Untouchables and Chameleons

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

Postby coloin » Sun Nov 09, 2008 10:17 am

I dont think there will be a rush to find these "rare" chameleons !

I generated a few [17000] minimal 25-puzzles from the mc grid. 2% were chameleons......
Code: Select all
+---+---+---+
|..3|4.6|...|
|..9|.2.|45.|
|.56|...|1..|
+---+---+---+
|2..|5.4|.9.|
|...|8..|..1|
|8..|.3.|...|
+---+---+---+
|...|...|..8|
|..5|97.|3..|
|...|3..|.4.|
+---+---+---+ 25 clues


Why this should be is unclear.
coloin
 
Posts: 2384
Joined: 05 May 2005
Location: Devon

Postby Mauricio » Fri Nov 14, 2008 6:20 am

A couple of 9-mutables, r1c1 can be any digit:
Code: Select all
+-------+-------+-------+
| . . . | . . . | . . . |
| . . . | . . 1 | . . 2 |
| . . . | . 3 . | 4 5 . |
+-------+-------+-------+
| . . 4 | 6 . . | . . 3 |
| . . 2 | . 5 . | 7 6 . |
| . 7 . | . . 8 | . . . |
+-------+-------+-------+
| . 4 5 | 1 . . | . 9 . |
| . 2 . | . 7 . | 1 . . |
| . 6 . | 4 . 2 | . . 8 |
+-------+-------+-------+
Code: Select all
+-------+-------+-------+
| . . . | . . . | . . . |
| . . . | . . 1 | . 2 3 |
| . . . | . 2 . | 4 5 . |
+-------+-------+-------+
| . . . | . . 6 | 7 . 5 |
| . . 3 | 2 5 . | 8 9 . |
| . 5 4 | 9 1 . | 2 3 6 |
+-------+-------+-------+
| . 4 . | 1 . 2 | 3 8 . |
| . 3 . | . 8 . | 1 . . |
| . 9 . | 4 7 3 | 5 6 2 |
+-------+-------+-------+
Mauricio
 
Posts: 1175
Joined: 22 March 2006

Postby coloin » Fri Nov 14, 2008 7:30 am

VERY Well done:!:

I didnt think it was possible.........

These are definitely rare puzzles.

So, if you didnt find them by random generation [im assuming you didnt] please tell us how you made them....

C
coloin
 
Posts: 2384
Joined: 05 May 2005
Location: Devon

Postby Mauricio » Fri Nov 14, 2008 7:33 am

A couple more, but these two have the extra property that setting r1c1 any digit, it always gives a minimal puzzle (.. bug found, sudokus deleted)

How I did it?
I generated some sudokus that did not have clues in either r1,c1 or box 1, had few solutions (<30) and that after setting r1c1 any clue, it created no contradictions (ie, it had at least 1 solution). Then I did {+1-1} on these puzzles, filtered those sharing the previous properties, and let it run overnight and in the morning there were already some 9-mutables! To be honest I did not expect it to be that easy. In all it took less than 10 hours, between programming, finding some seeds and then finding some solutions.
Mauricio
 
Posts: 1175
Joined: 22 March 2006

Postby JPF » Fri Nov 14, 2008 8:30 am

Nice job, congratulations:!:

JPF
JPF
2017 Supporter
 
Posts: 6127
Joined: 06 December 2005
Location: Paris, France

Postby Mauricio » Mon Nov 24, 2008 5:20 pm

Chameleons with 20 clues (I hope):
Code: Select all
..............1..2..3.4..56.....6....4........5..3.7..1.....8..2...7....6..25..74
..............1..2..3.4..56.....6....4........5..3.7..17....8..2...7....6..25...4
..............1..2..3.4..56.....6....4........5..7.3..1.....8..2...3....6..25..34
..............1..2..3.4..56.....6....46.......5..3.7.....25..741.....8..2...7....
..............1..2..3.4..56....7.....6.25...471....8...2...6...4........5...3.7..
Mauricio
 
Posts: 1175
Joined: 22 March 2006

Postby coloin » Wed Nov 26, 2008 4:04 am

They would be "exceedingly rare puzzles":!:

Unfortunately every clue is not mutable.

Its quite likely there would be a 24 clue chameleon in one of a series of randomly generated puzzles from the mcgrid.

i couldnt quite get gsfs program to do the filtering

Code: Select all
-e 'mutable()==clues'


but easily found a randomly generated 23 clue chameleon from the mcgrid !

Code: Select all
+---+---+---+
|.2.|...|.8.|
|..9|1..|...|
|4..|7..|.23|
+---+---+---+
|..1|5..|...|
|...|.9.|23.|
|...|.3.|..4|
+---+---+---+
|3..|.45|9..|
|..5|...|...|
|.78|...|6..|
+---+---+---+  23 clues


This 24 clue came out after the 23 !

Code: Select all
+---+---+---+
|...|...|7..|
|.89|1..|...|
|...|7..|.23|
+---+---+---+
|2..|..4|.9.|
|5..|.9.|.3.|
|..7|...|5..|
+---+---+---+
|.1.|6.5|9..|
|.4.|...|...|
|..8|31.|.4.|
+---+---+---+ 24 clues


and a 22

Code: Select all
+---+---+---+
|...|.5.|...|
|7..|...|.56|
|4..|.89|1..|
+---+---+---+
|..1|...|8..|
|.6.|...|...|
|...|23.|..4|
+---+---+---+
|3..|6..|.7.|
|...|9..|.12|
|..8|.1.|...|
+---+---+---+  22 clues


Some of the resultant puzzles are automorphic it would seem.

C
coloin
 
Posts: 2384
Joined: 05 May 2005
Location: Devon

Postby gsf » Wed Nov 26, 2008 7:13 am

coloin wrote:i couldnt quite get gsfs program to do the filtering

Code: Select all
-e 'mutable()==clues'


this should work
are there any error messages?
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Postby coloin » Wed Nov 26, 2008 7:38 am

Actually......it does work !

I took your command instructions too literally last week.

Code: Select all
 -e 'mutable()==0 || mutable()==clues'


Code: Select all
C:\temp>sudoku -e 'mutable()==clues' mcgrid22.txt
sudoku: <<<'mutable()==clues': invalid token

C:\temp>sudoku -e "mutable()==clues" mcgrid22.txt
....5....7......564...891....1...8...6..........23...43..6...7....9...12..8.1.... #     9 FN C22.m


Well.... much quicker than looking at the colour of the squares in Havard's gui program !

C
coloin
 
Posts: 2384
Joined: 05 May 2005
Location: Devon

Previous

Return to General