Puzzle Suggestions

[bsguedes]

I've played a board game that made me have an idea of a puzzle. I don't know how it's complexity grows, but I think it's exponentialy. It's quite similar with Modulo and Tetris. I also don't know if there is a known algorithm that solves this problem directly.

Well, the idea is that we have a rectangular field, divided into squares. We have pieces (in my case it was wood pieces, of course that is not the situation here ) that must be positioned into that field to fill all the squares. The difference to Modulo is that pieces can be rotated and flipped here, so each piece can be attached to the board in eight different ways. Also, pieces cannot be placed over other pieces.

I think that this puzzle would be hard to solve by brute force, as we have a lot of possibilities for each piece. However, it is not that hard using the brain in low levels (like Tapeworm).

[adum]

that's an interesting idea. i wonder if people have analyzed this type of puzzle before?

[bsguedes]

I don't think so.

[bsguedes]

Hey, another puzzle idea, this time easier to implement, I guess.

When I was young, I've used to play a puzzle made of a board 4x4 with 15 squares named from 1 to 15 (and a free slot). The goal is to place the squares into the crescent sequence.

For example, if we start with:

9 7 1 3
6 10 4 8
2 15 11 13
14 5 12

We must get:

1 2 3 4
5 6 7 8
9 10 11 12
13 14 15

A little randomness can be inserted by doing board that are not necessarly squares (like 3x2 or 5x3) and with wall squares (squares where it is not possible to move).

All levels should have only one free square (although maybe in harder boards it could be interesting having 2 or even more free squares).

The level generation is not hard too, we need only to stard with a solved board to mix the pieces.

The solution submission is not too complex. With only one free square, all that we need is the direction of the piece that is being moved in relation with the free square.

[tails]

I think the puzzle is too famous (Wikipedia article), and too easy for computers to solve, even a larger board with walls.

It can be fun, though, to compete for the minimum number of moves in the King of the Hill style.

[nighthalk]

i found another possible challenge to insert, unfortinatly i didnt come up with the idea, the link is at http://www.kongregate.com/games/Eketek/eketekonnect

you have to route wires with the exact length, and crossovers, between blocks, its easy to verify but brute forcing would have solutions in 3^(30*20) range(each block only has 3 states, and not all of them are used), unfortinatly the puzzles themselves have to be designed, you cant have a very good scaling random puzzle generator. its difficulty is the exactness and packedness of the boards, with enough space/length its trivial to make a solution

[uws8505]

I like Tetris, so I thought of a puzzle similar to it.

There is a given 'field' partially filled with some blocks.
There are also some Tetris blocks(or some others) in a queue(like Modulo).
The objective of this puzzle is to clear the field using all Tetris blocks.
But there is a restriction: You should decide where and in which direction to drop each block, and a block which already started to fall cannot move horizontally nor rotate further.

The answer has two numbers(horizontal position and rotation) for each block.

For example, if the field is given as

Code: Select all

x...xxxx..
xx.xxxxxx.
xxxxxxxxx.

and the blocks are

Code: Select all

xxx
.x.

xxx
x..

then the answer would be 1081 or (1,0),(8,1), assuming horizontal position starting from the left as zero and clockwise rotation.

This puzzle scales really fast with the number of blocks, at most x36 if the horizontal size is fixed to 10 (Think of L-tromino).

The generation of levels should not be so hard, since it can be done by (optionally) adding full line(s) and cropping suitable blocks from it.

[MrTerry]

since uws8505 revived this thread i really have to ask : will there be some new puzzles?

they are really fun! for me, more fun than the challenges and most of the time hard to tackle... i still stuck on the most of them (modulo, crossflip etc) but i would really appreciate new stuff... but i could understand if this would be to time consuming..anyway...i ll looking forward to the answer


cheers
Terry

[adum]

there will be new puzzles someday, but it's an awful lot of work to add one. all the level generation, client code, art work, site integration, server verification etc... so i'm only going to add a puzzle if it really is something special at this point. specifically, i want puzzles where there's a wide range of solver strengths. something like Runaway Robot or One of Us, you basically either solve it or you don't. puzzles like Mortal Coil, Modul, and Bricolage have more depth.

[CodeX]

is everything for a game just PHP possibly with MySQL (along with anything extra for the graphical side of things)? Are custom PHP extensions or executables an option for things like on the fly level generation/solution checking?

[bsguedes]

sokoban is a cool puzzle, but the main issue with it is that there already exist freely downloadable solvers. programs people have already put a lot of effort into, and they do a good job. so the joy of coming up with a cool new algorithm to solve a puzzle would probably not exist. generating boards is another possible challenge... but i think it's doable.

sokoban has been proven NP-hard, by the way. (many of our problems probably are. this isn't a problem in itself. you can still come up with really creative algorithms for np-hard puzzles.)

however, sokoban _variants_ are a very promising direction. there are a ton of ways to tweak or add rules to the basic sokoban game which change the gameplay significantly. so these things we've been thinking about... but haven't come up with anything firm yet.

thanks for the suggestion,
adum

There is a Sokoban variation I've seen in a Zelda game long time ago (maybe it was the game boy color ? I don't remember). The objective was to position boxes to identified spots, but the boxes were three-colored cubes ; a pair of opposed faces had a color, and the spots were marked with a unique color : you should position the cube with the good color in that spot; that made the problem incredibly hard... I think it could be a nice idea of puzzle, not so hardly generated and that scales well (number of boxes, number of colors, board size, etc...)

[Adzeye]

A Rush-Hour like sliding block games could be interesting, if it hasn't been suggested before.

[Hckr]

Another game from Kongregate:
http://www.kongregate.com/games/MrJinx/ ... nture-ep-1

The game is about an adventurer running through a dungeon according to fixed rules. You can influence his behaviour by placing gold to lure him from the designated path. Also you can put one-way gates to block a path once it was passed.

This looks very interesting, what do you think?

[Allosentient]

I could picture this picture appearing easy at first, then becoming insanely hard later on. But I wouldn't know off the bat. It seems like you'd need a much more complex and/or a larger game board to be worthy of this site

Note: I didn't play the game all the way through so maybe it already gets to that point later on

[Allosentient]

Yeah I made it to the first bonus room and it already looks pretty damn hard for both a computer and a person to solve, so might work