puzzle with uniqueness type 3?

Advanced methods and approaches for solving Sudoku puzzles

puzzle with uniqueness type 3?

Postby Havard » Thu Feb 02, 2006 10:35 am

Hi.

I have been trying to find a puzzle that "needs" the UR type 3, and was wondering if anyone got one lying around?

To be more specific: A puzzle that won't solve using the just the "normal deductions", strong-links based patterns (x-wing, sword, jelly, turbot), xy, xyz etc... with no chains, nishio, tabeling etc allowed, that will crack open with a UR type 3. Have examples of type 1,2 and 4 that fulfills this, but not 3.

Code: Select all
Needs UR type 1 and 2:
. . . | . . . | 2 9 .
9 . . | . . 8 | 4 7 .
. . 7 | . . 9 | . 1 5
------+-------+------
. 4 2 | 9 . 6 | . . .
. . . | . 3 . | . . .
. . . | 7 . 4 | 8 5 .
------+-------+------
7 9 . | 1 . . | 5 . .
1 2 8 | 4 . . | . . .
. . 5 | . . . | . 8 .


Code: Select all
Favorite of mine!
Needs "everything" (xw, sf, tf, xy etc) and a UR Type 4
. . . | . 9 . | 4 . .
8 6 . | 7 . . | . . .
7 . . | . 6 . | 8 . .
------+-------+------
4 . . | . 8 . | 9 . .
. 5 . | 1 . 6 | . 7 .
. . 3 | . 7 . | . . 4
------+-------+------
. . 9 | . 2 . | . . .
. . . | 9 . . | . 8 .
. . 8 | . 3 . | 2 5 .


But I am missing one that needs UR type 3! I should add that I have found a few puzzles where the UR type 3 CAN be used, but in those cases it has ALSO been a type 4 (which is easier to deal with)

Can anyone help?

Havard

edit: took out a puzzle with (juck) multiple solutions...:)
Last edited by Havard on Thu Feb 02, 2006 5:36 pm, edited 1 time in total.
Havard
 
Posts: 378
Joined: 25 December 2005

Postby Wolfgang » Thu Feb 02, 2006 11:08 am

I dont have a definition for UR type 1,2,3,4. Is there a link for it or can you explain it please? What type is tso's here?
Wolfgang
 
Posts: 208
Joined: 22 June 2005

Postby Carcul » Thu Feb 02, 2006 11:41 am

Hi Wolfgang.

Try here for a good explanation (The Ultimate Guide to Unique Rectangles - version 0.2, by MadOverlord).
And the puzzle of Tso have a Type-3 Unique Rectangle.

Regards, Carcul
Carcul
 
Posts: 724
Joined: 04 November 2005

Postby Havard » Thu Feb 02, 2006 2:58 pm

Hi.

If it is this one you mean:
Code: Select all
8 . . | 2 . 5 | . . 1
. . . | 1 . 3 | . . .
. . 3 | . 7 . | 8 . .
------+-------+------
6 3 . | . . . | . 7 5
. . 8 | . . . | 2 . .
9 1 . | . . . | . 4 8
------+-------+------
. . 5 | . 9 . | 1 . .
. . . | 7 . 6 | . . .
3 . . | 5 . 2 | . . 9


...then it can be solved with a UR type 4.

The hunt for a type 3 is still on!:)

Havard

edit: I said it needed 2 T.Fish, but it don't... There are two of them, but they are not needed...:)
Havard
 
Posts: 378
Joined: 25 December 2005

Postby Havard » Thu Feb 02, 2006 3:56 pm

I dont have a definition for UR type 1,2,3,4. Is there a link for it or can you explain it please?


Sure I'll try:)

The unique rectangle itself is defined by four cells forming a rectangle that satisfy these criteria:
1: all four cells must have the same two candidates in them
2: two parallell sides of the rectangle must share a box and a line(row or column), the other two sides need only to share the same line.
3: One pair of cells that shares one line (or box and line) must contain ONLY the two candidates, the other pair of cells can contain other candidates. (these "other" candidates is what defines the type of rectangle...:) In other words you can't have the two cells with "other" candidates across from each other!

So lets use this puzzle as an example:

Code: Select all
. . . | . . . | 2 9 .
9 . . | . . 8 | 4 7 .
. . 7 | . . 9 | . 1 5
------+-------+------
. 4 2 | 9 . 6 | . . .
. . . | . 3 . | . . .
. . . | 7 . 4 | 8 5 .
------+-------+------
7 9 . | 1 . . | 5 . .
1 2 8 | 4 . . | . . .
. . 5 | . . . | . 8 .


In this one there is both a Type 1 and Type 2. Now let's start with the Type 1 (though the type 2 comes first)

look at this:
Image

As you can see, the UR is marked with the red lines, and has the candidates 2 and 4. Now in a Type 1 there is only ONE of the cells that has any other candidates then 2 and 4, and here that is the one with 6 and 9 marked with green. The logic goes that if that cell is either 4 or 2, then the puzzle has multiple solutions, hence we can eliminate them (marked in red).

now lets look at a Type 2:
Image

Again the UR is marked with red, and here it has the candidates 1 and 7.
Now the definition of a Type 2 is that there are one extra candidate in two cells, and that those candidates are the same. Here a 9, marked in green. Now the logic goes: We don't know which one of them is a 9, but one of them HAS to be a nine to avoid the "multiple solution" dilemma. This means that we can kill of any other nines in the same sectors (sectors marked in blue, other nines marked in red)

and a Type 4:

using the sudoku from a previous post:
Code: Select all
8 . . | 2 . 5 | . . 1
. . . | 1 . 3 | . . .
. . 3 | . 7 . | 8 . .
------+-------+------
6 3 . | . . . | . 7 5
. . 8 | . . . | 2 . .
9 1 . | . . . | . 4 8
------+-------+------
. . 5 | . 9 . | 1 . .
. . . | 7 . 6 | . . .
3 . . | 5 . 2 | . . 9

Image

A type 4 is very handy, because unlike type 1-3 it does not matter how many extra candidates you have in your two cells.
Again the UR is marked with red lines, and is here 4 and 9.
Now if you look in column 4 (marked with blue) there are no other possible place for the number 9 than in our UR (marked with green). Hence ONE of those cells HAS to be a 9. Now say you pick one of them and assign a 9 to it... Then if the OTHER cell becomes a 4, you have a multiple solution sudoku... Now apply the same logic to the other cell. As you can see, the number 4 (marked in red) can not be in either of those cells, and you can eliminate them both! Notice also how the "other" candidates (in this case 6 and 3,6) does not play a part at all!:)

I have not separated between A and B types, but A is basically when your two cells with other candidates in them shares both a line and a box, and B types is when they only share a line.

Hope that is clear!:)

now give me a sudoku that requires type 3!:) (and can not be solved with 1,2 and 4 instead):)

havard
edit: typos...
Havard
 
Posts: 378
Joined: 25 December 2005

Postby Carcul » Thu Feb 02, 2006 4:25 pm

Hi Havard.

Havard wrote:Notice also how the "other" candidates (in this case 6 and 3,6) does not play a part at all!


Your example of a type-4 UR is also an example of a type-3 UR, where the extra candidates in two of the cells of the UR form a N-tuple with N-1 cells in the same unit. In your example, we know that "3" or "6" must be in one of r3c4/r5c4 to prevent a multiple solution situation. But we have a bivalue node on "3,6" in r6c4, and so the cells r3c4/r5c4/r6c5 function as a naked pair on "3,6", which allows elimination of "3" from r7c4.

BTW, here is a puzzle where you can use a type-3 UR:

Code: Select all
 4 . 1 | . 6 . | . . .
 3 . . | . . . | 2 . .
 . . . | . . . | 8 . .   
-------+-------+------
 1 5 . | 2 . . | . . .
 6 . . | . . . | . 1 .
 . . . | 9 . . | . . .
-------+-------+------
 . 2 . | 7 . 8 | . . .
 . . . | . . . | . 4 3
 . 7 . | . . . | . . .

Regards, Carcul
Carcul
 
Posts: 724
Joined: 04 November 2005

Postby CathyW » Thu Feb 02, 2006 5:25 pm

I'm puzzled by the second puzzle in Havard's post:!:

Have got to this point:
Code: Select all
 
{8}   {7}   {6}   {4}   {5}   {3}   {2}   {1}   {9}   
{3}   {9}   {4}   {1}   {8}   {2}   {6}   {7}   {5}   
{2}   {5}   {1}   {79}  {6}   {79}  {4}   {3}   {8}   
{6}   {4}   {7}   {2}   {19}  {8}   {19}  {5}   {3}   
{5}   {1}   {2}   {79}  {3}   {6}   {79}  {8}   {4}   
{9}   {3}   {8}   {5}   {17}  {4}   {17}  {2}   {6}   
{7}   {6}   {3}   {8}   {4}   {1}   {5}   {9}   {2}   
{4}   {2}   {59}  {3}   {79}  {579} {8}   {6}   {1}   
{1}   {8}   {59}  {6}   {2}   {59}  {3}   {4}   {7}   

Now - uniqueness would indicate that r8c6 = 7, but according to Simple Sudoku that's an invalid move; BUG would indicate that r8c6 = 9, but again that's an invalid move. So how is this one solved (apart from having made invalid moves that mean r8c6=5!)?
CathyW
 
Posts: 316
Joined: 20 June 2005

Postby Wolfgang » Thu Feb 02, 2006 5:30 pm

CathyW wrote:I'm puzzled by the second puzzle in Havard's post:!:

This puzzle has 3 solutions.
Wolfgang
 
Posts: 208
Joined: 22 June 2005

Postby CathyW » Thu Feb 02, 2006 5:39 pm

Thanks Wolfgang! That explains it - I obviously didn't notice the warning when I first pasted it in to Simple Sudoku.
CathyW
 
Posts: 316
Joined: 20 June 2005

Postby eclark » Thu Feb 02, 2006 6:03 pm

I can't even find ones that need type 2. I only see type 2b's everything else I see can be solved with other methods.....


for example The one you have posted for type2 has a type one with <42>

can anyone give me one ?
eclark
 
Posts: 14
Joined: 26 January 2006

Postby Wolfgang » Thu Feb 02, 2006 6:23 pm

Hi Havard,

thanks for your explanations, together with MadOverlord's definitions i hope, i understood it now. Then this Angus' unlimited021 should be a "pure" type 3:
Code: Select all
 *-----------*
 |92.|61.|..3|
 |...|.32|...|
 |...|49.|1..|
 |---+---+---|
 |85.|..7|...|
 |..6|.4.|7..|
 |...|9..|.52|
 |---+---+---|
 |..3|.79|...|
 |...|15.|...|
 |4..|.63|.81|
 *-----------*

Wolfgang
 
Posts: 208
Joined: 22 June 2005

Re: puzzle with uniqueness type 3?

Postby flip » Thu Feb 02, 2006 7:17 pm

Havard wrote:I should add that I have found a few puzzles where the UR type 3 CAN be used, but in those cases it has ALSO been a type 4 (which is easier to deal with)

A type 4 reduction destroys the unique rectangle (see the MadOverlord reference given by Carcul). You should do any other possible UR reductions first, then type 4. As you observed, type 3 is often also type 4.
Last edited by flip on Fri Feb 03, 2006 2:18 am, edited 2 times in total.
flip
 
Posts: 17
Joined: 08 January 2006

Re: puzzle with uniqueness type 3?

Postby Wolfgang » Thu Feb 02, 2006 9:34 pm

Havard wrote:[code]...
Needs "everything" (xw, sf, tf, xy etc) and a UR Type 4

My friend just solved it with the UR and simple coloring alone. Only if you use coloring for another number, you come to the swordfish and xy-wing:)
Wolfgang
 
Posts: 208
Joined: 22 June 2005

Postby Havard » Thu Feb 02, 2006 9:35 pm

hi

Carcul:
thanks for your feedback! I know that a lot of type 4 is also a type 3, but I wanted a puzzle that was only type 3. The puzzle you posted CAN also be solved by a type 4:)

I'm puzzled by the second puzzle in Havard's post


Thanks for finding that CathyW and thanks for pointing that out Wolfgang. It slipped past me:) I have edited the post and taken it out!

eclark:
In the explanation I wrote, I use that exact sudoku as an example of the Type 2. (and the Type 1 you talk of) Just look at that post!:)

Wolfgang:
That puzzle is great! just what I wanted. Thank you!:)

flip:
good point, but I have not yet come across a puzzle where you would get stuck if you did the Type 4 before the type 3. Maybe you can find me one?:)

thanks guys!

Havard
Havard
 
Posts: 378
Joined: 25 December 2005

Postby aeb » Thu Feb 02, 2006 11:53 pm

Havard wrote:In other words you can't have the two cells with "other" candidates across from each other!

Because otherwise it is of type 5?
aeb
 
Posts: 83
Joined: 29 January 2006

Next

Return to Advanced solving techniques