I think first you need to scramble it! And thats a hard job too!

How about some good old Pochmann-style super stickers?

I'm thinking more of a Wacky Woolies 17x17x17 at the moment.

Hi Twisty Puzzles fans,

Here are a YouTube video and some photos of what Over The Top- 17x17x17 looks from the inside.

Enjoy!

Oskar

Amazing! LHC comes to mind. Out of interest, how long was the assembly process?

It took me a bit less than an hour to sort the orange pieces and to get them in place again.

Oskar

Why did you make the corners and edges all black?
To make the challenge a bit easier perhaps?

All pieces are colored by putting them into boiling dye. This results in a single even color. I chose black for the corners and edges, because that is the most neutral color for them.

Oskar

I'm blown away Oskar. Why do you assemble the puzzle outside? I think if you ever did a V4, you should design the edges and corners with recessed faces and design tiles for them, so that it is a 17x17x17 both mechanically and solving wise.

NICE!!! Very cool pics!

I have a question. I understand the holes the the edges and corners are there to cut down on the material costs. What about the holes in the long anchors? I see each has one and they are at various depths. Was this intended to help with the sorting process?

Also are the parts hollowed?

And I know we just picked the puzzle of the year for 2010 but I can't help but think I'm already looking at the puzzle of the year for 2011. If not, I can't wait to see what beats this.

Carl

I agree. The edges should either have tiles or stickers, so it's a real 17x17x17, instead of a babyface 15x15x15.

This is going to end up in a museum.

Correct.Yes, also the floating-anchor pieces are hollowed to a wall thickness of 0.7 mm. Unfortunately, most of the volume of these pieces is in the walls, so it does not save too much material ... Unlikely. By the end of 2011, all the excitement over the 17x17x17 will be over


Oskar

Wow Oskar, that mech is unbelievably complex! It was genius to extend the parts inward for stability...man what an awesome puzzle.

How can you distinguish all the pieces? Is there somekind of mark on each pieces or something?
They look so similar to me.

Oskar, is this structure pagoda-style or linear-style? It seems to me that it's pretty linear.

All pieces have tiny holes at different heights at their stems. This is how I distinguish them. Next version will have 3D-printed numbers in them.It is a linear style that I call "floating anchors", see here. The anchors keep all pieces straight up.

Oskar

Oskar, I have a question: why did you choose such center piece form - with plate under all the middle pieces? Why did not use simple idea like to put anchors onto the corner`s down-plate elements? - in this case it will be better turning quality, because of outer layer almost always turns with at least one inner layer...
The only thought I have of your such decision - to mush easier assembling/disassembling.

Adding to Tesseract's question, would the puzzle quality improve with a much larger core and thus shorter center-stalks?

No. The long stalks, which I call "floating anchors", are essential to keep the pieces upright. The notches prevent the pieces from sliding out, whereas the floating anchors prevents a slice of pieces from buckling while turning. This is why this design is absolutely unpopable. The pieces are even hard to remove when the screws are loosened.I apologize, but I do not understand your question.

Oskar

I believe he's asking about the plate where the floating anchors 'rest' on the center piece, where the big edges hold on to. Maybe if you had the edges hold on to the anchors so they go along with the center pieces, the puzzle will turn nicer overall because then the edges wouldn't 'fall behind' during the turns.

@ Oskar,
Yes, I mean: why did you use 01.png variant instead of 02.png variant?

This means there is a small 3x3 inside - so I think this gives it a better stability.

@ EMI94100
really? so why v-cube use 02.png-version?
note: my sketches are simply schematic - for illustrate Center`s vs Corner`s plate to "rest" floating anchors only.

The only thing keeping the pieces in in PNG 2 is the "notches" in the pieces.

The internal 3x3x3 keeps the edge pieces in, as far as I can see. Without them, the puzzle would be more prone to: major pops (disintegration)

This is as far as I can tell, I know that you and Oskar have much more CAD experience so please tell me if I am wrong.

Yes, U`re absolutely right! I`m so ashamed that didn`t surmise by myself...
My question is closed now.

p.s.
> I know that you and Oskar have much more CAD experience so please tell me if I am wrong.
Thanks! I`m in progress now a huge CAD presentation. You`ll see it just later. hush!

Like on the v-cubes, the centers also have those rounded "plates" below the surface - so both versions would work in my opinion (stability reasons).
also "why didn't the v-cube company use this mech?" Is the same kind of question as "Why didn't they use Pagoda?"

Tony Fisher asked me to make a video demonstrating the turning of the puzzle by making it checkerboard. Of course I accepted Tony's challenge. However, as Tony had not specified the checkerboard pattern, 3x3, 5x5 or 15x15, I decided to do them all. It took me over six minutes to make all turns. Checkerboarding feels like a mild form of torture, as I had muscle cramps afterwards. Scrambling the 17x17x17 is definitely a good workout.

Watch the YouTube video of me checkerboarding the Over The Top 17x17x17.
Check out the photos below.

Enjoy!

Oskar

WOW! If you make an 18x18, you can do an 8x8 checkerboard and play!

Oskar, you rock!

-Eitan

I'll do a video of the scramble and solve if only you'd ship it to me.

EDIT: No, seriously, I would! If only.

Well Oskar, maybe a disassembly - assembly video?

That checkerboard pattern looks awesome.

Actually I can't wait to see the edges and corners stickered. Until then the number of positions is down to a lousy

437059247861723352482166578027668501027919735551491448460575409482049418464396
259728373782477454074300356464873536472734484804444772825344909728389465792623
609533511530713122731737484803955289910697710525871201397675130040743557351649
582724131665165714984023126594901985843595552065084335377553196598443445516561
167986552679283392368194334161794997856412780669680115456946532090174764612412
821153646574386047359083443286825871264676308869659061063448130263646310869693
295405331958046825636151686900913601182075962004946587231584028248388141314235
491261797710191496564580911613862038209192238624600871161088789112346142894819
687996549747475141761000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000
00000000000

It's difficult to get an exact time from Oskar's videos but I estimate a typical slice move to be at least 30 seconds to complete. Multiply this by the number of moves required to solve one (anyone know this?) and you have a very very long video.

That's pretty slow, even by a non-speed-cuber's standards (I don't speedcube). Is this due to friction the pieces, or because performing the slices on such an overly large is cumbersome? Either way, some silicone lubricant should speed that up - I'd bet that puzzle's got more than enough material in it to soak up at least half a can!

Another thing that makes this difficult to turn is the complete innablility to corner cut.

That is why this mechanism is really not suitible for speedsolving, even if it was completely smooth. Every piece must be lined up perfectly or it will not turn.

I loved the checkerboard video, and I had another idea. If you don't want to scramble it, just make a 3x3x3 solve video with it, like the 3x3x3 checkerboard. Then to to the 5x5x5 solve, etc.

I can pretty much guarantee you, it's more than 20.

I think it's somewhere like 5-15 thousand moves, based off some simulators estimates ive seen.

In a video the re-aligning will be pretty dull after 20 moves. A stop-motion clip in which the re-aligning is removed will probably bring down a solving vid to 5000-15000 seconds (1.5-4 h).
I think I woulf still fast forward that.

Based on Tony's estimate of turning speed, that's between 52 and 157 solid days. Or between 78 and 234 16-hour days (with 8 hours for food and sleep).

N.B. I know you can all work this out... Also, this assumes one slice at a time, but ignores search time