17-clue and 18-clue Sudoku update

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

Postby Red Ed » Tue Dec 20, 2005 9:56 pm

<deleted stupid mistake re number of 17s>
Last edited by Red Ed on Wed Dec 21, 2005 4:02 pm, edited 1 time in total.
Red Ed
 
Posts: 633
Joined: 06 June 2005

Postby Wolfgang » Mon Jan 09, 2006 10:00 am

The last weeks i tried two modifications of my greedy search. First i restricted it to an area round an arbitrary chosen 17-clue pattern, which i searched rather intensively (700000 clue changes, the last 50000 i did not even find a trivial 18-clue there, 445 before). All the about 250 17-clues i found have been in the list.
After altering the method slightly again i now found 31 17-clues so far, only this one is not in the current list:
Code: Select all
120300000400200000000000500607000800000120000000000000005008000000047010000000003
Wolfgang
 
Posts: 208
Joined: 22 June 2005

Postby coloin » Sat Jan 13, 2007 3:31 am

Code: Select all
003000009050008200000007000609000300000501070000000000070000010500060000000090000
003000009050100000700200000040000030800700000000009002602000100000043000000000000
003000009050100000800200000040000300700800000000009002602000010000043000000000000
003000009050200000800700000200000050000043006000090000004000000030000000000800270
003000009050700000700200000040000300800100000000009002602000070000043000000000000
003000009050800000800200000040000030700100000000009002602000800000043000000000000
003000009070800000800200000046007000000000850000000020500000000000063000000040003
003000080050009000700200000200000003000038010000004000048000000000600207000000000
003000080050009000700200000200000003000038090000004000048000000000600207000000000
120400000500000000000900630030000090000005070000001000200000004090300000000000001
120400000500000000000900630030000090000005070000001000600000004090300000000000001
120400000500000000000900630030000090000005800000001000008300000090000000000004001
120400000500000000000900630030000900000031800000005000090600000000000001000000020
120400000500000000000900630030000900000081004000005000800000001090600000000000000
120400000500000000000900630030007900000015800000000000090600000000000001000000002
120400000500000000000900630030700902090000000000000001400001000000005000000000300
120400000500000000000900630039000800000005001000000000200001004080000900000000000
120400000500000000000900630040000800000031070000005000090670000000000001000000000
120400000500000000000900630060000970000015000000000000001003000000700092000000000
120400000500000000000900630070800900060000000000000001000000802000035000000010000
120400000500000000000900630079000000000003502000000001040000090000025000000000000
120400000500000000000900630090000002004700000000000800007000900000025001000000000
120400000500000000000900630090002000030000900000005800800000000000000021000300000
120400000500000000000900630096000800000005001000000000200001000000000904000800000
120400000500000000000900630096000800000005001000000000200301000080000900000000000
120400000500000000000900630200000004090300000000000001030000090000005070000001000
120400000500000000000900630200001000000000904000800000096000800000005001000000000
120400000500000000000900630200001800004000300000005000006300000400000001000000000
120400000500000000000900630200301000080000900000000000096000800000005001000000000
120400000500000000000900630300001000000000470000000900007600000009000000000005001
120400000500000000000900630300001000004000070000030900009000000007000000000005001
120400000500000000000900630300005800040000900000001000009600000000003001000000000


Papy has forwarded these 17-sudoku puzzles.

They might be new - Gordon's list hasnt changed for some time.......currently 36628.

Thanks

C
coloin
 
Posts: 1637
Joined: 05 May 2005

Postby JPF » Sat Jan 13, 2007 8:28 am

coloin wrote:Papy has forwarded these 17-sudoku puzzles.
They might be new - Gordon's list hasnt changed for some time.......currently 36628.

No news.
#9 is invalid.
Here are the Gordon's equivalents :
Code: Select all
    # Papy  #  gfroyle                                                           
                                                                               
        1       20187                                                           
        2       15211                                                           
        3       15211                                                           
        4       35626                                                           
        5       15210                                                           
        6       15210                                                           
        7       12496                                                           
        8        5125                                                           
        9     invalid                                                           
       10        7881                                                           
       11        7880                                                           
       12        4805                                                           
       13       35133                                                           
       14        6998                                                           
       15       24695                                                           
       16       27857                                                           
       17       10939                                                           
       18       31031                                                           
       19       33648                                                           
       20       35581                                                           
       21          17                                                           
       22       14309                                                           
       23         464                                                           
       24        7247                                                           
       25       28148                                                           
       26        7881                                                           
       27        7247                                                           
       28       10788                                                           
       29       28148                                                           
       30       34723                                                           
       31        1696                                                           
       32       16894                                                           

JPF
JPF
2017 Supporter
 
Posts: 3752
Joined: 06 December 2005
Location: Paris, France

Postby Papy » Sat Jan 13, 2007 9:52 am

It strange : I have all the solution for my puzzle.
I don't know what soft you use to test them and compare with the Gordon file
I try with the gsf code and C14N them : no error

Code: Select all
123400000500000000000700600200005900090000800000001000060900000000000031000000000
123400000600000000000070500047000900000102000000000000080090000000000013000000002
123400000600000000000070500047000900000201000000000000080090000000000013000000002
123400000600000000000070500080095000000000014000000000059000800000100060000000000
123400000600000000000070500080095000000000014000000000095000800000100060000000000
123400000600000000000070500090085000000000041000000000058000900000100006000000000
123400000600000000000070500090085000000000041000000000085000900000100006000000000
123400000700000000000800500050600000900000040000000071000031000000000400000070000
123400000700000000000800500080300000000000016000000007600000000000000430000017000
123400000700000000000900500050600000800000040000000017000031000000000400000070000
123400000700000600000100000608000007000502010000000000010000040000080000000060000
123400000700900000000000500400000001000005020000080000059000800000107000000000000
123400000800000500000900000090000040000078010000005000040300020000000800000000000
123400000800000500000900000090000060000078010000005000060300020000000800000000000
123400000800000600000090050060308000000100002000000000050060000000000803000000000
123400000900000000000000500040000800000309000000710000050068000000000010000000030
123400000900000000000000506040000008000035000000010000670800000000000130000000000
123400000900000000000000506670800000000000130000000000040000008000035000000010000



Try and THANKS FOR TRYING

I solve the 9 and send you the solution (if I find it!!:!)

I think hat when you compare with the Gordon file yours test don't take care of the value of the digit because one of my method to generate was to eliminate intouchables and find all combinaisons
I eliminate isomorh and c14n the rest

I also find new dispositions.

Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Postby Papy » Sat Jan 13, 2007 11:00 am

Sorry 1000 times !!!!
The problem 9 is really bad :

I have make and extraction for Colin but noty in the good file

I controle the file again
My grid are all sorted and the firest rox is always 123456789
just for test. It's easy to scramble them every solver do it bt it's more easy
for working
Sorry

Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Postby coloin » Sat Jan 13, 2007 6:59 pm

Well done Papy

I was going to say - you have obviously got a method to make 17s.....so why dont you post the whole lot to see if you have found any new ones....

gfroyle ran his program for a full year......

But I believe the list that you posted above "look" to be new !

It depends on the normalization you use........just because they begin 1234 doesnt mean they are new of course....and many of Gfroyles 17s which have 3 consecutive clues in a box can be made to start 1234

C
coloin
 
Posts: 1637
Joined: 05 May 2005

Postby Papy » Sat Jan 13, 2007 10:01 pm

I will tell you my methd but I prefer other control on the result to be sure
of it.
But I use intouchables....

Papy

The new game of papy make grid with only 3 box
(a block) with 17 clues...
Hard...
Papy
 
Posts: 131
Joined: 15 August 2006

Postby gsf » Mon Jan 15, 2007 9:36 pm

Wolfgang wrote:After altering the method slightly again i now found 31 17-clues so far, only this one is not in the current list:
Code: Select all
120300000400200000000000500607000800000120000000000000005008000000047010000000003

this is isomorphic to this one in gordon's catalog
Code: Select all
000570200308000000000000000074000030020010000000000006600308000800400000000000100
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Postby gsf » Mon Jan 15, 2007 9:45 pm

coloin wrote:Well done Papy

I was going to say - you have obviously got a method to make 17s.....so why dont you post the whole lot to see if you have found any new ones....

gfroyle ran his program for a full year......

But I believe the list that you posted above "look" to be new !

It depends on the normalization you use........just because they begin 1234 doesnt mean they are new of course....and many of Gfroyles 17s which have 3 consecutive clues in a box can be made to start 1234

C

JPF showed these puzzles are not new
one cannot determine sudoku isomorphism based on "normalization" unless the normalization itself is based on isomorphism

papy has been known to use non-isomorphic techniques to determine non-equivalence (as he described in a recent "Isomorph" post)
but these techniques fail to verify equivalence
e.g., if two puzzles have different #clues then they are not equivalent does not imply that
two puzzles with the same number of clues are equivalent
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Postby ravel » Mon Jan 15, 2007 10:30 pm

gsf wrote:this is isomorphic to this one in gordon's catalog
Code: Select all
000570200308000000000000000074000030020010000000000006600308000800400000000000100

But it has not been in the list one year before - Wolfgang was the first, who published it.
ravel
 
Posts: 998
Joined: 21 February 2006

Postby gsf » Mon Jan 15, 2007 10:43 pm

ravel wrote:
gsf wrote:this is isomorphic to this one in gordon's catalog
Code: Select all
000570200308000000000000000074000030020010000000000006600308000800400000000000100

But it has not been in the list one year before - Wolfgang was the first, who published it.

aha
and how many documents have I dated 2006 this january ...
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Postby Papy » Tue Jan 16, 2007 9:29 am

Hi,

Exploring and analysing a block I make a very strange discovery.
A new latin square magic secret

My algo was this:

If I cannot solve a bloc with 16 clues and 27 cells I will cannot solve a grid with 81 celle ans 16 clues.
So I take blocks and try with 16 clues: no solution: It's not possible to solve a 16 clues block

I beleive that my theory was good!!! And a big surprise controling my result To be sure I try to solve the block with 17 clues Impossible also You need 18 clues
So the reult is here

to solve a block you need 18 cluesbut to solve 3 blocks only 17!!!!

There is a reason but the result is fun! (I let you search a little)


Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Postby coloin » Tue Jan 16, 2007 11:25 pm

Yes Papy, "unluckily" your 17s were all in the list . Wolfgang demonstrated how hard it was to find 17s a year ago- and he eventually found one which at the time was not on the list - which was good going.

If I understand how you were trying to do it, by finding mutable clues in Gordons list....it is underestimating how hard Gordon's program worked.

If you want to find more 17s [if there are many more that is ] you would have to do rather more than remove one clue and search around......

I suspect Gordon was working at the 15 clue stage and multing a furthur 2 clues - I say this because doing it with 14 clues and multing 3 increases the possibilites rather a lot......... He may even have done deeper searches !

Your last post is interesting, if not rather similar to a reply I have given you in a recent post.

I agree 17 clues within a band can never define a band but........
Code: Select all
+---+---+---+    +---+---+---+
|7.3|8..|...|    |753|842|6..|
|...|6..|4..|    |891|657|432|
|2..|...|...|    |264|319|7..|
+---+---+---+    +---+---+---+
|...|...|.73|    |186|294|573|
|.4.|1..|...|    |547|1..|2..|
|...|...|...|    |932|765|8..|
+---+---+---+    +---+---+---+
|61.|...|9..|    |615|4..|927|
|...|.2.|1..|    |379|526|1..|
|..8|.7.|...|    |428|971|3..|
+---+---+---+    +---+---+---+  17 clues solves the first vertical band


I am not sure how common a collection of 17 clues fully solves a band [but not the full grid].....but it has to be pretty rare......

I also agree that if we can prove 16 clues cant define a band then it therefore means 16 clues cant define a full grid.
C
coloin
 
Posts: 1637
Joined: 05 May 2005

Postby gsf » Wed Jan 17, 2007 6:49 pm

Ruud and I have been pm'ing with papy
there have been some communication problems
resulting a a few posts with invalid puzzles and/or dups in gordon's catalog

but in a list papy sent to ruud there were 2 not in gordon's catalog
Code: Select all
120000080400900000000007003097000500000084000000000000800020010000500900000000000
120006000400000010000700500005010000006000900000004000008900000600000000000000042

these are the first new entries in over a year
congratulations papy
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

PreviousNext

Return to General