The New Sudoku Players' Forum

[urhegyi]

I had a lucky day 2 difficult ones with the same pattern. Why does this pattern allow creating nice puzzles in a limited batch and others produce nothing special?
number 1:

Code: Select all

...7..3.1..5.......3.21..........5......9..27.8...4....6..2.8...7..........8..... 122223333112233433142224453144444555166665555167765888117766888777776888999999999

number 2:

Code: Select all

4.......5...18.........1......4...1......2..8......35.164...8....75.......3..4..2 122223333112233433142224453144444555166665555167765888117766888777776888999999999

Rate them and convince yourself that it's possible. I couldn't believe my eyes.
If anyone is capable of creating a morph of the first one. Or even better find another example. I like yellyfishes.
I tried to make a morph of the first example myself:
number 3:

Code: Select all

...1..6.7..3.......6.57..........3......4..51.8...2....9..5.8...1..........8..... 122223333112233433142224453144444555166665555167765888117766888777776888999999999

or another:
number4:

Code: Select all

...7..9.1..6.......9.31..........6......2..37.8...5....4..3.8...7..........8..... 122223333112233433142224453144444555166665555167765888117766888777776888999999999

[Hajime]

I had a lucky day 2 difficult ones with the same pattern. Why does this pattern allow creating nice puzzles in a limited batch and others produce nothing special?
number 1:
...7..3.1..5.......3.21..........5......9..27.8...4....6..2.8...7..........8..... 122223333112233433142224453144444555166665555167765888117766888777776888999999999
number 2:
4.......5...18.........1......4...1......2..8......35.164...8....75.......3..4..2 122223333112233433142224453144444555166665555167765888117766888777776888999999999
Rate them and convince yourself that it's possible. I couldn't believe my eyes.
[/code]

Are the red digits in the same cell ?? Sensei-1 ??

[urhegyi]

I had a lucky day 2 difficult ones with the same pattern. Why does this pattern allow creating nice puzzles in a limited batch and others produce nothing special?
number 1:
...7..3.1..5.......3.21..........5......9..27.8...4....6..2.8...7..........8..... 122223333112233433142224453144444555166665555167765888117766888777776888999999999
number 2:
4.......5...18.........1......4...1......2..8......35.164...8....75.......3..4..2 122223333112233433142224453144444555166665555167765888117766888777776888999999999
Rate them and convince yourself that it's possible. I couldn't believe my eyes.
[/code]

Are the red digits in the same cell ?? Sensei-1 ??

It's not an overlapping jigsaw. Just 2 different examples of an ordinary jigsaw, but with a difficult level.

[Hajime]

It's not an overlapping jigsaw. Just 2 different examples of an ordinary jigsaw, but with a difficult level.

Aha, #1 has a jellyfish (SER 5.2) and #2 a turbotfish ((7) r1c5=r3c4-r3c8=r6c6 => (-7)r1c6, with SER 4.3) as most difficult technique.

[1to9only]

Using same jigsaw layout as OP.

Code: Select all

.26...83........17........2...487........2........5.....9..4.....72...8...4...... 122223333112233433142224453144444555166665555167765888117766888777776888999999999

Generalized X-Wing: R3C3,R3C6,R2C7,R2C4: -3r4c2


Code: Select all

.....2...9..................4.7....83.2.9............6..1...4......6..3..8....1.. 122223333112233433142224453144444555166665555167765888117766888777776888999999999

Generalized X-Wing: R1C7,R3C9,R3C4,R1C2: -9r3c6, -9r3c8


Note: I'm running a later version of JigsawExplainer - I've reworked some old changes and this has got rid of the UR bug reported elsewhere, also means JE now solves/rates slightly differently from previous versions, usually lower ratings. For Generalized X-Wings to show up, a newer version of JigsawExplainer is required. I'll try to do a new pre-release later today or tomorrow.

[urhegyi]

I have another example with the UR bug. It can only be analyzed and solved correctly when you disable UR in the old version. I'm interested to hear if this is fixed in the new version too.

Code: Select all

.......2...1......9..5...16.....6.5..54...76..2.1.....34...1..7......5...7....... 111111222333331112344555552334455222344455266777448886797888666797788866799999996

[Hajime]

I have another example with the UR bug. It can only be analyzed and solved correctly when you disable UR in the old version. I'm interested to hear if this is fixed in the new version too.

Code: Select all

.......2...1......9..5...16.....6.5..54...76..2.1.....34...1..7......5...7....... 111111222333331112344555552334455222344455266777448886797888666797788866799999996

Two skyscrapers needed.
Skyscraper (2) r2c6=r2c7-r3c7=r3c6 => (-2)r9c6
Skyscraper (8) r2c7=r3c7-r3c2=r4c2 => (-8)r2c4 (-8)r2c5

New puzzle, same layout but also with diagonals , only 12 clues
X-chain of length 6 needed (SER 6.5-6.8)

Code: Select all

#1//B4,XJS
9.............92......................3....7..5.....6.8..5.....6......4.2........
122223333112233433142224453144444555166665555167765888117766888777776888999999999


[urhegyi]

I have another example with the UR bug. It can only be analyzed and solved correctly when you disable UR in the old version. I'm interested to hear if this is fixed in the new version too.

Code: Select all

.......2...1......9..5...16.....6.5..54...76..2.1.....34...1..7......5...7....... 111111222333331112344555552334455222344455266777448886797888666797788866799999996

Two skyscrapers needed.
Skyscraper (2) r2c6=r2c7-r3c7=r3c6 => (-2)r9c6
Skyscraper (8) r2c7=r3c7-r3c2=r4c2 => (-8)r2c4 (-8)r2c5
With skyscrapers it works, but in the old version of JigsawExplainer it had some problems too!
New puzzle, same layout but also with diagonals , only 12 clues
X-chain of length 6 needed (SER 6.5-6.8)

Code: Select all

#1//B4,XJS
9.............92......................3....7..5.....6.8..5.....6......4.2........
122223333112233433142224453144444555166665555167765888117766888777776888999999999


[1to9only]

I have another example with the UR bug. It can only be analyzed and solved correctly when you disable UR in the old version. I'm interested to hear if this is fixed in the new version too.

Code: Select all

.......2...1......9..5...16.....6.5..54...76..2.1.....34...1..7......5...7....... 111111222333331112344555552334455222344455266777448886797888666797788866799999996

Try the new version.See here.
The 2nd jigsaw I posted earlier today requires 3 X-Wings and 3 Generalized X-Wings to solve.

[urhegyi]

I have another example with the UR bug. It can only be analyzed and solved correctly when you disable UR in the old version. I'm interested to hear if this is fixed in the new version too.

Code: Select all

.......2...1......9..5...16.....6.5..54...76..2.1.....34...1..7......5...7....... 111111222333331112344555552334455222344455266777448886797888666797788866799999996

Try the new version.See here.
The 2nd jigsaw I posted earlier today requires 3 X-Wings and 3 Generalized X-Wings to solve.

UR bug is fixed!!! Works fast and accurate now.
How to generate examples like the one below?(not harder because generating in diabolic mode generates all 9.8 to 10.0 puzzles with almost max rating)
What's the exact difference between generate after loading a specific grid layout but an empty grid and use solution(after generate solution)? Generated jigsaws are both random.

Code: Select all

8.......1.7.....4.2.....5.7..6...7...62..598...7...4..9.8.6...2.8.......7.......6 111123333111224333122244443125544443225555566788885569788886669777866999777769999

[1to9only]

How to generate examples like the one below?(not harder because generating in diabolic mode generates all 9.8 to 10.0 puzzles with almost max rating)

You can generate Fiendish puzzles (ED=1.0-6.2 only), but most of the puzzles will be rejected (ED=20.0, i.e. > 6.2), and the rest will be a mixed bag. I usually generate Diabolical puzzles, put them in a file (one jigsaw per line), and run diuf.sudoku.test.Tester on them - this analyzes the puzzles (same as Analyze in the GUI), and list the techniques needed to solve each one, I then pick and choose from these.

What's the exact difference between generate after loading a specific grid layout but an empty grid and use solution(after generate solution)? Generated jigsaws are both random.

Yes, all generated jigsaws are random.

'Generate...' will, on each click, generate a random solution grid, and then create a jigsaw from that solution grid - all generated jigsaws have different solutions.

'Use solution' saves the solved grid, and the next Generate dialog uses this as solution grid - all generated jigsaws have the same solution. Saved solution grid is discarded after exiting the Generate dialog.

''Generate Solution' is used to generate a solution grid from a partially filled grid, e.g. if you already have a solution grid for a samurai grid 3, to generate a solution for the samurai grid 1, you would fill grid 1 block 9 grom grid 3 block 1, and select Generate Solution.

[urhegyi]

A very hard to generate pattern(number 18 in the list of JigsawExplainer). Finally after 15 minutes it came with this:(solvable with bug type 1 or turbotcrane and skyscraper):

Code: Select all

...3......5..........68.4.......2.5....8..29...8.6.1.......5...6..4.............. 111223333112223433112225433611525444666555444666575884996577788996977788999977888

longtime.png (29.44 KiB) Viewed 853 times

A possible solvepath:

sp642648874.pdf

(45.97 KiB) Downloaded 147 times

edit: What has happened in your latest version of JigsawExplainer?
To generate 3 more examples it took 35, 40 end even 45 minutes each starting with loading the format definition and an empty grid. With a backup of the version before only seconds.

[urhegyi]

As I posted before it takes now 2 hours to generate 3 examples of the latest H pattern, nr 18 in the list. As with the version before only seconds to minutes. And generating jigsaws with the use solution possibility can't produce jigsaws with the same solution. Solutions seems to be different all the time like with a normal generate. It seems the jason file isn't updated anymore. It uses now the registry? When that's the case, it can be the problem of all things mentioned before.

[1to9only]

I dont have issues here, delete the json file and a new one will be created on the next restart. The json file does not get updated if no setting has changed. The registry is not used as the code has been removed, unless youre using a very early version. Otherwise 'There is NO support for this program!'.

[urhegyi]

I dont have issues here, delete the json file and a new one will be created on the next restart. The json file does not get updated if no setting has changed. The registry is not used as the code has been removed, unless youre using a very early version. Otherwise 'There is NO support for this program!'.

I deleted the JSON file and on restart it isn't re-created in the same directory as the program like it was before and use the latest version!
edit: I moved the program to another directory and on restart here the JSON is created.