## Jigsaw windoku

For fans of Killer Sudoku, Samurai Sudoku and other variants

### Jigsaw windoku

Moonlotus windoku: anyone capable to solve?
Sudoku_Hyper_Speels_E_29-30-pages-2-1.png (67.39 KiB) Viewed 333 times
urhegyi

Posts: 539
Joined: 13 April 2020

### Re: Jigsaw windoku

Solution by SiseSuSo:
Code: Select all
`#1//B4,JSW.2.........8....2....3..5....1.4..............9....4..7...........8.....5...7....112222233111222333411123336441555366444555666447555966477789996777888999778888899`

SiSeSuSo_20210206_033014.png (21.69 KiB) Viewed 325 times
urhegyi

Posts: 539
Joined: 13 April 2020

### Re: Jigsaw windoku

The idea of moonlotus jigsaw combined with windoku inspired me to creating thisone:
Code: Select all
`#1//B4,JSW....8.......4.1......3.6.....1...8....5...2....7...4.....6.9......8.2.......7....112222233111222333411123336441555366444555666447555966477789996777888999778888899`

SiSeSuSo_20210206_162001.png (22.23 KiB) Viewed 309 times
urhegyi

Posts: 539
Joined: 13 April 2020

### Re: Jigsaw windoku

Or thisone: disable turbotfish in SiseSuSo when solving you get a solution with x-wing/yellyfish.
Code: Select all
`#1//B4,JSW...7.5.....8.9.3...........3..2.7..5.........2..5.3..8...........4.3.7.....1.4...112222233111222333411123336441555366444555666447555966477789996777888999778888899`

SiSeSuSo_20210206_170918.png (19.24 KiB) Viewed 300 times

Code: Select all
`Eliminated candidates per Method and per SudokuMethod   \  Sudoku |   SER |     1                   |-------|------Not counted elims  |     0 |    89Naked Singles      |   0.1 |    36Hidden Singles     |   0.2 |    99Naked Pair    [2]  |     3 |     4Naked Triple  [3]  |   3.6 |    11Naked Quad    [4]  |     5 |     3Locked Singles[2]  |   2.8 |    20Locked Singles[3]  |   2.9 |    15Locked Singles[4]  |     3 |    12X-Wing        [2]  |   3.2 |     2Jellyfish     [4]  |   5.2 |     5                   |-------|------Eliminated Cand's  |   296 |   296Sum(SER * Cand's)  | 257.9 | 257.9Initial Candidates :   297Maximum SER rating :   5.2 - Approach Labour rating      : 257.9 - Experimental ratingCheck (conflict)   : 297-297=0`

A few questions :
Code: Select all
`X-Wing, SwordFish, JellyFish  | ColsX-Wing (4)c89r37 => (-4)r3c7 (-4)r7c7 | Cols Jellyfish (8)c1468r3578 => (-8)r7c2 (-8)r8c2 (-8)r5c5 (-8)r3c7 (-8)r7c7 | Cols Singleton (8)c2r9 => (-8)r9c7`

It seems SiSeSuSo limits it's search to colums.
Perhaps there's a shorter alternative for yellyfish in rows?
BTW What is a Singleton?
When you did not disable Turbotfish SiSeSuSo does the elimination of the 7 candidates with Turbotfish instead of normal fish, because they are rated lower than 5.2 in the program.( between 4.4 and 4.6)
urhegyi

Posts: 539
Joined: 13 April 2020

### Re: Jigsaw windoku

urhegyi wrote:It seems SiSeSuSo limits it's search to colums.
- no, but maybe this is first to do...
Perhaps there's a shorter alternative for yellyfish in rows?
- can sure be
BTW What is a Singleton?
- during search of fish and elimination of candidates there is a single candidate found in row or column during fish-finding method:
jellyfish=4 digits locked in 4 rows/cols; swordfish in 3; X-wing in 2; singleton in 1 !
When you did not disable Turbotfish SiSeSuSo does the elimination of the 7 candidates with Turbotfish instead of normal fish, because they are rated lower than 5.2 in the program.( between 4.4 and 4.6)
- SiSeSuSo does not search per SER-rating, does a DFS (Depth First Search) and can stumble on a jellyfish (on digit p) before a x-wing (on digit q), where p<q.
- SiSeSuSo does a full sweep search per method in the methods list ordered from top to bottom.

You can also disable the jellyfish and Naked/hidden Quads in Help&Settings and solve the puzzle without the Yellyfish / Quads
Code: Select all
`Method   \  Sudoku |   SER |     1                   |-------|------Not counted elims  |     0 |    53Naked Singles      |   0.1 |    36Hidden Singles     |   0.2 |    37Naked Pair    [2]  |     3 |     2Naked Triple  [3]  |   3.6 |    14Locked Singles[2]  |   2.8 |    20Locked Singles[3]  |   2.9 |    15Locked Singles[4]  |     3 |    12X-Wing        [2]  |   3.2 |     6                   |-------|------Eliminated Cand's  |   195 |   195Sum(SER * Cand's)  | 222.1 | 222.1Initial Candidates :   197Maximum SER rating :   3.6 - Approach Labour rating      : 222.1 - Experimental ratingCheck (conflict)   : 197-197=0`

Hajime

Posts: 733
Joined: 20 April 2018
Location: Netherlands

### Re: Jigsaw windoku

I printed out the moonlotus windoku from the first post and solved it by hand. It looks easy, singles only, but doing it by hand was more challenging than I thought!
I hope to find more examples like this in the future.
I'm currently working on an algoritm to produce more moonlotus-windoku solution grids:
my first attempts:
Code: Select all
`#1/W/B4,JS...9.2.....6...9....2.5.1...............7...............5.1.4....8...3.....5.6...112222233111222333411123336441555366444555666447555966477789996777888999778888899`

solution:
Code: Select all
`184962735756183924932457186379825641861374592543691278295718463618249357427536819`

Code: Select all
`#1/W/B4,JS..........5..3..1.9..4.5..6...........1...2...........8..3.7..9.1..4..2..........112222233111222333411123336441555366444555666447555966477789996777888999778888899`

solution:
Code: Select all
`184256793756839412932415876579182634461593287328674951845327169617948325293761548`
urhegyi

Posts: 539
Joined: 13 April 2020

### Re: Jigsaw windoku

urhegyi wrote:I'm currently working on an algoritm to produce more moonlotus-windoku solution grids:
Any suggestions how to generate the solution grids?
urhegyi

Posts: 539
Joined: 13 April 2020

### Re: Jigsaw windoku

This windoku variant is not only possible with the moonlotus jigsaw layout, now i found another one:
Code: Select all
`#1/W/B4,JS......7.682..9..614.........3...........2.......8.3.....2...8.3.9.....5.......2.5711182229733182669731182269731182269731485269734485569734485569733485669744485559`

Solved this yesterday by hand which was a challenge:
hardest step pointing/claiming
Law of leftovers helps to do it quicker.
Also 2 hidden pairs and an UR can help.
Estimated rating 2.8 (or 3.4 with hidden pairs)
solution:
Hidden Text: Show
Code: Select all
`158249736825794361486132597634958172713425689561873924942567813297316458379681245`

update:
Thanks to the method described by Hajime I can now quick generate more solution grids like:
Code: Select all
`#1/W/B4,JS761492583146573892423981756918734625852367914394625178685149237537218469279856341711182229733182669731182269731182269731485269734485569734485569733485669744485559`

Example:
Code: Select all
`#1/W/B4,JS..1...5..1.......2.2.....5.9.8...6.5....6....3.4...1.8.8.....3.5.......9..9...3..711182229733182669731182269731182269731485269734485569734485569733485669744485559`

Rating:
Code: Select all
`Eliminated candidates per Method and per SudokuMethod   \  Sudoku |   SER |     1                   |-------|------Not counted elims  |     0 |    98Naked Singles      |   0.1 |    33Hidden Singles     |   0.2 |    94Locked Singles[2]  |   2.8 |    29Locked Singles[3]  |   2.9 |     3Locked Singles[4]  |     3 |     3XY-chain      [2]  |   3.6 |     1XY-chain      [4]  |   4.6 |     1                   |-------|------Eliminated Cand's  |   262 |   262Sum(SER * Cand's)  | 129.2 | 129.2Initial Candidates :   262Maximum SER rating :   4.6 - Approach Labour rating      : 129.2 - Experimental rating`

When I see the analyze log, I think it's possible to solve with pencil and paper!
hints:
Hidden Text: Show
Code: Select all
`XY_Chain r  |  (7=4)r4c4 (4=2)r1c4 (2=6)r6c4 (6=7)r6c8 [4] => (-7)r4c8XY_Chain J  |  (4=2)r4c8 (2=4)r8c7 [2] => (-4)r8c8`

Or even without Generalized Intersection but with additional Turbot Fish(es):
Code: Select all
`Eliminated candidates per Method and per SudokuMethod   \  Sudoku |   SER |     1                   |-------|------Not counted elims  |     0 |    89Naked Singles      |   0.1 |    20Hidden Singles     |   0.2 |   111Naked Pair    [2]  |     3 |     2Locked Singles[2]  |   2.8 |    29Turbot-fish   [4]  |   4.2 |     9XY-chain      [2]  |   3.6 |     1XY-chain      [4]  |   4.6 |     1                   |-------|------Eliminated Cand's  |   262 |   262Sum(SER * Cand's)  | 157.4 | 157.4Initial Candidates :   262Maximum SER rating :   4.6 - Approach Labour rating      : 157.4 - Experimental rating`

hints:
Hidden Text: Show
Code: Select all
`2-string-Kite,SkyScraper,TurbotCrane  | 2-string Kite (2) r4c6=r4c8-r8c7=r7c7 => (-2)r5c6 | X-Chain [4]  (2) r4c8=r8c7-r7c7=r9c8 => (-2)r1c8XY_Chain r  |  (7=4)r4c4 (4=2)r1c4 (2=6)r6c4 (6=7)r6c8 [4] => (-7)r4c8XY_Chain J  |  (4=2)r4c8 (2=4)r8c7 [2] => (-4)r8c8Naked/Hidden Pairs,Triplets,Quads  | NSS (24)J6e48 => (-4)r2c7 | NSS (24)c8r49 => (-4)r1c82-string-Kite,SkyScraper,TurbotCrane  | X-Chain [4]  (7) r1c8=r6c8-r7c7=r5c6 => (-7)r1c6 | X-Chain [4]  (7) r2c7=r6c8-r1c8=r5c6 => (-7)r2c6 (-7)r7c7 | 2-string Kite (7) r3c7=r2c7-r6c8=r6c5 => (-7)r3c5 | Skyscraper (7) r3c7=r2c7-r6c8=r1c8 => (-7)r4c6 | Skyscraper (7) r6c5=r6c8-r1c8=r1c1 => (-7)r7c1 | Skyscraper (7) r6c8=r6c5-r7c5=r7c9 => (-7)r6c5`
Last edited by urhegyi on Wed Feb 10, 2021 2:40 pm, edited 9 times in total.
urhegyi

Posts: 539
Joined: 13 April 2020

### Re: Jigsaw windoku

urhegyi wrote:
urhegyi wrote:I'm currently working on an algoritm to produce more moonlotus-windoku solution grids:
Any suggestions how to generate the solution grids?

In SiSeSuSo
• Press "Clear" to erase all givens
• In tab-page Speedy: check speedy checkbox
• Check method BruteForce
Label X:
• Press button "All Steps"
• Wait 2 seconds
• Select Tab Output or press Refresh there
• Select the solution lines and paste/append it in some file
• Press button "All back"
• Goto X
Due to a random function in Bruteforce you get different puzzles.

Here 3 puzzles in 5 seconds manually, no batch generate yet possible.
Hidden Text: Show
Code: Select all
`#1/W/B4,JS761492583952738461283167954594671238678243195417859326346925817825314679139586742112222233111222333411123336441555366444555666447555966477789996777888999778888899#1/W/B4,JS197523648546918327973846152352184976614237589721695834839752461285461793468379215112222233111222333411123336441555366444555666447555966477789996777888999778888899#1/W/B4,JS291548673584973126316725489153287964928654317462319758745196832679831245837462591112222233111222333411123336441555366444555666447555966477789996777888999778888899`

Hajime

Posts: 733
Joined: 20 April 2018
Location: Netherlands

### Re: Jigsaw windoku

I think this layout JSW (with windoku) can be interesting for the human solver!
Code: Select all
`#1/W/B4,JS4.7...1.5..........5.....7...8.4.5......2......3.1.9...7.....5..........8.5...7.9111122233111222333112223333445556666444555666444455566777788899777888999778889999`

It needs a lot of quads! How many are effectively needed, I hope to find out when trying to solve!
analyze:
Code: Select all
`Eliminated candidates per Method and per SudokuMethod   \  Sudoku |   SER |     1                   |-------|------Not counted elims  |     0 |    58Naked Singles      |   0.1 |    41Hidden Singles     |   0.2 |    85Naked Pair    [2]  |     3 |    18Naked Triple  [3]  |   3.6 |    10Naked Quad    [4]  |     5 |     8Hidden Quad   [5]  |   5.4 |     3Locked Singles[2]  |   2.8 |    37Locked Singles[3]  |   2.9 |    17Locked Singles[4]  |     3 |    13                   |-------|------Eliminated Cand's  |   290 |   290Sum(SER * Cand's)  | 359.2 | 359.2Initial Candidates :   290Maximum SER rating :   5.4 - Approach Labour rating      : 359.2 - Experimental rating`

When you disable quads(and wings max WXYZ) it solves with XYZ-wings which is lower in rating!
Code: Select all
`Eliminated candidates per Method and per SudokuMethod   \  Sudoku |   SER |     1                   |-------|------Not counted elims  |     0 |    55Naked Singles      |   0.1 |    43Hidden Singles     |   0.2 |     6Naked Single  [1]  |   2.5 |     1Naked Pair    [2]  |     3 |    27Naked Triple  [3]  |   3.6 |     6Locked Singles[2]  |   2.8 |    19Locked Singles[3]  |   2.9 |     7Locked Singles[4]  |     3 |     7XYZ Wing      [3]  |   4.4 |     3                   |-------|------Eliminated Cand's  |   174 |   174Sum(SER * Cand's)  | 218.3 | 218.3Initial Candidates :   174Maximum SER rating :   4.4 - Approach Labour rating      : 218.3 - Experimental rating`
urhegyi

Posts: 539
Joined: 13 April 2020