The hardest sudokus

Everything about Sudoku that doesn't fit in one of the other sections

Postby daj95376 » Sun Oct 29, 2006 8:08 pm

[Edit: deleted. Some how, an earlier post became duplicated in this thread.]
Last edited by daj95376 on Tue Oct 31, 2006 3:59 pm, edited 2 times in total.
daj95376
2014 Supporter
 
Posts: 2624
Joined: 15 May 2006

Postby RW » Sun Oct 29, 2006 8:35 pm

tarek wrote:What is the status of backdoors for this one ???

I can find several backdoors of size 2 (example r2c3 & r8c8). If you want a backdoor size one, there is only one possible that don't require further tabling in Sudocue, r8c6. Solved after that in Sudocue with one ALS, in SE with two forcing chains (SE rating 7.1), or one 7.4 Forcing chain...

I think the gsf rating for this puzzle also deserves to be mentioned: 99920!

RW
RW
2010 Supporter
 
Posts: 1010
Joined: 16 March 2006

Postby gsf » Sun Oct 29, 2006 8:44 pm

RW wrote:
tarek wrote:What is the status of backdoors for this one ???

I can find several backdoors of size 2 (example r2c3 & r8c8). If you want a backdoor size one, there is only one possible that don't require further tabling in Sudocue, r8c6. Solved after that in Sudocue with one ALS, in SE with two forcing chains (SE rating 7.1), or one 7.4 Forcing chain...

I think the gsf rating for this puzzle also deserves to be mentioned: 99920!

RW

the backdoor size up to and including multicoloring is 2, with 36 backdoor pairs
my solver doesn't do ALS
the nonstandard addition to rate this puzzle (mentioned above) is to use
all methods as the proposition constraint
based on that this is the toughest puzzle for the methods I've implemented
as usual, adding in more techniques will whittle the ratings down
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Postby Ocean » Mon Oct 30, 2006 7:41 am

ravel wrote:Our good connections to Finland made it possible that i was able to see another exraordinary puzzle, published in a local newspaper there. It is the first puzzle, that neither could be solved by SE nor by my program. gsf's program needed other than the default settings (adding coloring) to give a rating.
Since it was already published i hope, the author, ArtoI, is not offended to see it here:

AI Etana (Snail)
Code: Select all
 
+-------+-------+-------+
| 1 . . | . . 7 | . 9 . |
| . 3 . | . 2 . | . . 8 |
| . . 9 | 6 . . | 5 . . |
+-------+-------+-------+
| . . 5 | 3 . . | 9 . . |
| . 1 . | . 8 . | . . 2 |
| 6 . . | . . 4 | . . . |
+-------+-------+-------+
| 3 . . | . . . | . 1 . |
| . 4 . | . . . | . . 7 |
| . . 7 | . . . | 3 . . |
+-------+-------+-------+ 

Amazing puzzle! - not solved by Explainer, not rated by gsf's program (with default settings), and not rated by ravel's!

tarek wrote:I knew it,

The puzzle above follows the pattern which I submitted my top puzzle (The fluid drive)....... it is basically a diagonal pattern with an empty box. I was hoping that someone would genrate some more as I knew that pattern would harbour more difficult puzzles:D

Had a look into this pattern - hope it's ok with you. Far from finished with it, but here is a small taste of diagonal puzzles with an empty box. The other measures (ER, gsfr, suexrate) are not so extraordinary. Therefore I'm curious about the RBS. Also three puzzles that are not diagonal (one box deviates).

001002003020040010500600200006000004070000080300000900005001009080020070100900500 #M24 ER=9.2 gsfr=99396
001002003020040010500600700006000004080000090300000200003001009090060020700800500 #M24 ER=9.5 gsfr=99396
001002003000050020300600100006000004070000060500000800003004001060020050900700400 #M23 ER=9.2 gsfr=99367
000002003020040010500100600007000004010000080300000200005003002080070090600200500 #M23 ER=9.4 gsfr=99567
001002003000040010500600700003000004060000080900000200005007002080030070700200500 #M23 ER=9.4 gsfr=99385
001002003020040010300500600005000004070000080900000500000001009080050070600700300 #M23 ER=9.5 gsfr=99677
001002003020040050600700800005000004090000070300000200006008000010050090800200600 #M23 ER=9.8 gsfr=99316
001002000020030040400500300003000005060000070800000100002003009070080060000100800 #M22 ER=9.4 gsfr=99387
001002000020030040400500600006000005030000070800000100002005009070010030000900800 #M22 ER=9.5 gsfr=99338
001002000020030040500600700007000006030000080900000100002006004080040030000100900 #M22 ER=9.4 gsfr=99355
001002000030040050500600700007000008040000090600000100003008002090060040000100500 #M22 ER=9.5 gsfr=99346
001002300020030040500600007006000008070000090200000100009001005030040070000800400 #M23 ER=9.4 gsfr=99318
001002300040050060600300007002000008070000030800000100004001003050060070000400900 #M23 ER=9.3 gsfr=99356
001002300040030050500100006007000000060000080200000100009001004030050060000900700 #M22 ER=9.3 gsfr=99386


Another set: Puzzles with vertical symmetry (one or two might have been posted earlier, in a different isomorphic form). These have some interesting properties, but don't expect many with high RBS-scores.
The Explainer rating is higher than average, but not exceptional.
Two with extreme suexrates (above 1600, close race with the "current highest known"). Several above 1000.
Some high gsfr (Two above 99700. Fifteen with potential to enter the top-ten list if they pass the limit RBS>=8).

001000200030000040400050006000102000060070030000809000300000005050040060008000900 # gsfr=99731 ER=9.4
001000200030000040500060007000108000070030050000209000600000005050090060002000800 # gsfr=99719 ER=9.4
001000200030000040500060007000108000040070050000209000600000005070080060002000900 # gsfr=99696 ER=9.5
001000200030000040400050006000107000060040080000908000800000004050030060009000100 # gsfr=99689 ER=9.2
001000200030000040500060007000508000040030090000709000700000005090080030002000400 # gsfr=99687 ER=8.3
001000200030000040500030006000107000040060050000208000300000005060090030008000100 # gsfr=99680 ER=9.5
001000200030000040500030006000107000040060050000208000300000009060090030002000700 # gsfr=99678 ER=9.5 *suexrat=1429
001000200030000040500030006000107000040060050000208000300000009060090030008000100 # gsfr=99669 ER=9.4 *suexrat=1483
001000200030000040500060007000103000040080050000902000800000006050070080002000900 # gsfr=99668 ER=9.4
001000200030000040500030006000107000080040050000209000300000005040060030009000100 # gsfr=99668 ER=9.5
001000200030000040500030006000107000040060050000208000400000005060070030002000900 # gsfr=99667 ER=9.7
001000200030000040400050006000102000060030070000708000300000005050040030002000800 # gsfr=99661 ER=9.4
001000200030000040500060007000108000070030050000902000700000003050040060009000800 # gsfr=99649 ER=9.7
001000200030000040400050006000107000080030060000209000600000004050040080007000900 # gsfr=99649 ER=9.5 *suexrat=1630
001000200030000040500060007000108000040030050000209000700000006060070030008000900 # gsfr=99638 ER=9.7
#
001000200030000040500030006000107000080040050000209000600000004040060080002000700 # gsfr=99597 ER=9.6
001000200030000040400050006000106000060030070000208000500000001070040030009000800 # gsfr=99596 ER=9.7
001000200030000040400050006000107000060080030000209000300000008080040050002000900 # gsfr=99586 ER=9.5
001000200030000040500060007000108000070030050000209000700000003050040060002000900 # gsfr=99547 ER=9.6
#
001000200030000040400050006000107000060080030000209000300000008050040060009000100 # gsfr=99487 ER=9.3
001000200030000040500060007000108000040030050000902000300000004060070030002000900 # gsfr=99487 ER=9.3
001000200030000040500060007000108000070030050000209000400000003060040070009000800 # gsfr=99478 ER=9.4
001000200030000040500060007000108000070030050000209000400000006060040030008000100 # gsfr=99477 ER=9.3
001000200030000040400050006000107000060040080000908000800000004050030060007000100 # gsfr=99467 ER=9.3
001000200030000040500060007000108000070030050000902000700000003050040070002000800 # gsfr=99467 ER=9.4
001000200030000040500030006000104000070080060000902000300000008060050070009000100 # gsfr=99466 ER=9.3
001000200030000040500030006000102000040060050000708000600000004050090030007000800 # gsfr=99447 ER=9.5
001000200030000040400050006000107000060040030000802000300000004050080060002000100 # gsfr=99446 ER=9.4
001000200030000040400050006000107000060030080000902000800000005050040030009000100 # gsfr=99435 ER=9.3
001000200030000040400050006000102000070060030000805000700000003060040070002000800 # gsfr=99416 ER=9.3
#
001000200030000040500030006000107000040060080000902000300000008060050030002000700 # gsfr=99386 ER=9.4 *suexrat=1606
001000200030000040500030006000102000070050060000809000400000005060040030008000100 # gsfr=99385 ER=9.3
001000200020000030400050006000203000030070080000809000600000004080010070005000900 # gsfr=99377 ER=9.3
001000200030000040500030006000102000070040050000809000400000005050060030008000100 # gsfr=99376 ER=9.5 (But 9.4)
001000200030000040500060007000106000050070030000802000700000006040090050008000100 # gsfr=99375 ER=9.3
001000200030000040400050006000107000060030080000802000300000005050040030002000700 # gsfr=99368 ER=9.4
001000200030000040500060007000102000070040080000309000600000005040050060002000900 # gsfr=99365 ER=9.3
001000200030000040500060007000108000070030050000902000400000006060040030008000100 # gsfr=99359 ER=9.3
001000200030000040500060007000108000040030050000902000300000002060070030008000900 # gsfr=99357 ER=9.3
001000200030000040500030006000107000050060080000209000300000005080040030009000100 # gsfr=99357 ER=9.3
001000200030000040500060007000108000070040080000902000600000003040030070009000100 # gsfr=99339 ER=9.3
001000200030000040400050006000107000060030080000209000500000004080040050009000100 # gsfr=99336 ER=9.6
Ocean
 
Posts: 442
Joined: 29 August 2005

Postby ravel » Mon Oct 30, 2006 11:35 am

Wow, much to do for my program.

Having 65 puzzles now plus 10 from JPF's first 56 puzzles (highest rating 14 until now) plus the ones from the other 100 puzzles plus those from Ocean's lists (heavy weight for sure) i will probably reach 100 puzzles very soon.

Please be patient for my ratings.
ravel
 
Posts: 998
Joined: 21 February 2006

Postby tarek » Tue Oct 31, 2006 10:29 am

Thanx everybody regarding the backdoors issue.......

I was thinking, that if after SS gets stuck on a puzzle, that the backdoors at that stage of the PM grid would be indicative of the puzzle's difficulty.......

tarek
User avatar
tarek
 
Posts: 3762
Joined: 05 January 2006

Postby ronk » Tue Oct 31, 2006 1:06 pm

tarek wrote:The puzzle above follows the pattern which I submitted my top puzzle (The fluid drive)....... it is basically a diagonal pattern with an empty box. I was hoping that someone would genrate some more as I knew that pattern would harbour more difficult puzzles:D ...

You posted isomorphs of that base pattern here on Aug 30th with SER=9.2 and 9.3. Are you aware of anything earlier?
ronk
2012 Supporter
 
Posts: 4764
Joined: 02 November 2005
Location: Southeastern USA

Postby tarek » Tue Oct 31, 2006 3:31 pm

ronk wrote:You posted isomorphs of that base pattern here on Aug 30th with SER=9.2 and 9.3. Are you aware of anything earlier?

No.......... I noticed the diagonal pattern when I 1st rated the top 1465 then Vidarino reminded us all that the pattern does hold more toughies when he posted one of his monsters.....

I only tried to use an empty box with the pattern to insure that my puzzles would not be isomorphic to anything that has been published, because I haven't noticed any of those published before... The puzzles also should have a difficult 1st move.........not so many of those around either. ArtoI's puzzle was the 1st that i've seen with that pattern which I did not generate & I do not know exactly when he published it....

I attempted generating the pattern with 2 empty boxes (which has potential) -- but I managed to generate very few where the 2 boxes did not share a stack or band here, & DID not succeed in generating ANY that did share a stack or band:(

The short answer to your question is........NO, & I think that I'm PROBABALY the 1st to publish it, from the vast collections of hard puzzles that I browsed through before, I could be wrong though:!:

The way to look for them is easy......filter any minimal puzzle with an empty box [permuting any puzzle with an empty box still results in an empty box] then see if a diagonal pattern can be made....

tarek
User avatar
tarek
 
Posts: 3762
Joined: 05 January 2006

Postby JPF » Tue Oct 31, 2006 8:06 pm

I was wondering what is the meaning of a low suexrate (~250) and SE ≥9.0 ; gsfr ≥99250.

Here's an example :
suex=271 ; SE=9.1 ; gsfr=99262

Code: Select all
 . 5 . | . . 4 | 1 . .
 7 . 1 | . 8 . | . 4 .
 4 9 . | 1 . . | . . 6
-------+-------+-------
 . . 8 | 9 . . | . . .
 . 2 . | . 3 . | . . .
 6 . . | . . 5 | . . .
-------+-------+-------
 1 . . | . . . | 8 9 2
 . 8 . | . . . | 6 . .
 . . 9 | . . . | 3 . 5



here are some more :
Code: Select all
301007200080020050060500001007100000040080000200009000800000009020000013009000482   9,0   99314   238
098002400030040070600900005001800000040090000200001000400000097070000641006000800   9,0   99262   259
098002000030040070600900005001800000040090000200001000400000590070000040006000802   9,0   99252   239
008002100640090000705400006002100000070040000300008000200000901090000267007000350   9,0   99252   203
001009200950060040800500007005600000040070000300004000500000810030000490009000075   9,0   99344   256
 

JPF
JPF
2017 Supporter
 
Posts: 6132
Joined: 06 December 2005
Location: Paris, France

Postby Ocean » Tue Oct 31, 2006 10:33 pm

Two new interesting/extraordinary puzzles found in the diagonal pattern with an empty box (close to tarek's fluid drive pattern, but not isomorphic to that one).

Code: Select all
# "The Anemone"
# ER: "The Sudoku Explainer failed to solve this Sudoku using logical rules only."
# gsfr=99325 (3 2)
#
 *-----------*
 |..1|..2|...|
 |.3.|.4.|.5.|
 |5..|6..|3..|
 |---+---+---|
 |..7|...|..6|
 |.8.|...|.4.|
 |2..|...|1..|
 |---+---+---|
 |..6|..9|..7|
 |.4.|.7.|.8.|
 |...|5..|9..|
 *-----------*

Code: Select all
# "The Coral"
# ER:10.0
# gsfr=99736 (6 13)
#
 *-----------*
 |..1|..2|...|
 |.3.|.4.|.5.|
 |5..|6..|7..|
 |---+---+---|
 |..7|...|..6|
 |.1.|...|.8.|
 |9..|...|2..|
 |---+---+---|
 |..3|..7|..1|
 |.8.|.1.|.4.|
 |...|5..|9..|
 *-----------*


Thanks to Tarek for discovering the potential in these kinds of patterns, and to ArtoI for showing that there exists puzzles which Sudoku Explainer can not rate.
Ocean
 
Posts: 442
Joined: 29 August 2005

Postby tarek » Wed Nov 01, 2006 2:04 am

Ocean wrote:Two new interesting/extraordinary puzzles found in the diagonal pattern with an empty box
Nice ones

Ocean wrote:close to tarek's fluid drive pattern.
TOO CLOSE actually........ This forces me to post what I've got as they would show on whatever OCEAN is using in no time:D

This is the first list of fluid drives (the name gurth gave to this pattern in this configuration)....The first on this list is the 1st one which I posted earlier, followed by new interesting ones [more to come later]
All are minimal symmetric pearls
Code: Select all
Fluid Drive List #1
600001900080030000001700002002000007050000010400000300300009600000040080008200005,se9.9,gsfr99639,suexr687
600009800050030000002500007007000009010000020400000300300008600000040080009700001,se9.9,gsfr99667,suexr731
100005600060030000002900007007000009080000020400000300300009100000040060005700008,se9.9,gsfr99427,suexr719
900006800050030000001700009002000007070000020400000300300008900000040060005200001,se9.5,gsfr99638,suexr757
100004600060030000007500009009000002050000070400000300300008100000040020008200005,se9.5,gsfr99324,suexr704
600004900080030000005100002002000007010000050400000800300008600000040070007200009,se9.4,gsfr99486,suexr634
800006700050030000001400005009000002030000060400000100600007800000040050002900003,se9.4,gsfr99427,suexr569


tarek
User avatar
tarek
 
Posts: 3762
Joined: 05 January 2006

Postby RW » Thu Nov 02, 2006 3:43 pm

ocean wrote:# "The Anemone"
# ER: "The Sudoku Explainer failed to solve this Sudoku using logical rules only."

I'm very impressed. Took you only two days to find another puzzle with that special feature...

The puzzle actually has a backdoor cell that then let's you solve the rest with nothing but a finned x-wing and singles, r4c2=5. There's two other candidates of digit 5 in column 2. The first (r6c2) is eliminated by the shortest contradiction forcing chain found by Explainer, ER 9.2. So then all you need to do is get rid of the 5 in r7c2 and the puzzle is solved... But it doesn't seem to be a very easy task... Any suggestions of how to do it?

RW
RW
2010 Supporter
 
Posts: 1010
Joined: 16 March 2006

Postby tarek » Thu Nov 02, 2006 4:04 pm

RW wrote:Any suggestions of how to do it?
I checked the puzzle with rubylips's solver....

That elimination is in reach with a chain after about 24 placements....that solver on the way there had to guess 3 times........

I say within recach because placement #24 required a chain >15 nodes with stuff getting easier afterwards.......

The singles tail anyway is 22 placements long anyway. Happy hunting:D

tarek
User avatar
tarek
 
Posts: 3762
Joined: 05 January 2006

Postby ronk » Thu Nov 02, 2006 4:52 pm

Ocean wrote:
Code: Select all
# "The Coral"
# ER:10.0
# gsfr=99736 (6 13)
#
 *-----------*
 |..1|..2|...|
 |.3.|.4.|.5.|
 |5..|6..|7..|
 |---+---+---|
 |..7|...|..6|
 |.1.|...|.8.|
 |9..|...|2..|
 |---+---+---|
 |..3|..7|..1|
 |.8.|.1.|.4.|
 |...|5..|9..|
 *-----------*

Ocean, nice find. Noticed some time ago that you "sequentially normalize" the clues. Do you use a program for this?

Also curious as to your CPU and its speed. Or do you multitask on several?:)
ronk
2012 Supporter
 
Posts: 4764
Joined: 02 November 2005
Location: Southeastern USA

Postby tarek » Thu Nov 02, 2006 8:22 pm

gsf on another thread said that colouring was needed to establish the contradiction for AI Etana.......that is why there wasn't an RMS count for it & probably that is why SE can't, because it doesn't incorporate X-chains in the Chains+ or whatever they are called........

tarek
User avatar
tarek
 
Posts: 3762
Joined: 05 January 2006

PreviousNext

Return to General