The New Sudoku Players' Forum

[DoubleB72]

I have the solution, but what is the first step from here?

**8 427 395
59* 61* *4*
*2* *59 6*1

7** 2*4 538
8** 1*5 **9
25* *** **6

9*2 5*1 *6*
*8* 942 *5*
**5 *36 9**

[Shazbot]

You haven't indicated how far you've got in eliminating candidates, so I'm starting from scratch here...

Code: Select all

 *-----------------------------------------------------------*
 | 16    16    8     | 4     2     7     | 3     9     5     |
 | 5     9     37    | 6     1     38    | 278   4     27    |
 | 34    2     347   | 38    5     9     | 6     78    1     |
 |-------------------+-------------------+-------------------|
 | 7     16    169   | 2     69    4     | 5     3     8     |
 | 8     346   346   | 1     67    5     | 247   27    9     |
 | 2     5     1349  | 378   789   38    | 147   17    6     |
 |-------------------+-------------------+-------------------|
 | 9     347   2     | 5     78    1     | 478   6     347   |
 | 136   8     1367  | 9     4     2     | 17    5     37    |
 | 14    147   5     | 78    3     6     | 9     1278  247   |
 *-----------------------------------------------------------*


Your (16) naked pair in column 2 will allow you to eliminate those numbers from the rest of the cells in column 2.

In box 6, 4s are locked to column 7, so you can eliminate 4 from r7c7, leaving you with a naked pair in row 7, followed by a hidden single in column 2.

Everything will flow nicely from there with just naked and hidden singles.

[Brendan]

You have
R3C4 = 38
R2C6 = 38
R6C6 = 38
R6C4 = 378

Uniqueness Theorum:
This would have multiple solutions unless R6C4 = 7

This breaks open the puzzle

Of course, there could be easier ones in there

[Shazbot]

wouldn't that require an assumption that the puzzle HAS a unique solution?

[Brendan]

wouldn't that require an assumption that the puzzle HAS a unique solution?

Yes, it does. There is a strong school in Sudoku that consider that a genuine Sudoku only has a single solution.

Brendan

[MCC]

Yes.

Edit: Pipped at the post by Brendan.

[rubylips]

I don't think that's a valid Uniqueness pattern - the two '38' cells in Box 2 would have to lie in the same row. As things stand, the values in these two cells couldn't be switched independently of the 3s and 8s in Boxes 1 and 3. Please correct me if I'm wrong.

[Shazbot]

I agree that a VALID sudoku should have only one solution, but there are some floating around that don't. So unless you know where the puzzle is from, and know that the source ONLY releases unique puzzles, I'm not sure it's a great idea to make that assumption, is it?

Do you find it better to assume that it's unique until you get totally stuck and THEN consider there may be multiple solutions (if you don't have a program that tells you immediately), or to keep the possibility in the back of your mind that it may NOT be unique?

[Brendan]

I don't think that's a valid Uniqueness pattern - the two '38' cells in Box 2 would have to lie in the same row. As things stand, the values in these two cells couldn't be switched independently of the 3s and 8s in Boxes 1 and 3. Please correct me if I'm wrong.

I think that you may be right. I have over generalised the technique.

Brendan

[stuartn]

The 3's in B5 are in one row only - the 4's in B6 are in one col only. These both facilitate terminating exclusions.

stuartn