It is either me or you seem to get some great scrambles for some of those puzzles.

Oh and by the way Doug, I got 1:15 on the Trajber's Octahdron. Not the first in moves though. I am not the fewest moves challenge type of person. I may work at it in the future though.

Adam

Solved http://users.skynet.be/gelatinbrain/Applets/Magic%20Polyhedra/octa_f4.htm and
http://users.skynet.be/gelatinbrain/Applets/Magic%20Polyhedra/octa_f5.htm.

Method:

1. Pair up the 2-stickered edge pieces
2. Solve like 4.1.3 (ignoring all pieces in the triangle centers)
a. Solve edges and centers of the four non-adjacent faces (not corners)
b. Match corners with the triangular (or trapezoidal in the 4.1.6) pieces not used by the solved faces
c. Solve like offset skewb diamond
3. Solve centers

A few algorithms I found to solve centers:

4.1.5:
Swap two pairs of outside centers with (U r U' l') (U r' U' l) 3x(r l' r' l)
Swap two pairs of direct centers 3x(r l' r' l)

4.1.6:
Swap two pairs of rhombus centers with (U r U' l') (U r' U' l) 3x(r l' r' l)
Rotate three triangle centers (u [D U' D'] u' [D U D'])

All symbols refer to the triangle face on the left view. U=top L=left R=right D=bottom. Lower case symbols mean shift click.

4.1.4 is the master skewb diamond and can be solved the same way as the master skewb by the way.

I'm not understanding Cambell's solution description. For instance "directly between corners" I can't translate into anything meaningful. Use of words like: edge, face, triangular, square, 1-sticker, 2-sticker, diamond, trapezoid, tip, inner, outer - would be a ton more helpful. It's more descriptive and my brain is wired all mathy. I find your turn counts for those two to be very impressive.

I'm stuck at 4.1.3 (Dino Octa). I solved it once in 44:39 using 410 turns,whereas you used 5:43 and 123 turns.

In order to advance to the higer-order versions, I need to master this guy first. Solve it in under 10 minutes is the goal seeing as I know what is possible from the 3 times above mine.

First, am I correct in assuming it's Skewb-like? A face-turing Octohedron bijectively-maps (mathematically analogous) to a vertex-turning Hexahedron (say the Skewb). Is it true it is Skewb-like? So how important is it that I master that first? I have tried it a bunch of times and struggled with it and planned to skip it as long as possible.

Now the way I did 4.1.3 was first to solve a "cross" of three edge pieces on a layer (ignoring the inner face-pieces, but not the central-centers). And then finish solving all those corners and edges using whatever intuition. Then I came up with a very lengthy alg that tended up doing two 3-cycles of the face pieces (these 6 pieces located in really odd areas). The alg was 26 turns long. Basically I 180 a corner's orientation twice, then I move it over using a single turn, and undo that to the other corner. Setting up these were hell. To do a 3-cycle, I'd have to 3-cycle three identical color pieces as well (and place them in the right spots while leaving the ones I want to cycle in theirs).

How many of you agree that the 8-sided guys need more "views"? I want to look straight at a Face!!! The restricted views confuse me so much that I never tried solving them back in '07. I also want a view that gives me a corner on top, bottom, left, right, and center (so like 45-deg rotated in the plane it currently uses).

-Doug

Sorry for making my solution confusing. I edited it slightly. I hope that helps. What I basically do to solve it is reduce it to a Dino Octa, solve the Dino Octa, then reduce to an offset Skewb Diamond, solve the offset Skewb Diamond, and solve the centers at the end. Solving the centers could be done directly after pairing the edges, reducing even more to a Dino Octa. Steps a, b, and c are for solving the Dino Octa (4.1.3).

Yes, the Dino Octa is skewb-like. 3.2.2, 4.1.4, and 5.1.7 are all variations of a master skewb. To solve these, I reduce them to their versions of an offset skewb and solve that the same way as a normal skewb.

4.1.2 and 4.1.3 can also be solved the same exact way. They are simpler, however, because of their shallower cuts. When they are reduced to the offset skewb diamond, they are like a pyraminx, meaning they are offset skewbs, but there are only four corners and six edges to worry about. In the case of the reduced 4.1.2 and 4.1.3, there are four faces and six corners to worry about. Also, 4.1.3 can be described as 4.1.2 with added fixed centers.

It is important to know how to solve Dino Octa before going for the higher-order versions if you want to use my solution. And I agree, more views on the octahedra would be useful.

Perhaps a better question to Campbell, is: how do you solve the Dino Octa? You have a HUGE turn-count lead over everybody else (the 3 others including myself who have solved it).

And what is a good Skewb method? And what do you mean by "offset skewb"? Is it what I mean by "skewb dual"? (A dual of a platonic solid is when you map vertices to faces and faces to verticies, the number of edges remains the same. c.f. http://en.wikipedia.org/wiki/Dual_polyhedron)

With all these variations on the Applet site, we need some precise terminology, and I suggest we stick to using mathematically-precise wordings to avoid confusions and also coin new words, but strictly define them when doing so.

Anyhow, does anyone want me to write up a solution to the Spherical puzzles?


-Doug

Yes. I plan on catching up sometime soon.

Doug, what I mean by offset skewb is this: http://www.twistypuzzles.com/cgi-bin/puzzle.cgi?pid=1354. I couldn't tell you a good skewb method. I'm sure the one I came up with is horrendous. You would have more luck Googling a method hehe.

Ok, I will try to make the best description I can of how I solve the Dino Octa. It's a little complicated.

There are 8 faces. However, certain colors can only be on certain faces. For instance: red, bright green, dark blue, and teal can only be together. Dark green, yellow, gold, and light blue pieces cannot be on the same face as any of these colors.

1. The first thing I do is solve one set of these faces (red,bright green,dark blue,teal or dark green,yellow,gold,light blue) except for the corners.

2. Let us call the direct center pieces centers and the triangular pieces that touch the edges at one point outer edges. After solving these faces (without the corners), 6 corners and a bunch of outer edges remain unsolved (you shouldn't need to worry about the other centers if you solved the trapezoid side pieces correctly). Match the corners up with the two outer edges. This stage is similar to solving edges on a 5x5x5 cube.

3.When you have gotten this far, you have reduced the Dino Octa to an octahedron version of a pyraminx with fixed centers. On a pyraminx, there are 6 edges and 4 corners. With the reduced Dino Octa, there are 6 corners (matched with outer edges) and 4 rotatable faces (these faces are the faces of a color set I mentioned. They are your solved faces). If you know how to solve a pyraminx, you can solve the rest.

Campbell,

By any chance are you Dave Campbell from CanadianCubing.com?

~Noah

Nope. Never even been to CanadianCubing.com

That was a fairly good description, and it cleared up a lot of things that I thougth might be true but wasn't sure. I like the idea of reducing it to a pyraminx-like thing. I had a few questions though.

How do you do step 1 itself, I got down to 2 outer-edges. do you use 5x5-edge-type tricks here as well?

nevermind, I figured it out! (Dino Octa)

Congrats, Doug. Nice move count too

I still have some catching up to do, but I showed this puzzle who is boss.



I spent 3 hours on it, then I realized I had to redo something, so I started over.

I can probably drop that time butt loads. Same with the turn count.

Now I just need to do 1.1.8 and I finished the face turning dodecahedrons.

I have got to do that on the puzzle that I hate at the moment. Good job. Now I have to work more on these puzzles.

new player in town, please post these somebody:

pyraminx, [00:00:01]
d26843d3f26043c8fb603dc27689ec13
5ea2bc436e91dc237a71f40ad926b34c
798ef20eb95e739ae90ed31d6770ce33
e51acb34e619cd32a7584fb09d623bc4

2x2, [00:00:01]
f59d0e3aad951c2fb59a6437c89e613d
c27489ec135ea1bc439a91dd237a85f4
0bd126b04c6186e40da1467b8cab16d3
2c6798ce31e51acb34e619cd32a7584f

megaminx, [00:01:38]
8776e5ed7a5ed5bc260c91dc237a85f4
0b9426b14c7986f20d4d46728ce916d3
2c6f98cd31fc1aca34e819e0323b584c
b09d623bc497682fd09b6437c89e613d

~elite

WOW

Looks like qqwref has some new competition! Those are some ridiculously fast times! I think all someone can do now is tie with you.

I have got to get on solving unsolved puzzles. Here I go...

EDIT: I haven't seen your scores pop up so I assume they won't be posted.

1:38 megaminx is believable. The 1 second things, well, were you just very lucky? Because you can't really get that time unless it's 2 moves or less, which requires some ridiculous luck. I see you only have 1 post on this forum so it makes me think you might have tried to hack the certificates rather than playing the game itself, but of course I have no way to know.

I decoded these certificates. It looks authentic.

pyraminx time= 1 moves= 3 scrambles=500
2x2x2 time= 1 moves= 2 scrambles=500
megaminx time=98 moves=589 scrambles=500

As for pyraminx and 2x2x2, I have to say he was extremely lucky.
I don't know how much is the exact probability to encounter such scrambles in a row.
Probably rarer than winning a jackpot or being hit by a meteorite.
The possibilty of the program's bug is not excluded.

As for megaminx, it means about 6 times/sec turns.
This too is hard to believe, but the computer clock too, it could go wrong sometimes.

His records don't appear on the page because they haven't been submitted nor mailed to me.
This is not because I suspect a cheating or anything. I don't care about the contents.
If they were sent normally, they would appear.

Lately I get many spams on my mailbox. So let me repeat once again the precaution to submit records:
· mail me the certificate only when your record dosen't appear within 24 hours.
· don't put anything other than the certificate and blank chars (space,tab,CR,LF) in the mail.
· no signature, no attachment, no html mail.
If your mail don't satisfy these condifions, your post will be rejected.

Precautins for naming:
·Use always the same ID.
·Be careful for miss typing(capital/small etc).
·Confirm your ID is not already used by someone else.
·Choose a ID nor too long nor too short (10~20 letters).
·No garbage in the name field.
(post this-is-my-first-time or I-better-on-real-cube things on the forum, please).

Thanks,

This is a post to say, getlatinbrain, I encourage you to try solve atleast one of your own puzzles I know you have a lot of fun making them, but you should try solve one!

I've been getting the above message in the load bar. I was wondering what I should do.

I generally come back the next day and it is fixed.

It could be, and likely is, a bot.

A person that is skilled enough at writting programs that can read screen pixel colors (easy) and execute clicks (with animation turned off), could come up with a simple human-based or LBL-stratagy programmed in, that would run overnight or something and out of hundreds of scrambles see one that is 2-turns from solved on 2x2. At a recent competition there was a 2x2 that was 5-turns (2-gen) away from solved and was sub-2s for some people. Moreover, much more than 6-clicks per second is possible using a bot like this - the lag would be from doing the computations at the start.

If done on multiple machines in parallel over many hours, times that appear "lucky" are completely to be expected.

I once wrote a pixel-reading-capable C++ program for Pentultimate... but had to execute the clicks manaually. And I once wrote an assembly-language (PPC) program to solve a 3x3 for a robot I did using breadth-first cross searching and F2L lookup-table.


-Doug

Ok, it's working for me now.

I'm still kinda bummed that no one really commented on my 1.1.6 solve.

I must have missed that post somehow, I just scrolled back up... That puzzle is a demon, congrats on that!

I got 4 corners on the Penultimate, then I gave up!

I did. Scroll back up to read my post. I almost solved a puzzle on gelatinbrain today at school too. I had to switch two pieces and I couldn't do it. I had to leave (next class). I will give it another go though.

Solved! 3.3.7 - http://users.skynet.be/gelatinbrain/...hexa_e3.htm

This is the puzzle I had trouble with. But I did it finally!! It looked interesting so I gave it a go. I feel happy that I did one on my own!

WOW! AWESOME FUSION! You're the first to solve the Little Chop I think! Tell us your method!!!

I decided to go at it again and got 12 mins instead! Awesome! I showed this puzzle who is boss!

My Method:

1 - Solve two faces, top and bottom (intuitive). The colors you choose don't matter.

2 - Use (UR UF UB) (UR UL UR) UB UR to switch triangle pieces on three faces without disturbing the top or bottom.
Use (UR UF UB) UF UR to switch two triangle pieces on opposite sides.
Use [FR (UF UR UF)]x6 to cycle three little triangle pieces on three faces.

That is it. That is my method, not for speed though. I will leave qqwref to that. Most of it is intuitive with these algs. If you have any questions please ask.

Now to find another puzzle to conquer!

EDIT: I just got 4:52 on it!

EDIT2: I just got 3:30.

You should try 4.3.3 too.
This is a shape&sticker variant of the Little Chop, but functionally not quite the same. Each of its 24 pieces is a unique combination of 2 colors, while the Little Chop is made of 6 sets of 4 identical pieces.

I've just added a new one (2x2x2 & LittleChop hybrid), the one requested by Doug a while ago.
http://users.skynet.be/gelatinbrain/App ... xa_fe0.htm
Actually it existed for ages, but just was not in the menu.
I think this too is a very interesting puzzle. Just like the spherical one, it consists of 48 pieces. Since the cut partially overlap with the edge, many sticker variants are possible(like Dogic and the spherical one). They will come later.

gelatinbrain, I'm impressed by your work . But, is it possible to change the colors of your puzzles? I have a hard time telling apart some of the colors you used .

If you don't mind me asking you, what's your educational background and what do you do for a living? (I'm guessing you work making 3D graphics).

Hello. I have been using gelatin brain for a while and it is definitely my favorite virtual polyhedra site for many reasons. So far I have solved these puzzles:

1.1.1 (Megaminx)
1.1.9 (Gigaminx)
1.1.12 (Impossiball)
1.2.1
2.2.1 (Lee Tutt's Pyracosaminx)
2.2.2 (Icosaminx)
2.2.11 (Alexander's Star (10-color))
3.1.1 (Pocket Cube)
3.1.2 (Rubik's Cube)
3.1.3 (Rubik's Revenge)
3.1.4 (Professor's Cube)
3.2.1 (Skewb)
3.2.4 (Dino Cube)
3.6.1 (2x2x2 Maze Cube)
4.2.1 (Trajber's Octahedron)
4.2.2 (Magic Octahedron)
5.1.1 (Pyraminx)
5.1.2 (Halpern-Meier Pyramid)
5.1.3 (Master Pyraminx)


I would like to bring up a couple problems that arise on macs:

1) right click doesn't work (on a mac, you push control and then click to make a right click)

2) the submit screen appears behind the puzzle (it is still possible to submit though)


Also, 5.1.1 is not a pyraminx. In the score rankings, the "Old Pyraminx(obsolete)" is a pyraminx. Why was this changed, because you had it right the first time.

Mac still doden't support multi-button mouse? It's surprising!
I'm sorry, but in this case I can do nothing. I want to reserve the ctrl-key for the future update.
But what on earth is the advantage of single-button mouse ? Or is it just a nostalgie for old good time? I'd like to ask Mr.Jobs...

I don't understand well. Do you have a screenshot?

The corners of Pyraminx are trivial, they have no reason to exist, at least not for solvers.
I dropped it because I don't want functionally identic puzzles with only trivial differences of appearance.

Who is the first to solve this Super-Pentultimate(with oriented centers)?
http://users.skynet.be/gelatinbrain/App ... ca_f14.htm

Even if the right click did work, I would still far prefer a regular mouse so this is not a big deal.


I have many. They are attached. It is hard to see but if you look between the left and right screen, you can see the submit box. It is in the same place that it pops up on a PC, it is just behind the puzzle, instead of in front of it.


I realize that the tips are trivial. However, if the tips are removed then I think of the puzzle as being shaped like this. I didn't realize that yours functioned the same way. I now see that it does. Thanks for the clarification.



I'm not!

Still can't get the Half Chop for the life of me. Any more hints?

What part are you having problems with?

idk about Noah, but I can't even get the top and bottom faces.

We're on the same page.

I got the top and bottom once by extreme luck. I spent a couple hours after that and still couldn't get a third face.

Yeah, I played with it for a while and couldn't get more than one face either. I dunno about this "get two faces intuitively" step

2.2.14 Master Icosahedron
48:34 834 moves.
LBL (lol) Maybe if I try again, I'll attempt a different method.

edit:
I took a few mini breaks on this one, and I was speaking to IRC at the time. I also didn't figure out a method first, I just went for it.

I will make a better guide this weekend. I don't have the time right now. You will better understand it then.

Adam

Gelatinbrain, if your taking requests, could we get some 14-sided puzzles? http://twistypuzzles.com/cgi-bin/pdb-se ... y=fourteen

Do you mean the 14-color Rainbow Cube? It is functional equivalent of the Skewb Diamond. I'm sorry to repeat the same answer, but what is the interest to program the same puzzle with insignificant difference of details?
For physical puzzles, I agree that the aesthetic is very important. But not for simulations, at least not for me. I rather want to discover new puzzles.

I realized that the rainbow cube was the same as that. As for the other 14-sided ones, I knew nothing about them really when I requested them. I just went and looked at them more closely and read the descriptions and apparently, they are all just equivalents of other puzzles so forget the request lol.

What about a master rainbow cube?

I would like to see that too, although I can't visualise it right now.

This is what he means.

http://twistypuzzles.com/cgi-bin/puzzle.cgi?pid=669

Actually Darren, that is a Master Cuboctahedron.

This is a Master Rainbow Cube.
http://v3.espacenet.com/textdraw?IDX=JP ... JP60222076

Noah, could you screenshot that page for me? It doesn't work on my comp.

here you go, joey.