The New Sudoku Players' Forum

by **coloin** » Sun Jun 03, 2012 9:44 pm

Automorphisms aside. 2 great finds by eleven. I cant get close to making these puzzles.

looking back at what serg has done - and the puzzles we have.

If we look at any of the puzzles we see that there is always at least 8 clues in boxes 12358

Code: Select all: x x x x x

serg has shown that there cant be only 4 clues in total in these 4 boxes. [there can be 5 but that wouldnt give a symetric 18 puzzle]
whch leaves 6 clues in eqivalent boxes 12358
like this crude attempt with > 20 clues

Code: Select all: +---+---+---+ |...|5.8|...| |.4.|...|.9.| |...|.7.|...| +---+---+---+ |..6|...|1..| |7..|...|..6| |5..|...|..4| +---+---+---+ |1.3|.6.|..7| |.5.|...|82.| |.8.|...|.4.| +---+---+---+

i couldnt remove any more clues from boxes 7 and 9.
might be difficult to prove that is impossible though
except there cant be many more ED symm patterns with 6 clues in B12358 [with full B4679]
here is 8

Code: Select all: ...673......2.1.......8....439...287567...491812...563395...178741...625628...349 #060909909 ...6.3...............584...439...287567...491812...563395.2.178741...625628...349 #050909919 ...6.3......5.4.......8....439...287567...491812...563395.2.178741...625628...349 #050909919 ...6.3.............7..8..1.439...287567...491812...563395.2.178741...625628...349 #131909919 ....3.....7.....1....1.6...439...285567...491812...763395.2.178741...526628...349 #131909919 ...6.5............9...8...6439...287567...491812...563395.2.178741...625628...349 #131909919 ...185....1.....7..........123...956456.2.837789...214875...423931...765642...189 #131919909 ............1.4...9.......2439...287567...491812.9.563395...178741...625628.7.349 #121919919

C

by **eleven** » Mon Jun 04, 2012 8:21 am

Code: Select all: +-------+-------+-------+ | . . . | . . . | . . . | | . . 1 | . . . | 2 . . | | . 3 . | . 4 . | . 5 . | +-------+-------+-------+ | . . . | 1 . 5 | . . . | | . . 6 | . . . | 7 . . | | 2 . . | . . . | . . 4 | +-------+-------+-------+ | . . . | 7 2 6 | . . . | | . 5 . | . . . | . 8 . | | . 1 . | . . . | . 3 . | +-------+-------+-------+ ...........1...2...3..4..5....1.5.....6...7..2.......4...726....5.....8..1.....3. #110 eleven

Serg wrote:How many vertical simmetric puzzles do you know? (Some time ago you published 385 vertical simmetric puzzles. What is their actual number now?)Serg

Currently i have 393 listed below. [Edit: i made an error when i tried to sort them after the pattern, now sorted after puzzles strings]

Hidden Text: Show

by **Serg** » Mon Jun 04, 2012 7:39 pm

Hi, eleven!

eleven wrote:
Code: Select all
+-------+-------+-------+ | . . . | . . . | . . . | | . . 1 | . . . | 2 . . | | . 3 . | . 4 . | . 5 . | +-------+-------+-------+ | . . . | 1 . 5 | . . . | | . . 6 | . . . | 7 . . | | 2 . . | . . . | . . 4 | +-------+-------+-------+ | . . . | 7 2 6 | . . . | | . 5 . | . . . | . 8 . | | . 1 . | . . . | . 3 . | +-------+-------+-------+ ...........1...2...3..4..5....1.5.....6...7..2.......4...726....5.....8..1.....3. #110 eleven

Fantastic! You find new 18-clue symmetric patterns almost every day. Congratulations! I checked your pattern number 110 - it is new 18-clue vertical symmetric pattern.
Thank you for 393 symmetric puzzles' list. I checked them for minimality - all 393 puzzles (including 110 puzzles representing patterns) are minimal.
Could you say some details of your search?

I am updating my classification (clue over boxes distributions list).

Serg

by **ronk** » Mon Jun 04, 2012 8:52 pm

eleven wrote:
coloin wrote:Is this how you found this 18 ?
No, i made {-5+5} on known puzzles.

Did you use a [-5+5] generator specific to the vertical symmetry pattern? By this I mean, removal and addition of an odd number of clues from c5, and the vertically symmetry removal and addition of pairs of clues elsewhere.

by **eleven** » Tue Jun 05, 2012 8:56 am

Yes i wrote a special {-5+5} method using the set of possible patterns, Serg and i had calculated.

As said, it takes about 2 days/core for one puzzle and i dont know, if this is fast or slow compared to a classical way, as you described it.
Probably it could be done in the half time, but i dont have any time to optimise it.

In a first step i go through the patterns and list all transformations of the puzzle, which have at least 13 common cells.
Typically i get about 300000 sub-puzzles, where then i have to do a {-n+5} inside this pattern then. n is 0 for 13 common clues and 5 for 18 (the puzzle's pattern).

by **eleven** » Fri Jun 08, 2012 12:03 am

ALMOST all known of these symmetric 18's build a {-5+5} cluster, but these 2 are an island (pattern #60 by Ano1):

Code: Select all: ...........1...2....3.4.5.....5.6......7.1.....7...8...9.....7..2..9..6.6.......1 ...........1...2....3.4.5.....6.7......8.1.....8...6...9.....8..2..9..7.7.......1

With {-6+6} they connect to the others.

It was easy to implement, what Mladen advised me, using Michael Deverin's subcanon() function. The cluster calculation took about 4 hours then.

by **Serg** » Fri Jun 08, 2012 3:42 pm

Hi, eleven!

eleven wrote:ALMOST all known of these symmetric 18's build a {-5+5} cluster, but these 2 are an island (pattern #60 by Ano1):
Code: Select all
...........1...2....3.4.5.....5.6......7.1.....7...8...9.....7..2..9..6.6.......1 ...........1...2....3.4.5.....6.7......8.1.....8...6...9.....8..2..9..7.7.......1

With {-6+6} they connect to the others.

It's curious, but if we relabel the first puzzle by scheme 6 -> 7, 7 -> 8, 8 -> 6, we'll see that both puzzles differs by one clue only:

Code: Select all: ...........1...2....3.4.5.....5.7......8.1.....8...6...9.....8..2..9..7.7.......1 ...........1...2....3.4.5.....6.7......8.1.....8...6...9.....8..2..9..7.7.......1

So, these both puzzles are within {-1,+1} vicinity.

Serg

by **coloin** » Sat Jun 16, 2012 11:58 pm

Well .... false alarm
Here are 4 new puzzles to elevens list, but not new pattern unfortunately

Code: Select all: ............1.2....34...56.....3....7.......28.......7...8.7.....1...3....6.5.4.. ............1.2....34...56.....3....7.......28.......7...8.7.....1...4....6.5.3.. ............1.2....34...56.....5....7.......88.......1...8.7.....6...9....5.3.4.. ............1.2....34...56.....5....7.......88.......2...7.8.....6...1....5.3.4.. ............1.2....34...56.....6....7.......81.......7...7.8.....6...2....9.5.3.. ............1.2....34...56.....6....7.......82.......7...8.7.....6...9....1.5.3.. ............1.2....34...56.....5....1.......78.......3...9.8.....7...4....5.3.9..# new ............1.2....34...56.....5....7.......42.......8...7.9.....8...3....5.4.9..# new ...........1...2...3..4..5....4.6.....2...1....7.8.5.....2.7....6.....3..9.....4. ...........1...2...3..4..5....6.4.....2...7....6.1.8.....2.7....4.....1..9.....3.# new ............1.2.....3...4.....5.6.....1...5....7.4.8...6.....2..5..8..3.2.......7 ............1.2.....3...4.....5.6.....7...1....6.7.8...4.....5..2..3..6.1.......2 ............1.2.....3...4.....5.6.....7...1....6.7.8...4.....5..2..9..6.1.......2 ............1.2.....3...4.....5.6.....1...7....6.4.8...5.....1..4..7..5.2.......6# new

C

by **coloin** » Sun Jun 17, 2012 9:58 pm

A new pattern at last.

Code: Select all: +---+---+---+ |...|...|...| |...|1.2|...| |.32|...|45.| +---+---+---+ |...|.4.|...| |6..|...|..7| |8..|.3.|..1| +---+---+---+ |...|6.7|...| |..5|...|2..| |..4|...|9..| +---+---+---+ ............1.2....32...45.....4....6.......78...3...1...6.7.....5...2....4...9.. # 111 coloin

This is within {-5+5} of one of elevens known ones. But i actually found it from one of the new ones i found
The pattern is one clue [of 2 in column 5] exchanged.

Code: Select all: similarity 13 ............1.2....32...45.....4....6.......78...3...1...6.7.....5...2....4...9.. ............1.8....39...45.....4....6.......77.......8...6.7.....5...2....4.3.9..

Found by generating many symmetrical 19s [with 3 clues in column 5]

C

by **eleven** » Mon Jun 18, 2012 1:08 pm

Congratulations on the new puzzles and patterns.

In fact there were 17 puzzles (in 12 patterns) in the list, which are inside {-5+5} of the new pattern.

[Added] This is the list of the 398 known puzzles ordered by patterns.

Hidden Text: Show

by **coloin** » Fri Jun 22, 2012 11:21 pm

Another new pattern

Code: Select all: +---+---+---+ |...|.1.|...| |..2|...|3..| |.1.|4.5|.6.| +---+---+---+ |...|...|...| |.2.|...|.5.| |.4.|.6.|.1.| +---+---+---+ |...|3.2|...| |..3|...|7..| |7..|...|..8| +---+---+---+ ....1......2...3...1.4.5.6...........2.....5..4..6..1....3.2.....3...7..7.......8 #112 coloin

and two more of a known pattern

Code: Select all: ...........1...2...3..4..5.....5......6...7..7.2...8.1...2.8....1.....3..8.....4. ...........1...2...3..4..5.....6......7...8..8.2...9.1...2.9....1.....3..9.....4. ...........1...2...3..4..5.....5......6...7..2.7...8.1...2.8....8.....3..1.....4. new ...........1...2...3..4..5.....9......6...7..2.7...8.1...2.8....8.....3..1.....4. new

C

by **Serg** » Sun Jun 24, 2012 11:47 am

Hi, coloin!
I've checked your patterns #111 and #112 - both are new unique patterns. Congratulations!
Both published puzzles are minimal too.

I am updating my classification list (list of sym. 18-clue patterns, groupped by distributions of clues over boxes).

Serg

by **eleven** » Tue Jun 26, 2012 8:28 am

Well done, Coloin.

You discovered a new island, the puzzle with the new pattern is not within {-5+5} of any other known one (two were in {-6+6})
And there are more than 400 puzzles now.

I also updated the lists.

Posted new patterns:

Hidden Text: Show

Sorted patterns:

Hidden Text: Show

All puzzles sorted after patterns:

Hidden Text: Show

by **coloin** » Tue Jun 26, 2012 10:18 pm

Thanks serg and eleven
I would like to say there was subtlety to my search - but the -6+6 shows there isnt any !

for good measure
here is another new pattern

Code: Select all: +---+---+---+ |...|...|...| |.1.|2.3|.4.| |..5|...|6..| +---+---+---+ |...|657|...| |.3.|...|.8.| |...|.9.|...| +---+---+---+ |..9|...|2..| |..6|...|5..| |.2.|...|.3.| +---+---+---+ ..........1.2.3.4...5...6.....657....3.....8.....9......9...2....6...5...2.....3. # 113 coloin

C

by **eleven** » Wed Jun 27, 2012 8:20 am

Good job, keep them coming.
The last one had 20 puzzles in {-5+5}.

I will wait a while with the next list update

[Added:] How many 19's do you have so far ?

The New Sudoku Players' Forum

Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s

Re: Symmetric 18s