The New Sudoku Players' Forum

[urhegyi]

A lot of my previous puzzles where to hard for the p&p solver.
Here one everyone should try: only pointing/claiming and subsets up to quads required.

Code: Select all

#1/WX/B4,JS
7.......3....5.......................9.....6.......................8....2.......4
112333333112223344111125344612225544662555844665588894667589999667788899777777899

[creint]

Does not require subsets.

Spoilers:

Hidden Text: Show

You can even use the symmetry, it even solves with symmetry. So if you find one you know where you need to place the other.
Mapping 2-3, 4-7, 5-8, 6-9
After the first naked single you need some locked singles before you can make progress
After locked singles you could find 4x 4 and 7
After locked singles you could find 6x 2 and 3
After locked singles you could find 2x 4 and 7
After locked singles you could find 1x 5 and 8

[Hajime]

Does not require subsets.

Spoilers:

Hidden Text: Show

You can even use the symmetry, it even solves with symmetry. So if you find one you know where you need to place the other.
Mapping 2-3, 4-7, 5-8, 6-9
After the first naked single you need some locked singles before you can make progress
After locked singles you could find 4x 4 and 7
After locked singles you could find 6x 2 and 3
After locked singles you could find 2x 4 and 7
After locked singles you could find 1x 5 and 8

Get used to it. I learn a lot of him. Creint has the best solver, unfortunately it is private to him.

[creint]

I don't know what you mean by 'the best'.
There are more solvers out there, some have more tactics.
There are still things that can be improved to be more user-friendly.
You can test it if you want, how can I send it to you?

[urhegyi]

Inspired by the symmetrical solvepath proposed by Creint a somewhat harder variant which I hope symmetry will reduce the difficulty here too.

Code: Select all

#1/WX/B4,JS
............6.4...3..9.8..4.........5.......2.........4..1.5..8...5.3............
111122233144422233144422233144455539177555339175556669778886669778886669778889999

[Hajime]

Code: Select all

#1/WX/B4,JS
7.......3....5.......................9.....6.......................8....2.......4
112333333112223344111125344612225544662555844665588894667589999667788899777777899

It is also solvable using lots of Painting/Claiming (only 2 houses involved) and 2 XY-Wings
XY-wing : (2=3)r5c3 (3=2)r5c7 => (-2)r5c4 (-2)r5c6
XY-wing : (3=2)r5c3 (2=3)r5c7 => (-3)r5c4 (-3)r5c6
No Subsets and no Locked-Singles/Generalized-Intersection.

[Hajime]

I don't know what you mean by 'the best'.
There are more solvers out there, some have more tactics.
There are still things that can be improved to be more user-friendly.
You can test it if you want, how can I send it to you?

Yes, 'the best' is of course my personal opinion.

• But there are not much solvers that can solve overlapping puzzles with multiple grids.
• And lots of Single Grid solvers can not handle jigsaws and/or combined with diagonals,windoku etc.
• And your solver may not have 'all the tactics' but it spots eliminations that I have not seen in other solvers (including SiSeSuSo):
• in http://forum.enjoysudoku.com/halo-around-the-sun-t38218.html?hilit=halo your solver is not limited to grid-wise methods
• in http://forum.enjoysudoku.com/jigsaw-windoku-x-t38749.html#p301129 you make use of generic X-wings; never seen that before.

Features to endorse my opinion

[urhegyi]

Can symmetry help here too to solve this JSWX in an easier way?

Code: Select all

#1/WX/B4,JS
..1...5...........3..9.8..7...........................6..8.1..4...........2...9..
336666777366666777344444777344411222341111152333115552888555552888999992888999922