The New Sudoku Players' Forum

[tarek]

One stepper to finish this Latin Square puzzle from this point. Anything spectacular?
Found this while rating a group of Latin square puzzles

Code: Select all

+----------------------------------------------------------------+
| 69     8      67     1      3      4      2      5      79     |
| 7      59     468    2      48     59     46     3      1      |
| 139    12     5      679    479    8      3479   69     2347   |
| 2      357    9      4      78     37     1      68     356    |
| 349    6      347    79     5      2      8      1      3479   |
| 8      13579  137    579    2      6      3579   4      3579   |
| 156    4      12     3      89     59     569    7      258    |
| 3459   3579   3478   5789   6      1      34579  2      345789 |
| 3459   23579  2348   56789  1      37     34579  689    345689 |
+----------------------------------------------------------------+

original puzzle

Code: Select all

.8.13425.7..2...31..5..8...2.94..1...6..5281.8...26.4..4.3...7.....61.2.....1....


[SpAce]

Would this work?

Code: Select all

+-------------------------------------------------------------------+
| 69     8       67     1        3      4      2      5      79     |
| 7     b59      468    2        4-8   b59     46     3      1      |
| 139    12      5      679      479    8      3479   69     2347   |
| 2     a735     9      4      a[8]7   a73     1      68     56-3   |
| 349    6       347    79       5      2      8      1      3479   |
| 8      1379-5  137    579      2      6      3579   4      3579   |
| 156    4       12     3      c(8)9   c59     56-9   7      258    |
| 3459   379-5   3478   5789     6      1      34579  2      345789 |
| 3459   2379-5  2348   56789    1      37     34579  689    345689 |
+-------------------------------------------------------------------+

(8=735)r4c562 - r2c2 = r2c6 - (5=98)r7c65 - loop => -8 r2c5, -5 r689c2, -3 r4c9, -9 r7c7; stte

...or another POV for the same pattern and eliminations:

Code: Select all

+------------------------------------------------------------------+
| 69     8       67     1       3      4      2      5      79     |
| 7    *[59]     468    2       4-8  *[59]    46     3      1      |
| 139    12      5      679     479    8      3479   69     2347   |
| 2    *[357]    9      4     *[78]  *[37]    1      68     56-3   |
| 349    6       347    79      5      2      8      1      3479   |
| 8      1379-5  137    579     2      6      3579   4      3579   |
| 156    4       12     3     *[89]  *[59]    56-9   7      258    |
| 3459   379-5   3478   5789    6      1      34579  2      345789 |
| 3459   2379-5  2348   56789   1      37     34579  689    345689 |
+------------------------------------------------------------------+

MSNS (35789): {2N26 4N256 7N56 \ 37r4 9r27 5c26 8c5} => same eliminations; stte

--
This is the solution I got with that, but I don't know if it has mistakes (no software support for this):

solution: Show

Code: Select all

.---------------------------.
| 9  8  6  1  3  4  2  5  7 |
| 7  9  8  2  4  5  6  3  1 |
| 1  2  5  7  9  8  3  6  4 |
| 2  5  9  4  7  3  1  8  6 |
| 4  6  7  9  5  2  8  1  3 |
| 8  1  3  5  2  6  7  4  9 |
| 6  4  1  3  8  9  5  7  2 |
| 3  7  4  8  6  1  9  2  5 |
| 5  3  2  6  1  7  4  9  8 |
'---------------------------'

[tarek]

Hi SpAce & well done,

Here is what I was after:



I can see that the variants section is not as popular as the sudoku section which I'll take not of next time

tarek

[SpAce]

Hi tarek,

Thanks for confirming my result! I think it was the first Latin Square puzzle I've solved. Then again, it's simpler than sudoku so not much of a stretch.

Looks like your loop is practically the same as mine, except for using a hidden instead of a naked route. I considered that too, but it was a bit longer to write and the loop eliminations are harder to see if written in the packed format:

(8)r4c5 = (865)r4c892 - (5=9)r2c2 - r2c6 = (98)r6c65 - loop => the same eliminations

I can see that the variants section is not as popular as the sudoku section which I'll take not of next time

I like a bit of variance! The good (and bad) thing about variants is that there's less software support available. It's good because it forces to solve the whole thing manually unless one has specialized or really generic tools (I don't). However, results are harder to check and doing all basics manually is kind of mundane (but good practice every now and then).

In this case it was impossible to use Hodoku even for coloring because it killed candidates automatically from boxes. Thus my options were to either solve on paper or on the text-based grid directly. I chose the latter, which was kind of tedious, especially for the basics at the end.