Untouchables and Chameleons

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

Re: Puzzle with a six-mutable cell.

Postby gsf » Sat Nov 18, 2006 6:38 pm

you can use my solver to filter the max mutable cell
Code: Select all
-e 'mutable(sig) >= 6'

and then list the # solutions for each mutable cell (mutable(sig) "signature")
and a pencilmark grid with all of the mutable cell candidate values ( %(mutable(sig))h "hints")
Code: Select all
-f'%(mutable(sig))x %(mutable())h'


here's the mutable signature for ocean's 6 above
Code: Select all
6 2
6 2 2 2
6 3 3 2 2 2
6 2 2 2
6 2 2 2 2
6 3 2 2 2 2 2 2

i.e., first puzzle has two mutable cells, the first with 6 solutions, the second with 2

the solver also has functions for required() and superfluous() cells
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Twin

Postby Papy » Sun Nov 19, 2006 9:37 am

In the gordon file you have a clue disposition which accept 2 000 digits differents disposiitions!

I don't compute the candidats but if I have time I will give you this disposition
to study it.
Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Postby coloin » Sun Nov 19, 2006 10:26 am

Papy....we await !

Meanwhile ive found a 9-mutable set of puzzles.......or have I ?

Heres one puzzle which possibly works.....the potentially mutable clue at r9c9 :
Code: Select all
+---+---+---+
|7..|9..|4..|
|.6.|.5.|.7.|
|..9|..6|...|
+---+---+---+
|1..|4..|7..|
|.5.|.2.|.1.|
|..6|..3|...|
+---+---+---+
|3..|7..|...|
|.8.|.1.|...|
|...|...|..3|
+---+---+---+


Heres the rest

Code: Select all
7..9..4...6..5..7...9..6...1..4..7...5..2..1...6..3...3..7......8..1............1      123 sol.
7..9..4...6..5..7...9..6...1..4..7...5..2..1...6..3...3..7......8..1............2       47 sol.
7..9..4...6..5..7...9..6...1..4..7...5..2..1...6..3...3..7......8..1............3        1 sol.
7..9..4...6..5..7...9..6...1..4..7...5..2..1...6..3...3..7......8..1............4       72 sol.
7..9..4...6..5..7...9..6...1..4..7...5..2..1...6..3...3..7......8..1............5       33 sol.
7..9..4...6..5..7...9..6...1..4..7...5..2..1...6..3...3..7......8..1............6        0 sol.
7..9..4...6..5..7...9..6...1..4..7...5..2..1...6..3...3..7......8..1............7      278 sol.
7..9..4...6..5..7...9..6...1..4..7...5..2..1...6..3...3..7......8..1............8      123 sol.
7..9..4...6..5..7...9..6...1..4..7...5..2..1...6..3...3..7......8..1............9       29 sol.


Unfortunately they have many solutions !!!

Sorry but I just know you can do better !

C
Last edited by coloin on Sun Nov 19, 2006 7:52 pm, edited 1 time in total.
coloin
 
Posts: 1637
Joined: 05 May 2005

Twins...

Postby Papy » Sun Nov 19, 2006 12:59 pm

Ther e is an easy way to find this kind of grids

You take e file of grid an with an editor (Nottepad,Teachtext,...) you change all tyhe dogotd in one specific value

I/E 1..2....5...........1..9
After X..X....X...........X..X

You sort the result and the file will contains 10 lihnes 1 105 lines 2 1 Line 3....

I put the last tuch to my collection and Coloin will receice the magic link !!!

Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Re: Puzzle with a six-mutable cell.

Postby Ocean » Wed Nov 22, 2006 11:06 pm

gsf wrote:you can use my solver to filter the max mutable cell
Code: Select all
-e 'mutable(sig) >= 6'

and then list the # solutions for each mutable cell (mutable(sig) "signature")
and a pencilmark grid with all of the mutable cell candidate values ( %(mutable(sig))h "hints")
Code: Select all
-f'%(mutable(sig))x %(mutable())h'


Nice feature. Incredible how much functionality you can put into your program.

In my test-set several puzzles with a 4-mutable cell were found, a few with a 5-mutable cell, and only the one posted puzzle-set with a six-mutable cell.

Have you tried to run this function on your large puzzle collections?
Ocean
 
Posts: 442
Joined: 29 August 2005

Re: Puzzle with a six-mutable cell.

Postby gsf » Thu Nov 23, 2006 12:26 am

Ocean wrote:Nice feature. Incredible how much functionality you can put into your program.

Have you tried to run this function on your large puzzle collections?

thanks
and a good idea to burn idle cycles

my post has a typo, it should be
Code: Select all
-e 'mutable(max) >= 6'

and to get that to work you'll need the latest solver, just posted

its now filtering for puzzles with >=20 mutable cells or with one cell with >=6 mutable values

happy thanksgiving
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Mutable cell?

Postby Papy » Thu Nov 23, 2006 8:12 am

What do you camm 'mutable cell'?
What is the interet?
Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Re: Mutable cell?

Postby gsf » Thu Nov 23, 2006 2:17 pm

Papy wrote:What do you camm 'mutable cell'?
What is the interet?

a clue that can take on multiple values and still produce a puzzle with one solution
its described in detail earlier in this thread
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Postby Papy » Thu Nov 23, 2006 5:53 pm

Thanks coloin explain me
You s&ay thet you detect them :just trying each value for each clue?
Papy
Papy
 
Posts: 131
Joined: 15 August 2006

Re: Puzzle with a six-mutable cell.

Postby Ocean » Thu Nov 23, 2006 10:41 pm

gsf wrote:its now filtering for puzzles with >=20 mutable cells or with one cell with >=6 mutable values

Ok, waiting for result and statistics!
Ocean
 
Posts: 442
Joined: 29 August 2005

Postby coloin » Fri Nov 24, 2006 3:43 am

Papy wrote:You say that you detect them :just trying each value for each clue?


No I think its more of searching through millions of puzzles of a given pattern !

Some programmers have roving/clue iterating methods...Gfroyle made his 17s this way.

Are we going to have any advance on a 6-mutable? ...we will see

This pattern with a mutable clue on r9c9 looks good? [but isnt]
Code: Select all
+---+---+---+
|7..|9..|4..|
|.6.|.5.|.7.|
|..9|..6|...|
+---+---+---+
|1..|4..|7..|
|.5.|.2.|.1.|
|..6|..3|...|
+---+---+---+
|3..|7..|...|
|.8.|.1.|...|
|...|...|..3|
+---+---+---+


C
coloin
 
Posts: 1637
Joined: 05 May 2005

Postby gsf » Fri Nov 24, 2006 7:31 am

Papy wrote:Thanks coloin explain me
You s&ay thet you detect them :just trying each value for each clue?

yes, for a given puzzle, mutable clues are detected by checking all values for each clue

modifying a given puzzle to change the number of mutable clues or
the size of individual mutable clues is another question altogether
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Postby JPF » Fri Nov 24, 2006 9:51 pm

When a cell C is mutable in a puzzle P, the mutation gives a different puzzle P'.
but, it can happen that P and P' are equivalent.
Here is an example seen above from the Nick's list

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


has one mutable cell r5c5 = 1->8

The resulting puzzle P' :
Code: Select all
 1 . . | . 5 . | . . .
 . . . | 3 . . | . 6 .
 . . 8 | . . . | 2 . .
-------+-------+-------
 . 3 . | . . 2 | . . .
 9 . . | . 8 . | . . 5
 . . . | 6 . . | . 7 .
-------+-------+-------
 . . 5 | . . . | 1 . .
 . 6 . | . . 7 | . . .
 . . . | . 9 . | . . 8


is equivalent to P :

ravel wrote:Both are normalized to
Code: Select all
....5.6..4....9....8......12....4......8...9...5.1.3...6......5...2...4...1.3....
by gsf's program.

ronk wrote:Swap digits 1 and 8
Swap digits 3 and 7
Reflect diagonally

Any other example ?
What are the structure of such puzzles ?

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

Postby coloin » Fri Nov 24, 2006 11:44 pm

Code: Select all
+---+---+---+
|123|456|789|
|7..|328|.61|
|6.8|.7.|2.3|
+---+---+---+
|.37|..2|.16|
|9.6|7..|325|
|.12|63.|.74|
+---+---+---+
|..5|263|1.7|
|36.|8.7|..2|
|27.|.9.|638|
+---+---+---+


here is the pseudopuzzle with 2 grid solutions..

the two grid solutions
Code: Select all
123456789759328461648971253437582916986714325512639874895263147361847592274195638
123456789794328561658179243537942816946781325812635974485263197369817452271594638


the two grid solutions have the same "bands" and can be transformed into each other
Code: Select all
C:\suxx>index416 test3.txt
162 411 301  , 222 290 255
255 290 222  , 301 411 162 


Red Ed knows all the grids which do this !!!!! He counted them up and posted the Sudoku symmetry group

C
coloin
 
Posts: 1637
Joined: 05 May 2005

Postby JPF » Sun Nov 26, 2006 10:31 pm

coloin wrote:...
the two grid solutions have the same "bands" and can be transformed into each other
Code: Select all
C:\suxx>index416 test3.txt
162 411 301  , 222 290 255
255 290 222  , 301 411 162 

Red Ed knows all the grids which do this !!!!! He counted them up and posted the Sudoku symmetry group

Thanks coloin,
I must say that your post is a bit crytic for me.
Sorry if I'm the only one !

What I understood :
The 2 following grids G1 and G2 are equivalent :
Code: Select all
123456789759328461648971253437582916986714325512639874895263147361847592274195638
123456789794328561658179243537942816946781325812635974485263197369817452271594638

I checked that by the only canonicalization programm I know how to use :
gsf's program :
Code: Select all
sudoku -qFN -f%%#0c puzzles.txt > puzzles.can

btw, Is there a way to find one series of permutations which allows to go from G1 to G2 ?

Now,
My question* was related to one mutable cell in one valid puzzle.
How to use your observation for this specific question ?
Thanks,

JPF

*The question is : is there any specific properties for the puzzles P such that Mut(P)~P
Mut(P) is a (one cell) mutation of P.
JPF
2017 Supporter
 
Posts: 3752
Joined: 06 December 2005
Location: Paris, France

PreviousNext

Return to General