Sudoku Programmers Forum Index

 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister   ProfileProfile   Log inLog in          Games  Calendar

Log in to check your private messagesLog in to check your private messages   

Uniqueness based methods, are they allowed?

 
Post new topic   Reply to topic    Sudoku Programmers Forum Index -> Solving sudoku
View previous topic :: View next topic  

Do you agree with the use of solving techniques which presume a unique solution?
Yes, always
10%
 10%  [ 1 ]
Only when I trust the maker of the sudoku
30%
 30%  [ 3 ]
Only after my program has validated the sudoku
30%
 30%  [ 3 ]
Never
30%
 30%  [ 3 ]
Don't know or no opinion
0%
 0%  [ 0 ]
Total Votes : 10

Author Message
Ruud
Site Admin
Joined: 17 Sep 2005
Posts: 708
:
Location: Netherlands

Items
PostPosted: Sun Feb 19, 2006 10:10 pm    Post subject: Uniqueness based methods, are they allowed? Reply with quote

I will first show you a quote by Angus, in another topic:

angusj wrote:
Since I did mention 'Uniqueness' above, perhaps I'd better clarify my feeling on that. It is a very clever technique which is reasonably simple and seems prevalent enough to be quite useful for difficult puzzles. However, it does require the presupposition that a puzzle has a single solution, something I personally find less than satisfying as to me part of the charm of solving Sudoku is to prove there is just one solution.

This is a clear opinion. You can see the poll question, but replies to this topic with your motivations are also welcome.

Here is my opinion:

When I know the source of a sudoku is reliable, or when I tested the sudoku with my backtracking solver, I just know there is a unique solution, and there is no need to prove it. My task as a player is to find it, not to prove it.

One of the regular players of my daily nightmares asked me a very strightforward question here to be ensured he could use these techniques. For him, asking this question was part of the solving process. Many sources of sudokus guarantee that their sudokus have a unique solution that can be found with logic.

Here is an analogy: When I buy a car with ABS, traction control, airbags all around and dozens of other safety features, would I be wrong if I took a little more risk driving it, assuming that these safety devices will work?

So, using uniqueness tests and BUG variations are nice shortcuts, but with a warning sign "Use at your own risk!". I would certainly not keep them away from my users. In SudoCue, I've added an option that allows the user to choose whether or not to use them.

Ruud.
Back to top
View user's profile Send private message Visit poster's website
angusj
Site Admin
Joined: 18 Jun 2005
Posts: 406
:

Items
PostPosted: Mon Feb 20, 2006 12:00 am    Post subject: Re: Uniqueness based methods, are they allowed? Reply with quote

Yes, this is purely personal preference the same as "backtracking" is personal preference.
Back to top
View user's profile Send private message Visit poster's website
fermat

Joined: 05 Feb 2006
Posts: 25
:
Location: Melbourne

Items
PostPosted: Mon Feb 20, 2006 3:16 am    Post subject: Another option on the poll? Reply with quote

I would use it before guessing but after everything I usually do gets no further.

If it goes wrong? Caveat emptor!
Back to top
View user's profile Send private message
Carcul

Joined: 29 Dec 2005
Posts: 50
:
Location: Coimbra, Portugal

Items
PostPosted: Mon Feb 20, 2006 9:21 am    Post subject: Reply with quote

Ruud wrote:
When I know the source of a sudoku is reliable, or when I tested the sudoku with my backtracking solver, I just know there is a unique solution, and there is no need to prove it. My task as a player is to find it, not to prove it.


But by using Uniqueness you are not proving that the puzzle has only one solution! You don't need to do that because you already know that it has only one solution. Uniqueness is a logical argument that make use of the fact that the puzzle has only one solution. What is wrong with that?! Now, if someone don't find this technique "elegant" that is ok, but "elegance" is a subjective issue. Personally, I find a solution based on Uniqueness very elegant. When solving a puzzle, if you use Uniqueness you are not proving that it has just one solution, you are simply being clever by making use of that fact.

AngusJ wrote:
It is a very clever technique which is reasonably simple and seems prevalent enough to be quite useful for difficult puzzles.


First, sometimes it is not that simple to apply (I am thinking in the very advanced applications of Uniqueness, like Almost Almost Unique Rectangles, Almost Almost Unique Loops, and others Almost Almost Unique Patterns), and second, it can be not only usefull with difficult puzzles but also with extremely difficult puzzles.

Ruud wrote:
So, using uniqueness tests and BUG variations are nice shortcuts, but with a warning sign "Use at your own risk!".


And what is the risk?? The only "risk" that I can see is the one of getting a shorter and more elegant solution. But, of course, this is just a personal opinion.

Carcul
Back to top
View user's profile Send private message Send e-mail
evert

Joined: 30 Aug 2005
Posts: 68
:
Location: Amsterdam

Items
PostPosted: Mon Feb 20, 2006 11:35 am    Post subject: Reply with quote

Maybe one could create a puzzle with more solutions where uniqueness tests force into one specific solution:
Quote:
This puzzle has one unique solution if and only if it has one unique solution.
Back to top
View user's profile Send private message Send e-mail
Ruud
Site Admin
Joined: 17 Sep 2005
Posts: 708
:
Location: Netherlands

Items
PostPosted: Mon Feb 20, 2006 1:36 pm    Post subject: Reply with quote

evert wrote:
Maybe one could create a puzzle with more solutions where uniqueness tests force into one specific solution

When you apply uniqueness based methods on a sudoku with multiple solutions, it either leads to one of these solutions, giving the false impression that there is a single solution, or it leads to the false conclusion that the sudoku has no solution at all, because you end up with contradictions.

I've seen both happen several times.

Check this definition:

Wikipedia wrote:
Inference is the act or process of drawing a conclusion based solely on what one already knows.

We are using inference to solve sudokus. No argument about that. So, when I know there is only one solution, then I should be allowed to use that knowledge in my reasoning.

Ruud.
_________________
Meet me at sudocue.net
Back to top
View user's profile Send private message Visit poster's website
Display posts from previous:   
Post new topic   Reply to topic    Sudoku Programmers Forum Index -> Solving sudoku All times are GMT
Page 1 of 1

 
Jump to:  
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
Sudoku Programmers topic RSS feed 


Powered by phpBB © 2001, 2005 phpBB Group

Igloo Theme Version 1.0 :: Created By: Andrew Charron