The New Sudoku Players' Forum

[champagne]

The scan of solution grids producing 18 clues puzzles having a band with 2 clues is closed and produced 106 puzzles, only 3 more than the known by coloin.
Ill published later to-day the list of puzzles with more comments

[champagne]

The good new is that the run produced 8 new 18 having a band with 2 clues

Hidden Text: Show

Code: Select all

.....6..............9........7..8.5..6...3.4.......1..3...1.6....5.9...8.4.5....7
........9..............1....1...4...3....2..8.9......65..6..3....29...1...8.7..4.
........9..............1....1...4...3....7..89.......6..29...1...7.2..4.8..6..3..
.....6..............9.......3...7.5........1...8..2.9...41....8.1.9..6...7..4...3
.....6..............9.......3...7.5........1...8..2.9...21....8.1.9..6...7..4...3
......7..............2.....2...4...55....3.....8.7.1...1.8..9...6.5....2.7.....4.
..3......................6.2....5.9......8..2..73...4.5...2......4.9...39...6...8
....5..............8........3......8.6..1...3..7.6...2...9..4....2..8.9...5..3.1.


The bad new is that some known are missing although they have been processed successfully in check mode.
As I worked in slices, I restart the process in once to see if I made a handling error.
But it can be a bug!!!

[champagne]

It was a bug in the entry file generator.
The 5 missing 18 are now there, no fresh 18 came.

So at the end

the total number of 18 clues puzzles having 2 clues in a band is 111
103 known by coloin
8 more found here

I am nearly sure that this is now a bug free result.

The (dirty) corresponding code will be as process 500 in my skmppv2 next repository update.
I did not check that they all are minimal, but I'll do it to be safe. Assuming that they are minimal, we have here a code directly proving that no 17 clues puzzles having 2 clues in a band can be produced.

And more important for me, I have now some potential to speed up the 17 clues search. I started the corresponding implementation. I could have some ideas of the improvement potential (if any) within 2 weeks.

[blue]

Hi champagne,

I am nearly sure that this is now a bug free result.

I can confirm the count of 111.

And more important for me, I have now some potential to speed up the 17 clues search. I started the corresponding implementation. I could have some ideas of the improvement potential (if any) within 2 weeks.

Good challenge. So far I have nothing to do that, but after the results of the search of 18s having a band with 2 clues I see
a possible alternative design for the 17 clues search, this is for to-morrow
a possible design to explore a solution grid in the area 18-20 if the first idea works

Good luck, and have fun
Blue.

[champagne]

Hi blue,
Thank you first for confirmation.

Regarding the next variant of things you have done, as the main change in in the "valid XY" generator, I should have enough code tested in 2 or 3 weeks (not that much time to spend on this project currently) to have a better view of the validity of my positive feeling. As all test have to be re run, I'll change also things in the "phase 2", but nothing important

[m_b_metcalf]

the total number of 18 clues puzzles having 2 clues in a band is 111
103 known by coloin
8 more found here.

FWIW, 13 of these puzzles can transformed to have rotational symmetery. Here are 3 examples:

Code: Select all

 . . . . . 1 . 2 .
 3 . . . . 2 . 4 .
 5 . . . . 6 . 7 .
 4 . . . . . . . .
 . . . . . . . . .
 . . . . . . . . 1
 . 2 . 8 . . . . 9
 . 1 . 7 . . . . 3
 . 7 . 4 . . . . .   ED=1.5/1.2/1.2

 . 1 . . . 2 . . .
 3 . . . . 4 . 5 .
 6 . . . . 7 . 8 .
 5 . . . . . . . .
 . . . . . . . . .
 . . . . . . . . 4
 . 4 . 9 . . . . 1
 . 7 . 1 . . . . 2
 . . . 6 . . . 3 .   ED=3.4/2.0/2.0

 . . 1 . . 2 . . .
 3 . . . . 4 5 . .
 6 . . . . 7 3 . .
 5 . . . . . . . .
 . . . . . . . . .
 . . . . . . . . 7
 . . 7 5 . . . . 2
 . . 2 8 . . . . 1
 . . . 3 . . 9 . .   ED=4.4/1.2/1.2


Mike