The New Sudoku Players' Forum

[mantouq]

Code: Select all

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



Code: Select all

+-------------+----------------+--------------+
| 239 5  1269 | 12346 8  1234  | 3469 7   346 |
| 39  4  1689 | 13567 16 1357  | 2    359 568 |
| 7   68 268  | 23456 9  2345  | 3458 35  1   |
+-------------+----------------+--------------+
| 1   67 5    | 39    4  39    | 678  2   68  |
| 24  9  24   | 8     7  6     | 135  135 35  |
| 8   3  67   | 15    2  15    | 67   4   9   |
+-------------+----------------+--------------+
| 6   18 489  | 1249  3  12489 | 1459 159 7   |
| 5   17 3    | 14579 16 1479  | 149  8   2   |
| 49  2  4789 | 1479  5  14789 | 1349 6   34  |
+-------------+----------------+--------------+

according to the susser solver, i need to use forcing chains for the next step. is there an easier way of recognizing a forcing chain or are there other techniques which can solve it??

[vidarino]

Well, first of all, there is only one place for a 6 in row 8.

A bit later, you might need the hint from this thread too.

Vidar

[mantouq]

sorry there's an amendment to r8c4. the correct one is here:

Code: Select all

+-------------+----------------+--------------+
| 239 5  1269 | 12346 8  1234  | 3469 7   346 |
| 39  4  1689 | 13567 16 1357  | 2    359 568 |
| 7   68 268  | 23456 9  2345  | 3458 35  1   |
+-------------+----------------+--------------+
| 1   67 5    | 39    4  39    | 678  2   68  |
| 24  9  24   | 8     7  6     | 135  135 35  |
| 8   3  67   | 15    2  15    | 67   4   9   |
+-------------+----------------+--------------+
| 6   18 489  | 1249  3  12489 | 1459 159 7   |
| 5   17 3    | 14679 16 1479  | 149  8   2   |
| 49  2  4789 | 1479  5  14789 | 1349 6   34  |
+-------------+----------------+--------------+

[vidarino]

sorry there's an amendment to r8c4. the correct one is here: ...

Ah. In that case I'll simply refer you to the thread I linked to above.

Vidar

[tarek]

Attempting another approach,

My solver couldn't be solved without chains,

A total of chains & 2 Almost locked cells eliminations were used.

Tarek

[Carcul]

Alternatively, this puzzle can be solved in one step with a Single Implication Network. See the thread "Can't force chains on this one??" for further details.

Carcul

[Havard]

My solver cracked it with a few ALS-eliminations! I love that stuff!

[absolute beginner]

Hello Havard,

please show me the first ALS!
I am trying to implement ALS in my solver,
but have no strategy, to find them

[bennys]

look at http://forum.enjoysudoku.com/viewtopic.php?t=2510

[Havard]

Hello Havard,

please show me the first ALS!
I am trying to implement ALS in my solver,
but have no strategy, to find them

You were just given a link from the very person that thought of this stuff!

Bennys:
I think ALS is one of the greatest discoveries in Sudoku solving! Finally a alternative to forcing chains /loops that cracks the really hard puzzles. Kudos to you!

[absolute beginner]

Benny,
thank You very much.
I thought, xyz-wings were difficult to implement ;-),
but ALS will really be challenging.

I read the Eppstein-paper,
( and, yes, i did know, what bipartite graphs are )
but I think, thats nice for implementing a solver,
but not very good for paper&pencil.
And if you are really interested in a solution,
use a backtracking-alg and the work is done.
Thats not the chanllenge

[bennys]

Thanks