Page 1 of 1

Corrupted Sudoku

Posted: Tue Jan 23, 2007 10:38 pm
by Ruud
Here is a Sudoku where some of the given digits were lost in transcription. I then solved as far as possible:

Code: Select all

. . 5|4 . .|9 8 .
. . 7|. 1 .|3 . .
. . 3|. . .|. . .
-----+-----+-----
5 9 .|. . .|. . .
3 . .|2 . 6|. . .
. . .|. . .|. . 3
-----+-----+-----
. 5 9|. . .|4 . .
. . .|9 8 .|5 . .
. 3 1|. . 7|. . .
According to a reliable program, it has 3627 different solutions.

I challenge you to figure out how this puzzle has been corrupted, fix it and find the correct solution anyway. :twisted:

Hint: All solved cells were singles.

Good luck, you'll need it!

Ruud

Posted: Tue Nov 20, 2007 8:31 pm
by Para
Has anyone ever been able to figure out this puzzle?

Para

Posted: Wed Nov 21, 2007 12:52 am
by Ruud
Digging up old memories...

For more info, read this post on the Player's forum.

It can be done.

Posted: Wed Nov 21, 2007 11:30 pm
by Para
Maybe i misunderstood the puzzle. I thought there was some specific way this puzzle has been corrupted. For example: if two givens are a knight move from eachother, one of the two is removed(that was my first try). Now i am wondering is there a specific reason certain clues are removed or are they just "randomly" removed(for no specific reason, except for the uniqueness assumption) and that these digits are the only ones that give a unique solution.

Because what i had concluded so far, assuming a symmetrical givens pattern, was: R1C3, R2C7, R5C1 and R7C2 were the placed digits. R2C6, R3C7 and R6C8(and possibly more) were the removed givens.
But after my knight move thingy failed, i was out of ideas.

Para

Posted: Thu Jan 31, 2008 4:23 am
by Smythe Dakota
Para wrote:Has anyone ever been able to figure out this puzzle? ....
I suppose one could begin by assuming the usual newspaper symmetry among the givens. That is, if there is a given at row R column C, there is also a given at row 10-R column 10-C.

In the stated puzzle, in addition to the original givens, there would also be givens at the positions marked X below:

Code: Select all

. . 5|4 . .|9 8 .
. . 7|. 1 X|3 . .
. . 3|. . .|X X .
-----+-----+-----
5 9 .|. . .|. . .
3 . .|2 . 6|. . X
. . .|. . .|. X 3
-----+-----+-----
. 5 9|. . .|4 . .
. . X|9 8 .|5 . .
. 3 1|. . 7|X . .
Maybe that's cheating, but I couldn't help but notice the large number of symmeterically placed givens in the puzzle.

Bill Smythe

Posted: Sat Feb 02, 2008 10:46 pm
by Para
Smythe Dakota wrote:
Para wrote:Has anyone ever been able to figure out this puzzle? ....
I suppose one could begin by assuming the usual newspaper symmetry among the givens. That is, if there is a given at row R column C, there is also a given at row 10-R column 10-C.

In the stated puzzle, in addition to the original givens, there would also be givens at the positions marked X below:

Code: Select all

. . 5|4 . .|9 8 .
. . 7|. 1 X|3 . .
. . 3|. . .|X X .
-----+-----+-----
5 9 .|. . .|. . .
3 . .|2 . 6|. . X
. . .|. . .|. X 3
-----+-----+-----
. 5 9|. . .|4 . .
. . X|9 8 .|5 . .
. 3 1|. . 7|X . .
Maybe that's cheating, but I couldn't help but notice the large number of symmeterically placed givens in the puzzle.

Bill Smythe
Actually i think there are less givens, because as Ruud stated a few singles were already filled. So i have a feeling this is the original puzzles format. Otherwise it would be too many givens to figure out. I think this is enough to make it unique. 3 givens is enough to work with.

Code: Select all

. . .|4 . .|9 8 .
. . 7|. 1 X|. . .
. . 3|. . .|X . .
-----+-----+-----
5 9 .|. . .|. . .
. . .|2 . 6|. . .
. . .|. . .|. X 3
-----+-----+-----
. . 9|. . .|4 . .
. . .|9 8 .|5 . .
. 3 1|. . 7|. . .
Filling only singles here without the X's gives you the grid given by Ruud

My problem with this puzzle is that it seems to be completely random why a number has been deleted. At least i can't find a logical process which explains which number has been deleted. The only rule behind the puzzle is probably that only one combination of 3 numbers guarantees a unique solution. Then it is not really a fun puzzle as proving uniqueness by hand for each number combination is a completely annoying. So this really seems to be a programmers puzzle. That is why i have given up on it. Maybe there is some logical process behind it, but i can't detect it at least.

The link that Ruud gave was also based on uniqueness, which is no fun by hand.