|
View previous topic :: View next topic |
Author |
Message |
| daj95376
| Joined: 05 Feb 2006 | Posts: 349 | : | | Items |
|
Posted: Wed Feb 15, 2006 5:10 pm Post subject: Asymmetric vs. Symmetric Clues |
|
|
I understand there's been considerable discussion on whether or not the clues to a puzzle should be symmetric. I'm not here to dig up those discussions. However, I have an observation and (hopefully) a new side-issue to discuss.
I've collected a number of puzzles offered in the forums and from web sites mentioned in the forums. So, I decided to test the symmetry of their clues.
I further tested the symmetric puzzles to see how many had naked singles immediately after loading.
In one group of 121 puzzles, 42 were asymmetric and 79 were symmetric. Of the latter, 36 had naked singles.
Another group was a collection of puzzles with 17-clues each, all 35,396 were asymmetric.
In the final group of 1,465 puzzles, 1,449 were asymmetric and 16 were symmetric. Of the latter, 9 had naked singles.
So, my observation is that most of the puzzles that I've collected are asymmetric. Of the symmetric puzzles, many of them have naked singles showing immediately after loading.
My topic for discussion, comes in two parts:
(1) If you have a generator, does it generate symmetric or asymmetric puzzles?
(2) Should initial naked singles count towards whether or not a puzzle is considered symmetric?
Thanks! |
|
Back to top |
|
|
| Ruud Site Admin
| Joined: 17 Sep 2005 | Posts: 708 | : | Location: Netherlands | Items |
|
Posted: Wed Feb 15, 2006 7:18 pm Post subject: |
|
|
Hi,
the collection of 17-clue Sudokus is completely asymmetrical because nobody has found a symmetrical 17 clue sudoku yet. The set is not representative for all sudokus.
The top1465 contains a lot of 17 clue sudokus from Gordon's collection, so you're counting a lot of them twice.
I have a mixed collection of more than 600,000 symmetrical and asymmetrical sudokus. My generator can produce either of them, and also various types of symmetry (vertical, horizontal, diagonal, rotational)
So far, I have not found any correlation between symmetry and complexity. The only difference is that a symmetrical sudoku cannot be fully minimalized without losing symmetry.
A naked single to start of with is not very rare, but a sudoku that requires naked subsets or any other advanced technique right from the start is very rare.
The collections you checked are made for a special purpose. The 17-clue collection is obvious, but top1465 is assembled from most difficult to solve sudokus by a backtracking solver. These sudokus tend to be a little more difficult than average, but there are also some very easy sudokus in the collection.
I'm not sure where you got the group of 121, so I cannot say anything about it.
Ruud. _________________ Meet me at sudocue.net |
|
Back to top |
|
|
| vidarino
| Joined: 10 Feb 2006 | Posts: 38 | : | Location: Haugesund, Norway | Items |
|
Posted: Wed Feb 15, 2006 7:28 pm Post subject: |
|
|
Ruud wrote: | The only difference is that a symmetrical sudoku cannot be fully minimalized without losing symmetry. |
That's true in many cases, but not all. Here's a symmetric and minimal puzzle:
Code: | +-------+-------+-------+
| 9 3 8 | . 2 4 | . . . |
| . 1 . | . . . | . . . |
| . 5 6 | . . 7 | . . . |
+-------+-------+-------+
| . 2 . | 1 . 5 | 8 . . |
| . . . | . . . | . . . |
| . . 3 | 4 . 8 | . 1 . |
+-------+-------+-------+
| . . . | 7 . . | 2 4 . |
| . . . | . . . | . 3 . |
| . . . | 2 3 . | 1 6 5 |
+-------+-------+-------+
|
That was from a collection of 2682 symmetric puzzles. Only a dozen or two were minimal, so it's certainly not commonplace.
Vidar |
|
Back to top |
|
|
| Ruud Site Admin
| Joined: 17 Sep 2005 | Posts: 708 | : | Location: Netherlands | Items |
|
Posted: Wed Feb 15, 2006 8:04 pm Post subject: |
|
|
You are correct, Vidar.
It should have been: The only difference is that a symmetrical sudoku cannot always be fully minimalized without losing symmetry.
Ruud. |
|
Back to top |
|
|
| fermat
| Joined: 05 Feb 2006 | Posts: 25 | : | Location: Melbourne | Items |
|
Posted: Thu Feb 16, 2006 12:07 am Post subject: |
|
|
Ruud wrote: | You are correct, Vidar.
It should have been: The only difference is that a symmetrical sudoku cannot always be fully minimalized without losing symmetry.
Ruud. |
Cool. I expected that symmetry ALWAYS paid a price. |
|
Back to top |
|
|
| Cristian2006
| Joined: 25 Feb 2006 | Posts: 6 | : | | Items |
|
Posted: Mon Feb 27, 2006 12:36 pm Post subject: |
|
|
It's true what you said about minimizing (it's of course very easy to minimize if it doesn't matter what does it look like) but at least IMO there is a correlation between
making more easily a hard puzzle with (for example) 24 givens if it's symmetric
and
making equally hard puzzle with the same amount of givens if it's asymmetric
So I think if you want to compare symmetric and asymmetric puzzles, they all should be with the same amount of givens. At least I have seen very hard puzzles with more than 30 givens, but having equally difficult asymmetric puzzles are more rare, I think.
I think someone could also examine, what might have been the most minimal puzzles for example to be solved only with hidden and naked singles and opposite to that what might have been the most maximal puzzles for example requiring X-Wing. And at the same time check how many there are asymmetric, how many symmetric. I think there has to be more correlations in sudoku that has examined so far.
Me myself call only symmetric puzzles sudoku. Asymmetric is "Number Place" to me. I really like symmetric. _________________ My new website (doesn't include sudokus, at least yet):
http://www.cristianium.fi
(it's in finnish) |
|
Back to top |
|
|
| Richard
| Joined: 25 Feb 2006 | Posts: 2 | : | | Items |
|
Posted: Tue Feb 28, 2006 6:45 pm Post subject: |
|
|
Personally, I think it's good for a puzzle, however difficult, to have at least a couple of singles (naked or hidden) to start with. It helps the human solvers get into it. There's nothing more demoralising than looking at a new puzzle and taking five minutes or more just to work out your first number for one cell.
I always generate symmetric puzzles for my site because I like the symmetry and I think it's good to have some singles to get people started.
But talking about symmetry, one observation I have made is that it seems to be easier to make hard puzzles with diagonal symmetry than with horizontal or vertical symmetry. I haven't taken the time to figure out why but maybe someone else will find the observation interesting.
Richard |
|
Back to top |
|
|
| Cristian2006
| Joined: 25 Feb 2006 | Posts: 6 | : | | Items |
|
Posted: Fri Mar 03, 2006 12:12 am Post subject: |
|
|
Richard wrote: |
But talking about symmetry, one observation I have made is that it seems to be easier to make hard puzzles with diagonal symmetry than with horizontal or vertical symmetry. I haven't taken the time to figure out why but maybe someone else will find the observation interesting.
Richard |
That's true and the explanation is easy, I think. You know, sudoku as a game isn't diagonal at all. If you make puzzles with horizontal or vertical symmetry, already when you are making cells interfere much with each other. And, there's also other explanation: relations in those simple sudoku technics are technically ALL horizontal and vertical. But advanced technics can affect much more cells that simple leave out, and those related cells could be diagonally placed... _________________ My new website (doesn't include sudokus, at least yet):
http://www.cristianium.fi
(it's in finnish) |
|
Back to top |
|
|
| Richard
| Joined: 25 Feb 2006 | Posts: 2 | : | | Items |
|
Posted: Fri Mar 03, 2006 8:33 am Post subject: |
|
|
Ahh, of course. I hadn't thought about it that way. Thank's Cristian.
Richard |
|
Back to top |
|
|
| Lummox JR
| Joined: 07 Sep 2005 | Posts: 202 | : | | Items |
|
Posted: Sun Mar 19, 2006 9:27 pm Post subject: |
|
|
As a related thought here, there are four different ways to consider diagonal symmetry. One is the traditional crossword style which is simply a 180° rotation, which I presume is what you meant. Another is to mirror the clues along the NW-SE axis, and another along the NE-SW axis. Finally the last form does both.
Surprisingly, I've found that mirroring along both diagonal axes does tend to produce harder puzzles than you might expect, about the same as crossword symmetry--which is interesting considering there are more cells per symmetrical group.
My own generator is presently capable of assymetry, crossword-style, quad (90° rotations), horizontal, vertical, horizontal and vertical, and the dual diagonal form. |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
Igloo Theme Version 1.0 :: Created By: Andrew Charron
|