| Masyu please! |
ferkel
Kwon-Tom Obsessive
Puzzles: 536
Best Total: 39m 18s | Posted - 2010.05.28 17:22:11
Hello,
I've been a fan of Kwon-Tom Loop for a while now. Like many of us here I've wasted many a happy hour drawing lines. *Thankyou*
Recently I came across Masyu, which has all the appeal of Slitherlink but is just a bit different. It's still a loop-based game, so it fits with Kwon-Tom Loop, right?
Having exhausted Nikoli's sample puzzles and failing to find any decent software generator, I wondered if Kwon-Tom Loop might start providing Masyu puzzles.
I know I'm asking a lot. Thanks!
Ferkel |
Zyntax
Kwon-Tom Obsessive
Puzzles: 6534
Best Total: 13m 6s | Posted - 2010.05.29 00:37:10
Check out www.janko.at
It has lots of puzzles, including Masyu.
The site is in German but you should be able to find your way around. |
MondSemmel
Kwon-Tom Obsessive
Puzzles: 6159
Best Total: 7m 47s | Posted - 2010.05.29 03:20:28
Quote: Originally Posted by Zyntax Check out www.janko.at
It has lots of puzzles, including Masyu.
The site is in German but you should be able to find your way around. |
The section about line puzzles might be of interest:
Line puzzles
(Almost) All puzzles on this website (e.g. Masyu/1, the first Masyu version, here) display puzzles sorted by:
a) when they were uploaded (e.g. numbers in ascending order) ("Aufgaben und Lösungen");
b) sorted by size ("Sortiert nach Größe")
and c) sorted by difficulty ("Sortiert nach Schwierigkeitsgrad"); although "difficulty" only refers to solving time, nothing else - and because it's not weighed by size, bigger puzzles will always be considered "more difficult"...
I recommend trying to solve puzzles by their difficulty rating - that way, you can start learning a new puzzle type with very easy, almost trivial puzzles, and then the difficulty gradually increases as you solve more and more puzzles.
Last edited by MondSemmel - 2010.05.29 05:03:58 |
jamin
Kwon-Tom Obsessive
Puzzles: 1233
Best Total: 27m 9s | Posted - 2010.05.29 04:48:08
Thanks for the www.janko.at tip.
Masyu is a great game.
Last edited by jamin - 2010.05.29 04:48:31 |
foilman
Kwon-Tom Admin
Puzzles: 3614
Best Total: 24m 6s | Posted - 2010.06.02 08:32:55
I've played a few of the puzzles and I like them! As to putting them on here, it's a nice idea and if I can find enough time I might have a go... however, with a new baby about to arrive in the family in two weeks I might just have my hands full for a while!!
If I do put Masyu puzzles on here, I'll have to have a think about a site redesign too. I guess we'd need separate pages for just about everything... |
mathmaniac
Kwon-Tom Obsessive
Puzzles: 1293
Best Total: 20m 57s | Posted - 2010.06.02 13:42:22
Haha, yes I played those and I liked them, just be wary Foilman, or you will end up with Sudoku and every other Asian named logic puzzle on here. Not that I'm against it, but it might be a little much for you from the sound of it. That having been said, there are surprisingly few good sites for these games and I just know that you could improve upon them, so if you do have the time, I'm all for it. |
Brian
Kwon-Tom Obsessive
Puzzles: 4907
Best Total: 9m 6s | Posted - 2010.06.03 21:20:47
These are pretty neat. While searching for more of these puzzles, I came across a blog with some variations on slitherlink:
http://mellowmelon.wordpress.com/category/puzzle-types/slitherlink/
The slitherlinks and crosslinks are easy (compared to the puzzles here), but the other ones are interesting. |
MondSemmel
Kwon-Tom Obsessive
Puzzles: 6159
Best Total: 7m 47s | Posted - 2010.06.03 23:23:57
Quote: Originally Posted by Brian |
Thanks for the link. These puzzles look pretty interesting, but they obviously have the drawback of not being solvable online (like on janko.at or here at kwontomloop.com). How do you actually solve these puzzles? Do you print them? Do you use programs like Microsoft Paint to simply scribble the solution onto the image (e.g. on this one)? |
Brian
Kwon-Tom Obsessive
Puzzles: 4907
Best Total: 9m 6s | Posted - 2010.06.03 23:45:57
Quote: Originally Posted by MondSemmel How do you actually solve these puzzles? Do you print them? |
Quote: Do you use programs like Microsoft Paint to simply scribble the solution onto the image? |
|
Para
Kwon-Tom Obsessive
Puzzles: 1923
Best Total: 19m 28s | Posted - 2010.06.04 17:52:23
You can also try puzzlepicnic.com if you want to play more online. There's a whole bunch of blogs with nikoli style puzzles. http://mathgrant.blogspot.com/ features a whole bunch of Masyu, just named differently. Some of his are also found on puzzlepicnic in an online format.
Last edited by Para - 2010.06.04 17:52:53 |
Brian
Kwon-Tom Obsessive
Puzzles: 4907
Best Total: 9m 6s | Posted - 2010.06.04 22:06:58
If you do end up making a program, foilman, I'd urge you to not copy everyone else in using a gray grid. Why do they do that? All it does it make it harder to follow paths when you're trying to avoid closed loops. Instead you should just put dots in the center of the squares. Then you could also have crosses, for when you know the path doesn't go in a particular direction. This is what I mean:
 |
mathmaniac
Kwon-Tom Obsessive
Puzzles: 1293
Best Total: 20m 57s | Posted - 2010.06.05 01:04:59
Yes I personally was downtrodden by the lack of accomodation of crosses. Although the puzzles are, in my oppinion, easier to solve without crosses than slitherlinks, the harder ones could use the feature. |
Tilps
Kwon-Tom Obsessive
Puzzles: 6720
Best Total: 18m 37s | Posted - 2010.06.05 12:49:26
Proof of concept.
Silverlight version of Masyu
I hope to actually make it generate puzzles which are not only solvable, but have a controllable level of difficulty in the next couple of days, but for now it has a chance (<10%?) of giving you a solvable puzzle, displayed as Brian suggests (which I think is the best way).
(I'm calling my implementation Twist'N'Turn) I'd keep working on it tonight but I have google code jam in an hour.
If you want to challenge yourself, see if you can solve the puzzles under the assumption that non-marked intersections do not follow either of the two types of intersection marking. This will at least increase the percentage of puzzles which are solvable 
Last edited by Tilps - 2010.06.05 12:52:36 |
Brian
Kwon-Tom Obsessive
Puzzles: 4907
Best Total: 9m 6s | Posted - 2010.06.05 18:22:14
Unfortunately, I'm on Ubuntu which doesn't support Silverlight. Moonlight works with some sites, but apparently not yours, Tilps. But I'm sure others will have fun with your program. |
Tilps
Kwon-Tom Obsessive
Puzzles: 6720
Best Total: 18m 37s | Posted - 2010.06.06 01:12:38
Yeah, I'm using silverlight 4 - moonlight is only silverlight 3 compatible even using the latest beta. They claim they will have alpha silverlight 4 compatibility by the end of the year.
Last edited by Tilps - 2010.06.06 01:13:31 |
Para
Kwon-Tom Obsessive
Puzzles: 1923
Best Total: 19m 28s | Posted - 2010.06.06 01:19:05
If you made this generator like a slitherlink generator where you first draw a loop and then check if it forms a unique Masyu puzzle I can inform you that that isn't the way to go. You prob discovered the main problem in that, which is that not every loop has a unique Masyu puzzle opposed to slitherlink where almost every loop has a unique puzzle in it. It's probably one of the reason there aren't any Masyu generators out there already. The only way I can imagine it working is a generator that places balls one by one and checks if there still is a solution and keeps going till it finds a unique one.
Creating Masyus is really a deductive path and has very little hope of being unique with a recursive path.
I never liked these total Masyu puzzles where all balls are given and when there is no ball neither is true. The path always has this set curving path through open areas without a ball. Kinda predictable.
ps. I don't agree that the normal grid lay out is bad. If you give someone a square grid and ask them to draw a loop in it, most people will draw a loop that goes through the centers of squares opposed over the lines of a grid. The reason this is done different in slitherlink is because of the way the clues are given. Also you can easily use crosses as you can just put a cross on a line to indicate that line isn't crossed by the loop. Puzzlepicnic.com has crosses implemented in their script for Masyu. |
Brian
Kwon-Tom Obsessive
Puzzles: 4907
Best Total: 9m 6s | Posted - 2010.06.06 02:24:50
Quote: Originally Posted by Para ps. I don't agree that the normal grid lay out is bad. If you give someone a square grid and ask them to draw a loop in it, most people will draw a loop that goes through the centers of squares opposed over the lines of a grid. |
Here's a picture of what I mean by it being harder to follow the paths. Do the two red dots belong to the same path? It seems easier to determine that when the gray lines aren't there. (It's even worse on other sites, where the grid is black and thicker.) I'd also prefer to use a thin black line for the loop, and smaller black and white circles--probably something quite similar to how the slitherlinks look here. (I can't comment on Tilps' implementation since I haven't seen it.)
Last edited by Brian - 2010.06.06 02:25:04 |
Tilps
Kwon-Tom Obsessive
Puzzles: 6720
Best Total: 18m 37s | Posted - 2010.06.06 03:07:49
I may try a build up generator rather than a tear down, but since there are so few clues to potentially remove, the cost of repeatedly generating loops until you get one which sticks is probably fine for smaller grids. I guess I'll find out later. |
Brian
Kwon-Tom Obsessive
Puzzles: 4907
Best Total: 9m 6s | Posted - 2010.06.06 07:16:56
I tried my hand at constructing one of these puzzles. Here's the result:
 |
Tilps
Kwon-Tom Obsessive
Puzzles: 6720
Best Total: 18m 37s | Posted - 2010.06.06 09:25:11
My silverlight version is now generating actual solvable puzzles, generation time is acceptable for grids of size 6x6, 8x8 is pretty slow.
For the moment the puzzles are pretty easy, some of the game rules aren't even properly programmed in the logic solver, just in place by virtue of it refusing to close the loop if the puzzle is not correctly solved. Once I get the game rules in more fully, I'll go back and look at making the generation faster. |
