Page 3 of 3

Re: What puzzles can I add to the wiki ?

Posted: August 19th, 2011, 8:47 am
by Paradox
The tutorial is looking awesome so far!

Re: What puzzles can I add to the wiki ?

Posted: August 26th, 2011, 12:25 am
by Pio2001
Hello,
I've created the page for Philippe Dubois' Burr, but I then saw that it was linked as Gaby Games from the Aluminium Burrs page.

I'm used to call it the "Philippe Dubois", like in IBM's webpage. Do you know which name is best known in the puzzle community ?

Re: What puzzles can I add to the wiki ?

Posted: August 27th, 2011, 6:33 pm
by bluesign2k
I think its best left as "Philippe Dubois' Burr", I can't see many places listing it as Gaby Games :)

Re: What puzzles can I add to the wiki ?

Posted: September 14th, 2011, 10:43 pm
by Pio2001
Hello,
What do you think that the picture captions should look like ?

For the time being, we use to set <Puzzle by Manufacturer, photo courtesy of Photograph> for the main picture, and <Puzzle in Woods, by Manufacturer, photo courtesy of Photograph> for the galleries.

Isn't it giving more credit to the manufacturer rather than to the designer ? I used to set <Designer's Puzzle in Woods, by Manufacturer, photo courtesy of Photograph>, but it was too long, it seems.

Also, as a puzzlist, I'm a bit concerned with practically unsolvable puzzles. Some are completely unsolvable for a human being, disassembly as well as assembly, like the Harmonica or the Zig Zag 2, but most burrs recently published are just unsolvable in the assembling direction (Phelan, Xenon, Tiros (original color pattern), most burrs by Stéphane Chomine or Alfons Eyckmans...)

Strictly speaking, these puzzles are Take Apart puzzles ! Actually, the right category would be Take Apart Burrs, or Take Apart Interlocking.
What do you think ?

Re: What puzzles can I add to the wiki ?

Posted: September 14th, 2011, 11:19 pm
by Paradox
With regards to the picture captions I feel that stating the manufacturer is better than stating the designer. the designer is always the same and is mentioned in the main body of the page. The manufacturer however varies, and it's best to say who made that particular puzzle in preference as the designer can always be assumed.

What's your concern with the stupidly high level burrs? Is it with the category assignment? Although they are improbable to solve they are not impossible. So I feel they should still be classified to their normal burr category.

What do you reckon?

Re: What puzzles can I add to the wiki ?

Posted: September 15th, 2011, 12:21 am
by Pio2001
Ok for the pictures. Since I rarely post many of them for a given puzzle, I didn't think about the galleries involving many different manufacturers.
Paradox wrote:What's your concern with the stupidly high level burrs? Is it with the category assignment? Although they are improbable to solve they are not impossible. So I feel they should still be classified to their normal burr category.

What do you reckon?
My concern is with the game. Looking for a solution that can't be found is not interesting.

To illustrate this, I'll take two examples that i own in my collection : Croix de 36, by Jacques Frossard. It is a 36 pieces traditionnal cross (the size above 18, if you follow me). Opening is ok. It is level 1.
Rebuilding the puzzle is impossible. Analyzing thousands of configuration each second, infinitely more than a human could, Burrtools estimates the resolution time to more than one million years.
If the puzzle is open and I lose the solution, the puzzle can never be reassembled again !
It was designed assembled. If the designer then destroys the plan, taking the pieces apart becomes an irreversible process !

This doesn't stop there. Analyzing the Top Ten, by Frank Worrell, Burrtools eliminates false assemblies one after another, but when it comes across the right one, it takes it 20 minutes of calculus to open the puzzle !
Ok, this doesn't proves that it is out of reach because Burrtools doesn't look for a solution, but for the shortest solution.

The number of possible configurations of an N-piece puzzle is nonetheless of the order of N! (more than e to the power N !).
If we take the most complex 6-piece cross, for example, and imagine that it takes one hour to solve. With one more piece, if we keep things the most complex possible, there will be 7 positions to try, let's say times 4 for a piece with two possible backs. The resolution times jumps to 1 x 4 x 7 = 28 hours.
One more piece, 28 x 4 x 8 = 896 hours.
9 pieces, 32 000 hours
10 pieces, 1 million hours
etc

In practice these numbers are not met because only 6-pieces burrs have been completely analyzed (well, not completely, I'll talk about this in a dedicated thread), and the ones really needing one hour are just a few among 6 billions.

So yes, many puzzles published are really impossible, in the sense that it may take more than a human life to get a little chance to find the solution.

Donald Osselaer's Xenon, Fermium and Ultraburr are brillant examples of the contrary. Though Ultraburr is supposed to be a 610 moves puzzle, it is not so difficult to disassemble. Definitely easier than, say, a Revomaze.


You know, my proposition to create a "Take apart burr" sub-category was not very serious, but in a way, it would make sense. After all, a "Take apart" puzzle can be put together, though it's not part of the game. But that doesn't make it a Disentanglement puzzle.