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 

Puzzle 11/05/05: ~ Advanced (BBDB?)

 
Post new topic   Reply to topic    dailysudoku.com Forum Index -> Puzzles by daj
View previous topic :: View next topic  
Author Message
daj95376



Joined: 23 Aug 2008
Posts: 3855

PostPosted: Thu May 05, 2011 5:14 pm    Post subject: Puzzle 11/05/05: ~ Advanced (BBDB?) Reply with quote

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

Play this puzzle online at the Daily Sudoku site
Back to top
View user's profile Send private message
peterj



Joined: 26 Mar 2010
Posts: 974
Location: London, UK

PostPosted: Sat May 07, 2011 6:21 pm    Post subject: Reply with quote

Two steps with an interesting "transport" on a kite...
Quote:
kite(5) (5)r9c5=r9c2 - r7c1=r2c1 {pause} - (5=8)r1c2 - (8=4)r4c2 - (4=5)r4c9 ; r2c5<>5, r4c5<>5
kite(8) c2, r6 ; r1c6<>8
Back to top
View user's profile Send private message
Marty R.



Joined: 12 Feb 2006
Posts: 5061
Location: Rochester, NY, USA

PostPosted: Sun May 08, 2011 4:19 am    Post subject: Reply with quote

Type 3 UR (12), using pseudo cell 45
W-Wing (58); r1c6<>5
XY-Chain; r2c5<>8 + transport; r1c2<>8
Back to top
View user's profile Send private message
tlanglet



Joined: 17 Oct 2007
Posts: 2461
Location: Northern California Foothills

PostPosted: Sun May 08, 2011 3:17 pm    Post subject: Reply with quote

I had some free time this morning and spent it working on this puzzle. I found many, many two step solutions but they all seemed to be using similar paths. Like Peter recently posted, I became determined to find something OTT to find a better solution: in this case, an almost skyscraper.

almost skyscraper (8)r1c26 with pincers r4c2|r6c6 and fin (8)r4c6; r6c1<>8=1
If skyscraper is true: r6c1<>8=1
If fin is true: (8)r4c6-(6)r4c6=r8c6-(6=1)r8c2-r7c1=(1-8)r6c1;

Ted Very Happy
Back to top
View user's profile Send private message
daj95376



Joined: 23 Aug 2008
Posts: 3855

PostPosted: Sun May 08, 2011 4:37 pm    Post subject: Reply with quote

I'm glad this puzzle generated responses more interesting than the (default) solution provided by my solver. I just now ran my solver to see what else it could find.

From my solver! There are two URs with a cell in common: <12> UR r35c23 and <15> UR r56c36. In addition, each UR has an internal strong link based on two non-UR candidates. Finally, each strong link involves the common cell.

Code:
 (8=4)r4c2 - r5c2 =UR= (5-2)r5c3 =UR= (8)r6c6  =>  r4c56,r6c1<>8
 +-----------------------------------------------------+
 |  3    58   9    |  458  7    458  |  1    6    2    |
 |  58   7    6    |  158  158  2    |  3    4    9    |
 |  4   *12  *12   |  6    9    3    |  7    5    8    |
 |-----------------+-----------------+-----------------|
 |  7    48   3    |  2    56-8 56-8 |  9    1    45   |
 |  6   *124 *125# |  9    3    15#  |  8    7    45   |
 |  1-8  9    15#  |  7    4    158# |  2    3    6    |
 |-----------------+-----------------+-----------------|
 |  15   3    4    |  158  158  9    |  6    2    7    |
 |  9    16   7    |  14   2    46   |  5    8    3    |
 |  2    56   8    |  3    56   7    |  4    9    1    |
 +-----------------------------------------------------+
 # 38 eliminations remain

Ironically, this solution uses the same candidate and pincers as Ted's.
Back to top
View user's profile Send private message
peterj



Joined: 26 Mar 2010
Posts: 974
Location: London, UK

PostPosted: Sun May 08, 2011 5:09 pm    Post subject: Reply with quote

daj95376 wrote:
From my solver!

Danny, that's a really nice move! And impressed that your solver is now able to find such things...
Back to top
View user's profile Send private message
daj95376



Joined: 23 Aug 2008
Posts: 3855

PostPosted: Sun May 08, 2011 5:32 pm    Post subject: Reply with quote

peterj wrote:
Danny, that's a really nice move! And impressed that your solver is now able to find such things...

Thanks Peter !!!

My solver only has a "basic" internal UR SIS capability right now. I'm debating how far I want to go with internal/external SIS. If I add more candidates to the UR logic, then it can easily become Kraken cell/unit logic. Something I've tried to avoid because I never know what's a "reasonable" distance to follow false streams.

I really wish that I could add ALS logic to my chain() routine. Right now, that appears to be a nightmare to incorporate, even though my data structures might support it.

Regards, Danny
Back to top
View user's profile Send private message
Marty R.



Joined: 12 Feb 2006
Posts: 5061
Location: Rochester, NY, USA

PostPosted: Sun May 08, 2011 5:40 pm    Post subject: Reply with quote

For those of us (maybe just me) who don't see those big pictures, I see a Hidden UR (15) with strong link on 5 in r6, setting r5c3<>5, thus reducing the 12 UR to a Type 1.
Back to top
View user's profile Send private message
ronk



Joined: 07 May 2006
Posts: 397

PostPosted: Sun May 08, 2011 6:27 pm    Post subject: Reply with quote

daj95376 wrote:
My solver only has a "basic" internal UR SIS capability right now. I'm debating how far I want to go with internal/external SIS. If I add more candidates to the UR logic, then it can easily become Kraken cell/unit logic. Something I've tried to avoid because I never know what's a "reasonable" distance to follow false streams.

I believe every exclusion that can be done with UR "mixed" SIS can also be done with "100% internal" SIS and "100% external" SIS. So the issue would be relative complexity of the three classes, not their exclusion possibilities.
Back to top
View user's profile Send private message
peterj



Joined: 26 Mar 2010
Posts: 974
Location: London, UK

PostPosted: Sun May 08, 2011 7:07 pm    Post subject: Reply with quote

ronk wrote:
I believe every exclusion that can be done with UR "mixed" SIS can also be done with "100% internal" SIS and "100% external" SIS. So the issue would be relative complexity of the three classes, not their exclusion possibilities.

That is certainly my experience. However, one of the routes (internal/external) is likely to involve ALS logic to make the same eliminations - which I understand Danny's solver doesn't support (yet!).
I think most human solvers start with internal SIS as this seems easier to understand (only evidence for this is propensity to use internal SIS on posts in this and other forums)
Back to top
View user's profile Send private message
daj95376



Joined: 23 Aug 2008
Posts: 3855

PostPosted: Sun May 08, 2011 8:03 pm    Post subject: Reply with quote

Ron, here's an example UR where at least one of four premises: r2c4=3, r2c6=4, r8c4=5, and/or r8c6=6 is true.

Potentially, only one premise might be true. This leads to a forcing chain scenario where three of the four streams are based on a false premise. Since a stream based on a false premise may never result in a true condition that's useful, I could end up with numerous long streams that are useless. Since each premise might have several streams possible, the number of permutations to check all four premises could result in an ugly search with little prospect of finding a common elimination.

Obviously, we're no longer dealing with AICs. Thus, I'm staying clear of a UR scenario involving more than two internal, non-UR values. I may reconsider it should I decide to add forcing networks more complicated than a SIN.

Code:
 +-----------------------------------------------+
 |   .   .   .   |   .   .   .   |   .   .   .   |
 |   .   .   .   |  123  .  124  |   .   .   .   |
 |   .   .   .   |   .   .   .   |   .   .   .   |
 |---------------+---------------+---------------|
 |   .   .   .   |   .   .   .   |   .   .   .   |
 |   .   .   .   |   .   .   .   |   .   .   .   |
 |   .   .   .   |   .   .   .   |   .   .   .   |
 |---------------+---------------+---------------|
 |   .   .   .   |   .   .   .   |   .   .   .   |
 |   .   .   .   |  125  .  126  |   .   .   .   |
 |   .   .   .   |   .   .   .   |   .   .   .   |
 +-----------------------------------------------+
Back to top
View user's profile Send private message
ronk



Joined: 07 May 2006
Posts: 397

PostPosted: Mon May 09, 2011 7:42 am    Post subject: Reply with quote

daj95376 wrote:
Obviously, we're no longer dealing with AICs. Thus, I'm staying clear of a UR scenario involving more than two internal, non-UR values.

I can understand that. Perhaps the general approach should be 1) identify an AUR, 2) find the minimal strong-inference-sets for it using internal and/or external DP busters, and then 3) search for a useful chain using a minimal SIS.
Back to top
View user's profile Send private message
daj95376



Joined: 23 Aug 2008
Posts: 3855

PostPosted: Mon May 09, 2011 3:31 pm    Post subject: Reply with quote

ronk wrote:
I can understand that. Perhaps the general approach should be 1) identify an AUR, 2) find the minimal strong-inference-sets for it using internal and/or external DP busters, and then 3) search for a useful chain using a minimal SIS.

Right now, my solver has the following UR capability:

    *) reduced UR-values detection

    *) dedicated routines for UR types 1-6

    *) general routine for URs using strong links on the UR values

    *) extension to general routine called "+ Naked_Single"

    *) extension to general routine called "prevented" -- for UR+3k

    *) addition of (two-value) internal UR strong link to chain routine

    *) addition of UR detection as a contradiction to (network) SIN routine

What's glaringly missing is utilizing any ALS structures.

Note: "reduced UR-values" occurs when one (or both) of the UR values aren't present in all four UR cells.

Code:
 worst-case example of reduced candidate detection -- 4x DPs possible
 +-----------------------------------+
 |  . 12  .  |  . 34  .  |  .  .  .  |
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |  . 34  .  |  . 12  .  |  .  .  .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |-----------+-----------+-----------|
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |  .  .  .  |  .  .  .  |  .  .  .  |
 |  .  .  .  |  .  .  .  |  .  .  .  |
 +-----------------------------------+
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    dailysudoku.com Forum Index -> Puzzles by daj 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