dailysudoku.com Forum Index dailysudoku.com
Discussion of Daily Sudoku puzzles
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

An interesting one

 
Post new topic   Reply to topic    dailysudoku.com Forum Index -> Other puzzles
View previous topic :: View next topic  
Author Message
keith



Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA

PostPosted: Sun Aug 27, 2006 3:52 pm    Post subject: An interesting one Reply with quote

This is today's Mepham puzzle:
Code:
Puzzle: MM082706
+-------+-------+-------+
| . 6 . | 8 . 7 | 3 4 . |
| . . . | 5 . . | 8 . . |
| . . . | . 4 . | . . . |
+-------+-------+-------+
| . 8 9 | . . . | 5 3 . |
| 3 . . | . 1 . | . . 2 |
| . 4 6 | . . . | 7 1 . |
+-------+-------+-------+
| . . . | . 8 . | . . . |
| . . 7 | . . 4 | . . . |
| . 2 8 | 6 . 3 | . 7 . |
+-------+-------+-------+

Keith
Back to top
View user's profile Send private message
Marty R.



Joined: 12 Feb 2006
Posts: 5770
Location: Rochester, NY, USA

PostPosted: Sun Aug 27, 2006 5:13 pm    Post subject: Reply with quote

Here's my take: by the time I finished writing in the candidates, 24 cells had been solved. A scan of rows, columns and boxes solved five more. There was a Type 4 rectangle. However, I spotted no more rectangles or basic stuff and no XY-Wings were apparent, so I turned to looking for strong links and noting them outside the grid.

There were no X-Wings to come out of this, but there was a strong link (fork? skyscraper?). This strong link solved a cell, which in turn solved the puzzle. It's rather unusual for me to see a strong link make such a powerful impact.

I like that you post some of these puzzles. They're a welcome change of pace from the same old puzzles that I do on a day-in, day-out basis.
Back to top
View user's profile Send private message
keith



Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA

PostPosted: Sun Aug 27, 2006 7:58 pm    Post subject: Exactly Reply with quote

Marty,

Quote:
It's rather unusual for me to see a strong link make such a powerful impact.


which is exactly why I posted the puzzle! (I was going to title it "One for Marty", but changed my mind.)

Anyway, after you find the UR, there is sort of a brick wall. Here is the position I came up with:
Code:
+----------------+----------------+----------------+
| 59   6    12   | 8    29   7    | 3    4    159  |
| 7    3    4    | 5    6    129  | 8    29   19   |
| 8    59   12   | 3    4    129  | 1269 2569 7    |
+----------------+----------------+----------------+
| 1    8    9    | 2    7    6    | 5    3    4    |
| 3    7    5    | 4    1    8    | 69   69   2    |
| 2    4    6    | 9    3    5    | 7    1    8    |
+----------------+----------------+----------------+
| 459  1    3    | 7    8    29   | 249  259  6    |
| 6    59   7    | 1    259  4    | 29   8    3    |
| 459  2    8    | 6    59   3    | 149  7    159  |
+----------------+----------------+----------------+

The UR on <69> is there in R35C78. It eliminates <9> from R3C78, which is not much help.

Like you (I think) I found the strong links on <5> to be interesting. In particular, the strong links in R37 line up in C7. So the "fork" takes out the possibility <5> in R8C2, and the puzzle is solved.

There is another neat thing. Look at C19. Almost an X-wing on <5>. Except for the candidate <5> in R7C1. I believe this is a "finned" X-wing. Anyway, here is my reasoning:

Either the X-wing is true, or R7C1 = <5>.

If the X-wing is true, R9C5 is not <5>.
If R7C1 = <5>, R9C9 = <5>, and R9C5 is not <5>.

Then, you have an X-wing in R19, and R7C1 is not <5>.

Good stuff!

(If you do not exploit this coloring on <5>, the puzzle seems to be quite difficult to solve.)

Keith
Back to top
View user's profile Send private message
Marty R.



Joined: 12 Feb 2006
Posts: 5770
Location: Rochester, NY, USA

PostPosted: Sun Aug 27, 2006 9:40 pm    Post subject: Reply with quote

Quote:
Like you (I think) I found the strong links on <5> to be interesting. In particular, the strong links in R37 line up in C7. So the "fork" takes out the possibility <5> in R8C2, and the puzzle is solved.


When I saw the fork, I immediately took out the "5" from r1c1, having not noticed that it could also solve r8c2. But, of course, the effect is the same.

Quote:
There is another neat thing. Look at C19. Almost an X-wing on <5>. Except for the candidate <5> in R7C1. I believe this is a "finned" X-wing. Anyway, here is my reasoning:

Either the X-wing is true, or R7C1 = <5>.

If the X-wing is true, R9C5 is not <5>.
If R7C1 = <5>, R9C9 = <5>, and R9C5 is not <5>.

Then, you have an X-wing in R19, and R7C1 is not <5>.


Naturally, I didn't get that far, since the fork precluded the need to look for strong links in columns. I just learned the finned X-Wing a few months ago from the guys on this forum.

However, my understanding of that technique is that it removes one or more candidates from the box in which the fin occurs, because those candidates would be gone regardless of whether the X-Wing existed or the fin prevented it. In this case, the finned X-Wing couldn't be used because there was no candidate in the box that would have been eliminated had the X-Wing actually existed.

What you did, as I saw it, was to solve r9c5 with a DIC. The finned X-Wing technique was unable to eliminate anything, but the fin planted the seed in your brain that said, "this might be a good place to start a DIC."

Whaddya think? Different techniques or just semantics?
Back to top
View user's profile Send private message
Myth Jellies



Joined: 27 Jun 2006
Posts: 64

PostPosted: Mon Aug 28, 2006 2:34 am    Post subject: Colored x-wing is very similar to a finned x-wing Reply with quote

keith wrote:
Anyway, after you find the UR, there is sort of a brick wall. Here is the position I came up with:
Code:
+----------------+----------------+----------------+
| 59   6    12   | 8    29   7    | 3    4    159  |
| 7    3    4    | 5    6    129  | 8    29   19   |
| 8    59   12   | 3    4    129  | 1269 2569 7    |
+----------------+----------------+----------------+
| 1    8    9    | 2    7    6    | 5    3    4    |
| 3    7    5    | 4    1    8    | 69   69   2    |
| 2    4    6    | 9    3    5    | 7    1    8    |
+----------------+----------------+----------------+
| 459  1    3    | 7    8    29   | 249  259  6    |
| 6    59   7    | 1    259  4    | 29   8    3    |
| 459  2    8    | 6    59   3    | 149  7    159  |
+----------------+----------------+----------------+

There is another neat thing. Look at C19. Almost an X-wing on <5>. Except for the candidate <5> in R7C1. I believe this is a "finned" X-wing. Anyway, here is my reasoning:

Either the X-wing is true, or R7C1 = <5>.

If the X-wing is true, R9C5 is not <5>.
If R7C1 = <5>, R9C9 = <5>, and R9C5 is not <5>.

Then, you have an X-wing in R19, and R7C1 is not <5>.

Good stuff!

(If you do not exploit this coloring on <5>, the puzzle seems to be quite difficult to solve.)

Keith


Code:
+----------------+----------------+----------------+
|X59   6    12   | 8    29   7    | 3    4   *159  |
| 7    3    4    | 5    6    129  | 8    29   19   |
| 8    59   12   | 3    4    129  | 1269 2569 7    |
+----------------+----------------+----------------+
| 1    8    9    | 2    7    6    | 5    3    4    |
| 3    7    5    | 4    1    8    | 69   69   2    |
| 2    4    6    | 9    3    5    | 7    1    8    |
+----------------+----------------+----------------+
|x459  1    3    | 7    8    29   | 249  259  6    |
| 6   a59   7    | 1   A259  4    | 29   8    3    |
|X459  2    8    | 6   -59   3    | 149  7   *159  |
+----------------+----------------+----------------+

I've actually called this a colored x-wing, which is very closely related to a finned x-wing. It uses grouped coloring or multicoloring as in your case.

X and x are conjugate colors, with X combining with the starred cells in c9 to force an x-wing. A and a are also conjugate colors, and since x and a exclude each other, we know that either X, A, or both must be true. In any case, r9c5 cannot be a five.

This was a case where regular multicoloring (or strong links) also works, but there are some other examples where colored x-wing interactions and box-box interactions work when simpler coloring methods don't.
Back to top
View user's profile Send private message
keith



Joined: 19 Sep 2005
Posts: 3355
Location: near Detroit, Michigan, USA

PostPosted: Mon Aug 28, 2006 6:15 am    Post subject: What do I think? Reply with quote

Quote:
Whaddya think? Different techniques or just semantics?

Marty,

Definitely: "Not just semantics."

I never understood David's DIC's, because I could not see any rationale for how to look for them. If you now tell me they are based on strong links, I will have to go back for remediation! Maybe dcb can weigh in here.

What I do think is that these strong links are fundamental, especially for paper & pencil solvers like you and me. I don't think we need more names for techniques, but the idea that we can overlay strong links on other ideas (here, an X-wing) can be very useful.

As we have previously seen, with the overlay of strong links on Unique Rectangles.

Best wishes,

Keith
Back to top
View user's profile Send private message
Marty R.



Joined: 12 Feb 2006
Posts: 5770
Location: Rochester, NY, USA

PostPosted: Mon Aug 28, 2006 3:48 pm    Post subject: Reply with quote

Quote:
I never understood David's DIC's, because I could not see any rationale for how to look for them. If you now tell me they are based on strong links, I will have to go back for remediation!


I'm certainly not telling you that. Since I registered here in February, you have been one of my valued mentors and I don't think I'm in a position to try and tell you anything, other than to relate my thoughts based on what I see and to ask questions hoping to learn something.

The only rationale I use for starting chains, if indeed it's even considered a rationale, is to eyeball the potential starting point to see if both possibilities can eliminate something, in which case I pursue it. I'd also be interested in hearing how David finds his DICs.
Back to top
View user's profile Send private message
AZ Matt



Joined: 03 Nov 2005
Posts: 63
Location: Hiding under my desk in Phoenix AZ USA

PostPosted: Mon Aug 28, 2006 6:12 pm    Post subject: DICs Reply with quote

I will let David speak for himself on whether DICs ALWAYS entail strong links, but I will tell you that they (at least) very often do, and this is how I would have "discovered" this solution.

In looking at the set of <5> , I would note that it is a great candidate for holding a solution. The <5> is only a candidate twice in rows 1, 3, 7, and 8; columns 2, 8, and 9; boxes 1, 3, 8, and 9.

Conversely, <5> is a candidate three times ONLY in row 1, column 9, and box 7. Quite literally, then, all roads point to box 7.

In my admittedly limited experience, I all but know that not all the fives are going to fit, and box 7 holds the key, and in short order I see that if r7c1 is a <5> then r1c1 is a <5> and, consequently and in any event, r5c9 and r2c8 can't be <5>.

This is where David comes in and points out that it's just two connected DICs and then set out the logical equation...

At least I think that's how it works.


Last edited by AZ Matt on Mon Aug 28, 2006 10:42 pm; edited 1 time in total
Back to top
View user's profile Send private message
TKiel



Joined: 22 Feb 2006
Posts: 292
Location: Kalamazoo, MI

PostPosted: Mon Aug 28, 2006 11:35 pm    Post subject: Reply with quote

Here is the grid after removing the 9 from r3c78 from the 69 UR as noted by Keith and the 5 from r7c1 by a 'turbot fish'.

Code:
  59   6    12   | 8    29   7    | 3    4    159  |
| 7    3    4    | 5    6    129  | 8    29   19   |
| 8    59   12   | 3    4    129  | 126  256  7    |
+----------------+----------------+----------------+
| 1    8    9    | 2    7    6    | 5    3    4    |
| 3    7    5    | 4    1    8    | 69   69   2    |
| 2    4    6    | 9    3    5    | 7    1    8    |
+----------------+----------------+----------------+
| 49   1    3    | 7    8    29   | 249  259  6    |
| 6    59   7    | 1    259  4    | 29   8    3    |
| 459  2    8    | 6    59   3    | 149  7    159  |


There is a potential 49 UR in r79c17, which is also an X-wing on 4. I think that means we can remove the 9 from r9c17.

Code:
 
 *--------------------------------------------------*
 | 59   6    12   | 8    29   7    | 3    4    159  |
 | 7    3    4    | 5    6    129  | 8    29   19   |
 | 8    59a  12   | 3    4    129A | 126  256  7    |
 |----------------+----------------+----------------|
 | 1    8    9    | 2    7    6    | 5    3    4    |
 | 3    7    5    | 4    1    8    | 69   69   2    |
 | 2    4    6    | 9    3    5    | 7    1    8    |
 |----------------+----------------+----------------|
 | 49a  1    3    | 7    8    29   | 249  259  6    |
 | 6    59A  7    | 1    259  4    | 29   8    3    |
 | 45   2    8    | 6    59   3    | 14   7    159  |
 *--------------------------------------------------*


Which leaves us with a simple colouring chain that means r7c6 must be a 2, which leaves nothing but singles.

Or we can ignore the 49 AUR and colour the chain on 5:

Code:
 
 *--------------------------------------------------*
 | 59A  6    12   | 8    29   7    | 3    4    159a |
 | 7    3    4    | 5    6    129  | 8    29   19   |
 | 8    59a  12   | 3    4    129  | 126  256A 7    |
 |----------------+----------------+----------------|
 | 1    8    9    | 2    7    6    | 5    3    4    |
 | 3    7    5    | 4    1    8    | 69   69   2    |
 | 2    4    6    | 9    3    5    | 7    1    8    |
 |----------------+----------------+----------------|
 | 49   1    3    | 7    8    29   | 249  259a 6    |
 | 6    59A  7    | 1    259a 4    | 29   8    3    |
 | 459a 2    8    | 6    59A  3    | 149  7    159A |
 *--------------------------------------------------*


which makes all the (a) cells 5.
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    dailysudoku.com Forum Index -> Other puzzles 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


Powered by phpBB © 2001, 2005 phpBB Group