in text and we're
still feeling around for the best notations. I really appreciate
your gyro video clips, and I hope you can be convinced to upload a
video of showing a full solve with descriptions of what you're doing
at each step. Don't worry about length if that's a burden. I'm sure
that lots of us would love to watch it closely regardless of length.>


Regarding layers, we've tended to call them "slices" in MagicCube4d.
Twisting them is done by holding down number keys while twisting
left or right. The number '2' affects the second slice, '3' the
third, etc. You can hold down any combination to compose a kind of
chord or "slices mask" as it's called in the code. Those puzzles can
be solved layer-by-layer or by piece type. Neither approach really
applies to the physical 2^4 though.



Welcome and happy puzzling!

-Melinda

Skimming through the video above, it seems the mains
difference between out methods is that after what I call
pre-orientation, you solve one L/R cell directly, while I
immediately separate and solve the two cells almost as two
regular 2x2x2s. That means I also have to deal with the case
where I have to transfer corner twists between the L/R cells.

Also, I tend not to use any moves on=C2=A0 I/O cells except the
slice type moves. Here's how I approach pre-orientation.=C2=A0>

-Do L/R moves & a gyro-rotate so a favourable # of
pieces(you actually want a few) have the target colour "pointing
out".

-Get as many target colour stickers facing U/D, and move the
pointing out stickers to one layer of one cell.=C2=A0

-Twist non-sticker stickers so a gyro-rotate gives you a
2x2x2 oll case.

A few more questions I thought of:

What is RKT?

In 3d cubing, a distincti on can be made between a layer and
a face. How do you make this distinction in 4d?

--f403045dceb66f90d80565d54b0f
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

That's pretty cool. I was thinking about doing it that way, but
transferring the twists is a pain in my opinion. I don't know which is
faster though.

Yeah, those I/O moves can be awkward, but I think they're easier to
understand than constantly rotating back and forth.

RKT refers to manipulating one cell as a 3d puzzle by using only rotations
of it and one adjacent face. The classic example in MC4D is manipulating
the I cell of an n^4 as an n^3 by only using I and R moves. The R moves are
in the axis perpendicular to the I cell, which is achieved by clicking the
out side of the R cell in MC4D. This has the effect of turning a layer of
the I cell, and rotating the I cell places different "faces" of it on the
right side to be turned with the R cell. It's the equivalent of turning
RUR'U' into R z R z' R' z R' z', if that makes any sense. Try it on a 3d
puzzle using rotations and R moves and you should get a feel for it.

The distinction between a face and a layer is very similar on a 4d puzzle
as on a 3d puzzle. A face would be solving all the stickers on one cell,
for example getting all yellow stickers on the right cell. A layer would
also solve all the adjacent stickers, for example my solve on one 2^3
directly on the physical puzzle. If I had only separated out the outer
colours, it would be a face, and the 2^3 not yet solved is still a face. I
hope that's clear.

~Luna

On 22 Feb 2018 07:24, "pentaquark394@yahoo.ca [4D_Cubing]" <
4D_Cubing@yahoogroups.com> wrote:



Skimming through the video above, it seems the mains difference between out
methods is that after what I call pre-orientation, you solve one L/R cell
directly, while I immediately separate and solve the two cells almost as
two regular 2x2x2s. That means I also have to deal with the case where I
have to transfer corner twists between the L/R cells.

Also, I tend not to use any moves on I/O cells except the slice type
moves. Here's how I approach pre-orientation.

-Do L/R moves & a gyro-rotate so a favourable # of pieces(you actually want
a few) have the target colour "pointing out".
-Get as many target colour stickers facing U/D, and move the pointing out
stickers to one layer of one cell.
-Twist non-sticker stickers so a gyro-rotate gives you a 2x2x2 oll case.

A few more questions I thought of:
What is RKT?
In 3d cubing, a distinction can be made between a layer and a face. How do
you make this distinction in 4d?



--f403045dceb66f90d80565d54b0f
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

--001a1143fc26f4d68c0566c89d44
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Changing the slab move doesn't make much of a difference, true. I guess
it's up to you if you want to bother. Right to left or left to right is
just preference too. Although the scrambles won't just be mirrors of each
other.

I also wrote a python program to generate scrambles in my notation, but
could you explain the logic behind yours? I can tell that specifying the
pieces by colour, but what are the numbers? Orientation?

I personally don't have the patience to reassemble the puzzle as a
scramble. I find hunting for the right piece surprisingly time-consuming,
and considering I do a lot of solving on the go, it's not really an option.
It does seem like a good method for a thorough scramble though.

~Luna


On 7 Mar 2018 01:30, "pentaquark394@yahoo.ca [4D_Cubing]" <
4D_Cubing@yahoogroups.com> wrote:


[Attachment(s) <#m_-8870901422857357622_TopText> from pentaquark394@yahoo.c=
a
[4D_Cubing] included below]

I don't think changing the slab move face meaningfully adds to the
scrambling, so might as well always do U. I also prefer moving the right
slice to the left.

There are 24^2 distinct elements, so counting tells us that 10 minimum are
needed to reach all states. Probably ~15 elements should be enough for a
through scramble.

Speaking of which, I wrote some code in Python to assemble a puzzle into a
solvable state.

https://pastebin.com/kY2v4pPg

Pic 1 shows what a solved puzzle might look like if it happened to come up
as a scramble.
Pic 2 shows how to put the puzzle back into its normal shape to solve.
Pic 3 shows what the scramble below looks like when implemented.

yrgv-02 wogm-30 yrbm-13 wrbm-21 wogv-03 yobm-31 wrbv-02 wobv-20

wobm-30 yogv-12 yogm-21 wrgv-21 yrbv-31 yrgm-32 yobv-21 wrgm-32



--001a1143fc26f4d68c0566c89d44
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

--f40304378ed0cd1d890566c8a86e
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

I just stacked a whole bunch of cubes on a desk as a makeshift tripod, if
that's any help. With how long it can take to explain the solves though,
comfort is important.

All my magnets seem to be fine, despite several drops on hard surfaces. I
hope you're not doing anything too crazy. =F0=9F=98=81

~Luna

On 7 Mar 2018 01:35, "pentaquark394@yahoo.ca [4D_Cubing]" <
4D_Cubing@yahoogroups.com> wrote:



I'd consider doing a walk-through solve when I improve my camera setup. You
don't see it in my gyro rotations, but my right arm is reaching around
uncomfortably.

Also, it may just be the way I turn the puzzle, I had problems with the
magnets escaping their places in the pieces. You may want to check for dead
spots in your puzzles.



--f40304378ed0cd1d890566c8a86e
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

--001a1142ee22cb98d70566d86ec8
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

I haven't had any magnet issues myself but wanted to recommend using this
adhesive instead of superglue. It worked so much
better for me on a different printed model, one where I was using tiny disk
magnets in a similar fashion. I think it is the extra viscosity that helps,
whereas superglue just seeped into the roughness of the surface and failed
to hold. The only downside is it does take 24 hours to cure.

Roice

On Wed, Mar 7, 2018 at 2:06 PM, pentaquark394@yahoo.ca [4D_Cubing] <
4D_Cubing@yahoogroups.com> wrote:

>
>
> I've had half a dozen or so come loose. I have a tendency to fiddle with
> the puzzle and slide pieces against each other. Orange parts seemed
> particularly vulnerable.
>
>=20
>

--001a1142ee22cb98d70566d86ec8
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable