The New Sudoku Players' Forum

[Minopoli]

Hi everybody,

A new variant sudoku is on line here

http://joujoujou.over-blog.com/article-25839168.html

It is a new concept and I wish to read your thoughts about it.

Thank you for any comment

[HATMAN]

Minopoli

This is an excellent puzzle type, a bit like a multi-layer repeat killer with a bossonova effect.

Unfortunately you made an error when you were getting the puzzle ready for posting and transposed the 6s and 8s in the bottom right four blocks and the puzzle as presented is infeasible.

The puzzle is nice and simple for a first one with interesting diagonal techniques. I guess that it would be unique with a bit more than half the clues and human solvable with 2/3 to 3/4 of them.

Given that it is in killer style I suggest you post on DJApe's forum or Richard's SudokuSolver one.

Thanks for an excellent and quite derivative puzzle.

Maurice

[HATMAN]

I really enjoyed the puzzle so I thought I'd have a go at creating one. Quite difficult to create the solution - creating the puzzle was not then too difficult.

Rules:
The yellow squares are a double sudoku that obeys the row, column rules the red boxes also contain one to nine in the yellow squares.

The white squares are clues the value is the sum of the orthogonally adjacent four (or less if on the edge or corner) yellow cells.

For killer addicts remember the sums can represent repeats if a block border is crossed.

Not too difficult once you get the feel for it.

[Smythe Dakota]

Clarification, please.

.... The yellow squares are a double sudoku that obeys the row, column and box rules. ....

I assume this means that the grid formed by the odd rows and odd columns is a sudoku, and likewise the grid formed by the even rows and even columns.

.... The white squares are clues the value is the sum of the adjacent four (or less) yellow cells. .... For killer addicts remember repeats are allowed. ....

I assume this means that the four yellow cells adjacent to any white cell (including those white cells with clues) need not contain four different values (although, obviously, they must contain at least two different values, except for the corner white cells).

Bill Smythe

[HATMAN]

Bill

On your first point generally no:
Each of the eighteen blue outlined blocks contains in the yellow cells the numbers one to nine.
Each of the eighteen rows contains the numbers one to nine in the yellow cells.
Each of the eighteen columns contains the numbers one to nine in the yellow cells.

The white cells contain clues. The number in the white cell is the sum of the orthogonally adjacent yellow squares. So the top right and bottom left clues are the sum of the two adjacent yellow cells. the edge clues are the sum of the three adacent cells and the centre one are the sum of four cells.

As far as this puzzle and Minopoli's one are concerned your first point is true, but that is a factor of how the puzzle was created and is not required in the solution of either of them. I'll probably use it in my next puzzle to minimise the number of clues.

Your second point is correct. If you have a clue of 7 on the edge which overlaps two boxes then 124 is not the only solution, 115,223 and 331 can also be solutions.

Please note I did this as a quick translation from the French so I hope I've got it right!

Maurice

[Smythe Dakota]

.... Each of the eighteen blue outlined blocks contains in the yellow cells the numbers one to nine. ....

OK, thanks. I didn't even notice the blue lines before -- they are thin and sort of faint.

.... As far as this puzzle and Minopoli's one are concerned your first point is true, but that is a factor of how the puzzle was created and is not required in the solution of either of them. ....

So it just happens to be true in this particular puzzle? Or is it a theorem that this constraint, even though not one of the rules, is a consequence of the stated constraints in all such puzzles?

Bill Smythe

[HATMAN]

I do not believe it is a theorem - but I have not tried to disprove it. It is easier to build the grid this way, you just parallel solve two JSudokus ensuring that they obey the Blue boxes.

Thinking about it I could do the same with magic squares and see how it comes out.

[udosuk]

I do not believe it is a theorem - but I have not tried to disprove it. It is easier to build the grid this way, you just parallel solve two JSudokus ensuring that they obey the Blue boxes.

If we don't care about the existence of a valid puzzle then it's very easy to construct a solution grid which follows the rows/columns/6x3 blocks constraints but doesn't follow the double 3x3 blocks, such as:

Code: Select all

1.4.2.3.5.7.6.8.9.
.7.4.8.9.1.5.2.3.6
2.5.3.1.6.8.4.9.7.
.8.5.9.7.2.6.3.1.4
3.6.1.2.4.9.5.7.8.
.9.6.7.8.3.4.1.2.5
4.7.5.6.8.1.9.2.3.
.1.7.2.3.4.8.5.6.9
5.8.6.4.9.2.7.3.1.
.2.8.3.1.5.9.6.4.7
6.9.4.5.7.3.8.1.2.
.3.9.1.2.6.7.4.5.8
7.1.8.9.2.4.3.5.6.
.4.1.5.6.7.2.8.9.3
8.2.9.7.3.5.1.6.4.
.5.2.6.4.8.3.9.7.1
9.3.7.8.1.6.2.4.5.
.6.3.4.5.9.1.7.8.2

People who need a solution pic or text formats of Maurice's puzzle (& solution), please visit djape's forum.

Maurice, hopefully you'll provide the text format for the next puzzle.

[evert]

With my limited knowledge of the French language, I read the original comment of Minopoli, and
my conclusion is that the following is the constraint:

- Numbers are given on the white squares.
- Numbers on the coloured squares must be filled in.
- Every number on a white square is the sum of all numbers on the surrounding coloured squares.
Is this a correct interpretation?
Did I miss anything?

Of course I appreciate any variant of the same idea, but at the moment I lost track...