SudokuSolver Forum
http://www.rcbroughton.co.uk/sudoku/forum/

Near Sizedoku 2 & 4
http://www.rcbroughton.co.uk/sudoku/forum/viewtopic.php?f=13&t=1532
Page 1 of 1

Author:  HATMAN [ Wed Dec 11, 2019 11:55 am ]
Post subject:  Near Sizedoku 2 & 4

Near Sizedoku 2

David nice puzzles but I find the name cumbersome could we use something like 3-Futoshiki?

I've done this in JSudoku so the arrows are automatic. In any case I've no Idea how to do David's arrows.

The GT signs show that one cell value is greater than another but by no more than 3. All GT signs within a Nonet are given.

Corrected Puzzle

Image

I'm going to post solutions from now on as it is easy to make mistakes with these (three out of four were wrong and I have yet to check the fourth one).
Solution:
785496312623185497914372856872619534136254978459738261547921683368547129291863745



While the puzzles are interesting it is not easy in my view to make difficult puzzles as most Nonets practically solve on their own. The missing arrows give significant information, almost to the level of consecutive clues. Hence solve the puzzle again assuming that not all GTs are given. It is still not too hard but I'll make harder ones.

It is unique as a Futoshiki but I was not able to solve it, however I know I'm weak on this type of puzzle, perhaps one of you can solve it that way?

Author:  Quadrata [ Wed Dec 11, 2019 4:08 pm ]
Post subject:  Re: Near Sizedoku 2

It is sometimes hard to control the difficulty in these "if and only if" style puzzles where lack of a clue also gives some information. Maily because I can't just drop some clues to make the puzzles harder! Sometimes I'll come up with two symbols, like I'll use an arrow if the entries are near, a dot if they are far, and then let a blank line just mean "no information." Looking forward to trying this one. I use a program called yEd to do my arrows, but I like your arrows just as much if not more. They aren't too busy.

Author:  HATMAN [ Sat Dec 14, 2019 10:11 pm ]
Post subject:  Re: Near Sizedoku 2

Near Sizedoku 4

These give a lot more information than a futoshiki so if you start with a solvable futoshiki this puzzle will be easy. If the futoshiki is unique but not solvable it will probably be not to difficult. Hence we need to start with a solution that is not unique as a futoshiki. This will often be the case for a random solution, but I had the thought that a non-consecutive solution would normally fail as a futoshiki. This is the case in the ones I have tried, but I'm not sure why.

This one is based on an NC and I had to add 2 givens to make it solvable. It is unique with one given and may be unique without any givens but I cannot do the solving using JSudoku's recursively solve.

All arrows are given and it is quite hard. If you are feeling lazy or are stuck you can use the NC property.

Corrected Puzzle (LT r9c78 added) - the givens were not necessary


Image

Solution:
147528639693174852258639417572496183814753926369281574736915248481362795925847361

Author:  Quadrata [ Sun Dec 15, 2019 5:28 am ]
Post subject:  Re: Near Sizedoku 2 & 4

This new one was really fun. It was much easier than the one above it, because I need the "non-clue" arrowless lines to mean the entries are a certain distance apart in order to use some of my methods. I can't say whether or not I needed both clues, I should try it again without the 5. I will say that the cage with the 6 is a strong place to start, because all entries in that cage are immediately determined except for two of them, and then that cage determines the center cage completely. I enjoyed it a lot.

Page 1 of 1 All times are UTC
Powered by phpBB® Forum Software © phpBB Group
https://www.phpbb.com/