The New Sudoku Players' Forum

by **udosuk** » Fri Oct 27, 2006 6:49 am

Excellent job Uwe!

Here is the 777-code and complementary grids of the 623-candidate modification from Ruud:

Code: Select all: 777 777 777 777 777 777 777 422 022 775 775 775 521 775 775 775 777 775 164 777 777 777 777 777 777 777 777 164 777 140 777 777 777 340 777 777 777 777 777 442 777 777 240 777 777 775 775 775 777 775 775 775 777 775 220 777 777 777 264 777 210 777 777 777 777 777 777 414 777 777 412 777 777 420 777 777 777 777 777 777 777 777777777777777777777422022775775775521775775775777775164777777777777777777777777 164777140777777777340777777777777777442777777240777777775775775777775775775777775 220777777777264777210777777777777777777414777777412777777420777777777777777777777 . . . . . . . 234679 1234679 8 8 8 24678 8 8 8 . 8 12689 . . . . . . . . 12689 . 1256789 . . . 156789 . . . . . 235679 . . 1356789 . . 8 8 8 . 8 8 8 . 8 1346789 . . . 13689 . 1345789 . . . . . . 234589 . . 234579 . . 2346789 . . . . . . .

BTW Uwe only the xyz-wing is required to reduce it to singles... The shortest route ( :?:

) goes like this:

60 singles
1 naked/hidden triple: elimination in r1c3
1 xyz-wing: fixing r3c2
20 singles

I wonder if a simple short xy-chain could replace the 2 intermediate steps there... :?:

by **Ruud** » Fri Oct 27, 2006 12:26 pm

I slightly altered the approach by Pyrrhon, and found this grid:

628 candidates, 1 solution. It can be solved with basic techniques.

Code: Select all: 2346789 2346789 2346789 345789 345789 345679 456789 123456789 123456789 12346789 123678 1236789 13456789 13589 356789 1356789 123456789 12345789 12346789 123456789 12346789 123456789 135789 3456789 12456789 123456789 12345789 123456789 1235689 12356789 123456789 123456789 23456789 3456789 123456789 123456789 12356789 12356789 12356789 123456789 1235789 23456789 123456789 12356789 123456789 12356789 12356789 12356789 123456789 123456789 123456789 3456789 123456789 12345789 12345689 12345689 12345689 123456789 1234689 23469 123456789 1234589 123458 123456789 23456789 23456789 2346789 23789 234679 3456789 12345678 12345678 123456789 123456789 123456789 12346789 123789 1234679 13456789 123456789 1234578 342875916879136542651294837487962351916358274523741689264513798195687423738429165

only 101 candidates are disabled in this puzzle...

[edit]

The following puzzle also breaks a record. It has a mere 625 candidates, but each cell has 6 or more candidates.

Code: Select all: 12346789 13456789 123456789 1236789 1346789 1234678 123456789 12345679 2346789 123456789 13456789 123456789 123456789 1346789 1234678 123456789 123456789 123456789 123456789 3456789 123456789 123456789 3456789 2345678 23456789 345679 23456789 123456789 123456789 12345678 1235678 12345678 12345678 123456789 1234567 12345678 23456789 3456789 123456789 13456789 3456789 2345678 123456789 345679 23456789 123456789 3456789 12345679 123456789 1346789 123467 12345679 134679 124679 123456789 345789 12345789 1234789 134789 123456789 134579 134579 124789 234689 3456789 2345689 123689 123456789 12345678 123456789 234679 12346789 23456789 3456789 123456789 12356789 13456789 12345678 123456789 345679 123456789

No need to post the solution. It can be solved with simple techniques.

And then this remarkable puzzle with 628 candidates. Interesting to see that all candidates for digit 6 are present, and only a single candidate for digit 9 is missing.

Code: Select all: 13456789 1356789 13456789 1356789 156789 156789 145689 123456789 13456789 123456789 12456789 123456789 356789 123456789 256789 1245689 2456789 2456789 123456789 12346789 12346789 136789 12356789 126789 1234689 123456789 123456789 123456789 123456789 123456789 1356789 123456789 1256789 145679 1235679 123456789 2345679 12345679 123456789 123456789 123456789 123456789 123456789 12346789 12345679 2345679 1235679 12345679 123456789 12345679 123456789 124569 12345679 12345679 123456789 12356789 123456789 123456789 12456789 1246789 123456789 25678 2456789 23456789 123456789 123456789 1346789 12346789 12346789 2456789 23456789 23456789 2456789 256789 123456789 1456789 1246789 12456789 245689 2456789 2456789

Also a perfect puzzle for human solvers.

:!:

Another record puzzle: 633 candidates. Only 96 candidates eliminated.

Code: Select all: 123456789 2356789 123456789 1235679 12356789 123456789 12456789 1245789 123456789 123456789 123456789 12456789 135679 123456789 123456789 12456789 15789 12456789 134567 23456789 12456789 1345679 123456789 123456789 12456789 145789 12456789 12456789 123456789 2456789 135679 12356789 12356789 1245679 13579 1235679 3456789 3456789 456789 12345679 12356789 123456789 123456789 123456789 12345679 2456789 256789 2456789 123456789 1256789 123456789 12456789 12345789 123456789 12346789 236789 246789 12345679 1256789 2346789 12456789 12456789 123456789 12346789 123456789 12456789 1345679 123456789 123456789 1456789 123456789 12345689 123456789 2356789 123456789 1345679 123456789 123456789 123456789 134589 12345689

I wonder how far we can go...

by **udosuk** » Mon Oct 30, 2006 6:06 am

I envy those using a big laptop/monitor with 1280x1024 or wider displays... I couldn't see the full PM grids properly on my tiny laptop...

Here are presentations which could be shown properly on my screen... :idea:

Code: Select all: 628 candidates (1st grid): 357 357 357 167 167 175 077 777 777 757 716 717 577 523 137 537 777 767 757 777 757 777 527 177 677 777 767 777 733 737 777 777 377 177 777 777 737 737 737 777 727 377 777 737 777 737 737 737 777 777 777 177 777 767 773 773 773 777 753 351 777 763 762 777 377 377 357 307 355 177 776 776 777 777 777 757 707 755 577 777 766 357357357167167175077777777757716717577523137537777767757777757777527177677777767 777733737777777377177777777737737737777727377777737777737737737777777777177777767 773773773777753351777763762777377377357307355177776776777777777757707755577777766 15 15 15 126 126 128 123 . . 5 459 45 2 2467 124 24 . 6 5 . 5 . 246 12 3 . 6 . 47 4 . . 1 12 . . 4 4 4 . 46 1 . 4 . 4 4 4 . . . 12 . 6 7 7 7 . 57 1578 . 67 679 . 1 1 15 1456 158 12 9 9 . . . 5 456 58 2 . 69 625 candidates, 6 or more candidates in each cell: 757 577 777 717 557 756 777 775 357 777 577 777 777 557 756 777 777 777 777 177 777 777 177 376 377 175 377 777 777 776 736 776 776 777 774 776 377 177 777 577 177 376 777 175 377 777 177 775 777 557 754 775 555 655 777 167 767 747 547 777 565 565 647 353 177 373 713 777 776 777 355 757 377 177 777 737 577 776 777 175 777 757577777717557756777775357777577777777557756777777777777177777777177376377175377 777777776736776776777774776377177777577177376777175377777177775777557754775555655 777167767747547777565565647353177373713777776777355757377177777737577776777175777 5 2 . 45 25 59 . 8 15 . 2 . . 25 59 . . . . 12 . . 12 19 1 128 1 . . 9 49 9 9 . 89 9 1 12 . 2 12 19 . 128 1 . 12 8 . 25 589 8 258 358 . 126 6 56 256 . 268 268 356 157 12 17 457 . 9 . 158 5 1 12 . 4 2 9 . 128 . 628 candidates, candidate 6 in all cells, only 1 cell missing candidate 9: 577 537 577 537 437 437 473 777 577 777 677 777 137 777 237 673 277 277 777 757 757 517 737 617 753 777 777 777 777 777 537 777 637 475 735 777 375 775 777 777 777 777 777 757 775 375 735 775 777 775 777 671 775 775 777 737 777 777 677 657 777 236 277 377 777 777 557 757 757 277 377 377 277 237 777 477 657 677 273 277 277 577537577537437437473777577777677777137777237673277277777757757517737617753777777 777777777537777637475735777375775777777777777777757775375735775777775777671775775 777737777777677657777236277377777777557757757277377377277237777477657677273277277 2 24 2 24 234 234 237 . 2 . 3 . 124 . 134 37 13 13 . 5 5 245 4 345 57 . . . . . 24 . 34 238 48 . 18 8 . . . . . 5 8 18 48 8 . 8 . 378 8 8 . 4 . . 3 35 . 1349 13 1 . . 25 5 5 13 1 1 13 134 . 23 35 3 137 13 13 633 candidates (only 96 initially eliminated candidates!): 777 337 777 735 737 777 677 667 777 777 777 677 535 777 777 677 427 677 574 377 677 575 777 777 677 467 677 677 777 277 535 737 737 675 525 735 177 177 077 775 737 777 777 777 775 277 237 277 777 637 777 677 767 777 757 317 257 775 637 357 677 677 777 757 777 677 575 777 777 477 777 773 777 337 777 575 777 777 777 563 773 777337777735737777677667777777777677535777777677427677574377677575777777677467677 677777277535737737675525735177177077775737777777777775277237277777637777677767777 757317257775637357677677777757777677575777777477777773777337777575777777777563773 . 14 . 48 4 . 3 36 . . . 3 248 . . 3 2346 3 289 1 3 28 . . 3 236 3 3 . 13 248 4 4 38 2468 48 12 12 123 8 4 . . . 8 13 134 13 . 34 . 3 6 . 5 145 135 8 34 15 3 3 . 5 . 3 28 . . 23 . 7 . 14 . 28 . . . 267 7

by **tarek** » Mon Oct 30, 2006 10:44 am

udosuk wrote:I envy those using a big laptop/monitor with 1280x1024 or wider displays... I couldn't see the full PM grids properly on my tiny laptop...

What if I told you that I have a 1920*1200 wide screen which can 90 degrees rotate ??!!!

tarek

by **udosuk** » Mon Oct 30, 2006 11:25 am

tarek wrote:What if I told you that I have a 1920*1200 wide screen which can 90 degrees rotate ??!!!

Go ahead... Show off your b@tt... :twisted:

by **tarek** » Tue Oct 31, 2006 9:28 am

udosuk wrote:Go ahead... Show off your b@tt...

High definition is not for everyone

..........

Spirit of Oct 31

by **Pat** » Tue Nov 28, 2006 4:34 pm

history:

Bob Hanson (2005.Nov.16) wrote:no-given Sudoku -- twisted idea

A colleague of mine just pointed out that the real information unit in solving SuDoku is
"X is not possible in cell A."
Thus, it would be conceivable to construct a SuDoku puzzle with no givens at all -
one just has to provide enough "elimination clues" in the form
r3c2 is not 6
r4c5 is not 8
Thus my questions:
Has anyone ever presented Sudoku puzzles that have these sorts of "elimination clues" instead of givens?
Has anyone ever presented Sudoku puzzles that have a mixture of, say 15 givens and 5 "elimination clues"?
What is the fewest number of "elimination clues" that are necessary to solve an NxN SuDoku?
(Note that each "given" constitutes up to 28 "elimination clues".)
I would think it easy to construct an "unsolvable" 16-given 9x9 SuDoku,
then provide -- how many -- perhaps one? -- "elimination clue"
that would provide just the information necessary to, say, produce a hidden n-tuple and complete the puzzle.
That would be considerably less information than a 17-given SuDoku.
And maybe one or two choice clues with a 15-given SuDoku? ... Where does this end?

If you want to experiment with this, by the way,
there is a hidden feature of Sudoku Assistant
that if you click on the background of a cell and enter a NEGATIVE number, it removes a mark.
I'm not sure any other program has this feature or how useful it is. Maybe. I suppose they all do.

Sudoku Assistant
http://www.stolaf.edu/people/hansonr/sudoku

~ Pat

by **999_Springs** » Fri Mar 23, 2007 7:36 pm

Why can't you mix the order of the candidates to make it seem harder to the person solving the puzzle?

Or you could put in duplicate candidates to make it even harder...

so, say, 1235689 becomes 155692839.

25th post!

edit: wow when i was 13 some of the posts i made were really dumb

by **tarek** » Sun Mar 16, 2014 12:33 am

2 nasty puzzles

Code: Select all: +------------------------------+------------------------------+------------------------------+ | 12459 259 2345679 | 2356789 12345679 1245789 | 1245679 257 2357 | | 459 12346789 123589 | 234569 12345 124567 | 123578 12345679 567 | | 2345789 145679 168 | 1245789 258 2356789 | 348 345679 1256789 | +------------------------------+------------------------------+------------------------------+ | 3456789 245789 1234569 | 23456789 579 12456789 | 123567 256789 1456789 | | 12345789 124578 456 | 359 1379 157 | 456 235689 12356789 | | 1234569 123458 3456789 | 12345689 135 12345678 | 1456789 123568 1234567 | +------------------------------+------------------------------+------------------------------+ | 134589 134567 267 | 1234578 258 1235689 | 249 134569 1235678 | | 34 123456789 235789 | 345689 456789 145678 | 125789 123456789 156 | | 34578 358 14569 | 235689 13456789 234578 | 1345678 158 15689 | +------------------------------+------------------------------+------------------------------+ +------------------------------+------------------------------+------------------------------+ | 136789 1259 123569 | 1235689 1235789 1234578 | 123457 2357 134789 | | 1459 123456789 123589 | 25679 1235 24579 | 123578 123456789 3567 | | 145789 145679 1234579 | 2489 12346789 2678 | 235679 345679 356789 | +------------------------------+------------------------------+------------------------------+ | 1456789 34589 2469 | 23456789 13579 12456789 | 2467 15678 3456789 | | 1345679 1457 12346789 | 13579 12346789 13579 | 123478 569 1345679 | | 1234567 23459 3468 | 12345689 13579 12345678 | 1468 12567 1234569 | +------------------------------+------------------------------+------------------------------+ | 123457 134567 1345789 | 2348 12346789 1268 | 1356789 134569 123569 | | 3457 13456789 235789 | 13568 5789 13458 | 125789 123456789 1569 | | 123679 3578 356789 | 2356789 1235789 1245789 | 145789 1589 123479 | +------------------------------+------------------------------+------------------------------+

I think that I have managed to re-confirm using gsf's Sudoku that these are 2 of many Pencil-mark sudokus with singles backdoor size 4. Size 5 doesn't look far to reach :idea:

by **dobrichev** » Mon Mar 17, 2014 6:48 am

Hi tarek,

Have these 2 something in common with the backdoor size 4 ordinary sudoku family?
I can supply you with BD size 3 ordinary puzzles if you think they can be used as seed in your search.

Waiting for news...

by **tarek** » Mon Mar 17, 2014 10:48 pm

dobrichev wrote:Have these 2 something in common with the backdoor size 4 ordinary sudoku family?
I can supply you with BD size 3 ordinary puzzles if you think they can be used as seed

Thanks dobrichev but I'm not adopting a clear strategy where that would be useful.
The regular (vanilla) sudoku is a special case of a pencilmark sudoku. You can add candidates to the vanilla sudoku until there is multiple solutions. You can evaluate that minimal puzzle then as it may have become more difficult (I haven't adopted this strategy but it sounds promising)

I have a few singles backdoor size 5 but I haven't prepared them yet. I have a few symmetric puzzles as well

by **Smythe Dakota** » Tue Mar 18, 2014 11:51 pm

Ruud wrote: .... Is this the way to do it, udosuk?

I assume this is the old alt-255 trick, or something similar.

Bill Smythe

by **tarek** » Wed Mar 19, 2014 9:42 am

gsf wrote:here are, in row order, the singles <backdoor-size,number-backdoors>
for the 57 pm-only grids from a few posts back
it took 1h55m @ 3.2GHz and some reworked backdoor logic
there are some 5's but no 6's

Code: Select all
4 11 4 1 4 2 4 3 5 33 4 4 4 2 5 47 4 15 4 3 4 2 4 1 5 31 5 48 4 6 4 4 4 1 4 1 4 3 5 48 4 4 4 1 5 35 4 2 4 1 4 2 5 31 5 27 4 1 4 2 4 1 4 4 4 4 5 41 4 1 4 4 4 2 4 3 4 7 4 1 4 5 5 34 5 44 4 4 4 4 4 1 4 8 5 36 4 2 4 1 4 2 5 47 4 4 4 3 5 20 5 35 5 43

The link to these puzzles is broken. But this post suggest that the Singles backdoor size 5 barrier has already been reached ....

by **tarek** » Wed Mar 19, 2014 2:02 pm

Only one sweep produced this minimal singles backdoor size 5. There is only 1 set possible. The Size 6 therefore looks very very possible

Code: Select all: .-----------------------------------------------------------------------------------------------. | 136789 1234789 12345789 | 12356789 25789 1245789 | 12356789 136789 134789 | | 123679 124 1234578 | 12345679 12358 123457 | 235689 236 123479 | | 12345679 123456 12345678 | 123456 123458 1234567 | 12345689 123456 12345679 | :-----------------------------------------------------------------------------------------------: | 13456789 134589 12345789 | 123456789 123456789 12456789 | 2356789 356789 156789 | | 34569 1345679 123456789 | 12345689 24568 12345678 | 2345689 34569 145679 | | 1234569 123457 1234578 | 12345689 12345689 12345678 | 125689 1269 123567 | :-----------------------------------------------------------------------------------------------: | 13456789 456789 12456789 | 13456789 12356789 456789 | 23456789 456789 13456789 | | 136789 478 124578 | 1356789 1235789 1345789 | 2356789 689 134789 | | 123679 123479 12345789 | 12356789 12358 1234578 | 12356789 123679 123479 | .-----------------------------------------------------------------------------------------------.

by **dobrichev** » Wed Mar 19, 2014 8:36 pm

To me counting the sets is waste of time (unless it is the first puzzle of a kind).

BTW, do you stuck to some intermediate format of pencilmarks representation? Text file, one puzzle per line. Champagne's skfr can be modified to accept PM only puzzles...

The New Sudoku Players' Forum

Pencilmark only Sudoku

no-given SuDoku -- pencilmark-only SuDoku

Re: Pencilmark only Sudoku

Re: Pencilmark only Sudoku

Re: Pencilmark only Sudoku

Re:

Re:

Re: Pencilmark only Sudoku

Re: Pencilmark only Sudoku