HN Explorer

Show HN: Minesweeper on a chess board with chess pieces as mines

43 points by chedoku ↗ HN
rearrange the chess pieces so that the number of pieces threatening each square matches the large number in that square? pieces move freely

https://chedoku.com/#dailypuzzle

23 comments

[ 2.3 ms ] story [ 61.7 ms ] thread
Andugal
This is actually good! Well done :)
chedoku
I am very glad you like it and thanks for taking the time to write the comment :)
bspammer
This is really cool. I wish the pieces were off the board at the start though. I’d much prefer to build up a solution iteratively than try to fix a broken one.
chedoku
I totally agree. This was the technical limitation of the library that I was using (chessboardjs). If pieces are off board, there is no easy way to limit the number of pieces that you add to the board. Finding a better solution is on my todo list.
remarkEon
Do you think there’s a way to “rate” a particular solution? I’m wondering about how to gauge “efficiency” in placing the pieces. Fewest “wasted” threat squares maybe?
chedoku
I am not sure what this question has to do with the previous comment and why it is not asked in a separate thead! am I missing a point in your question?

Back to your question. There are two types of solutions: 1- Perfect solution, you get a green notification when you find the solution, in which other squares with no number have zero threats. Pressing the 'solution' button reveal the perfect solution 2- Imperfect solution, you get a yellow notification when you find one, in which other squares with no number may be under one or more threats.

As a simple example, for the puzzle below placing pawn on c2 is an imperfect solution but placing it on a2 is the perfect solution.

https://chedoku.com/#gid=3&d=1&

Is perfect solution in my terminology is what you mean by efficiency solution?

dado3212
Would also be awesome if there was deterministically only a single solution, though that may be hard to create.
smegsicle
there's a whole extra knight in today's puzzle

edit: i meant if you (r13: ebbx naq xavtug ba oynpx'f xvat'f ovfubc naq xavtug cbfvgvbaf) you only need to use two of the three provided

chedoku
It was promoted from a pawn ;)

joke aside, there can be: 1- any arbitrary number of pieces on the board 2- pawns in the first and last rows 3- missing king 4- missing opposite color pieces

chedoku
This is something that bugged me for a long time. There are many other challenging algorithmic questions that I will share in a separate blog post later (e.g. writing a efficient solver). If you have any idea how to do that I would love to hear about that.
shmageggy
Looks like this could be translated into a SAT problem and solved with a SAT solver. Have you tried that?
chedoku
I have actually used SAT solver before but I have no idea how to translate this into a SAT problem. I know it is too much to ask you to do it for me but maybe you can give me a hint (or link) on how to translate problems that deal with numbers into boolean expression that can be solved by a SAT solver.
shmageggy
I’m not an expert in it either, but this problem is similar to both the N-queens problem and Sudoku, both of which can be solved via SAT, so tutorials for those should be a good start. Good luck!
rcme
Also some explanation of what makes a solution "perfect" would be nice.
chedoku
Thanks for you comment. I am working on it now. 1- Perfect solution, you get a green notification when you find the one, in which other squares with no number have zero threats. Pressing the 'solution' button reveal the perfect solution

2- Imperfect solution, you get a yellow notification when you find one, in which other squares with no number may be under one or more threats.

A bit confusing so for now I try not to be strict about imperfect solutions.

[dead] akaa
[flagged]
shmageggy
This is really fun, and it definitely tickles that same part of the brain as sudoku.

One UI suggestion is to use a diverging color palette for the larger numbers so there are different hues for too many versus too few attackers.

chedoku
I am glad you like it. This is actually a great (and easy to implement) idea. I will fix this in the next release. Thanks for the suggestion
jarofgreen
I've been playing the daily puzzle for a week, it's fun. Well done :-)