The New Sudoku Players' Forum

[Morgoth]

Unique rectangle/loop type 3 with hidden quad?

I found with naked quad and I suppose it is possible to have also hidden quad if none of the cells in a group is solved.

[Morgoth]

Here it is

Code: Select all

000000000000200600314000000008094013060005740000003002000000800501000000400800029

This is extremely rare pattern.
It can be easily found if a pattern with empty group is used.

[udosuk]

Excellent puzzle!

After SSTS gets stuck:

 *-----------------------------------------------------------*
 |-2678 -2579 -2679  |*34   *147  #1789  |*13   #589  *145   |
 | 78    579   79    | 2     1347  1789  | 6     3589  145   |
 | 3     1     4     | 5     6    #89    | 2    #89    7     |
 |-------------------+-------------------+-------------------|
 | 27    27    8     | 6     9     4     | 5     1     3     |
 | 9     6     3     | 1     2     5     | 7     4     8     |
 | 1     4     5     | 7     8     3     | 9     6     2     |
 |-------------------+-------------------+-------------------|
 | 267   2379  2679  | 34    147   167   | 8     57    15    |
 | 5     8     1     | 9     37    2     | 4     37    6     |
 | 4     37    67    | 8     5     167   | 13    2     9     |
 *-----------------------------------------------------------*

UR {89} in r13c68 => r1c68 must have one of {157}
=> Hidden quad {2689} in r1c123+r1c(6|8)
(Or alternatively, naked quint {13457} in r1c4579+r1c(6|8))
=> 5 & 7 eliminated from r1c123

After that the puzzle is solved under SSTS.

This is extremely rare pattern.
It can be easily found if a pattern with empty group is used.

If naked quad is given a higher priority than hidden quad then you can only get the latter when there is an empty row/column/block and a "naked quint" in 5 of the cells...

[Morgoth]

Excellent puzzle!

This is extremely rare pattern.
It can be easily found if a pattern with empty group is used.

If naked quad is given a higher priority than hidden quad then you can only get the latter when there is an empty row/column/block and a "naked quint" in 5 of the cells...

Yes, I know. My idea was that using a pattern with empty group will give you high possibility to have empty group in the end
The puzzles I was looking for this UR were without pattern, but made from the first version of my generator which was usually left the first row without hints during the minimization