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 

Magic bullet

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



Joined: 09 Jan 2011
Posts: 24
Location: Devon, England

PostPosted: Sat Feb 26, 2011 3:08 pm    Post subject: Magic bullet Reply with quote

Menneske 2170801
Code:

+------------------+----------------------+----------------+
| 8      345  1    | 235    2345    9     | 6     235 7    |
| 9      345  2    | 3567   34567   356   | 1     35  8    |
| 356    7    36   | 8      1235    1235  | 234   9   245  |
+------------------+----------------------+----------------+
| 2367   1    5    | 236    236     4     | 23789 78  269  |
| 2346   239  3469 | 12356  8       7     | 23    235 1256 |
| 2367   8    367  | 9      12356   12356 | 237   4   1256 |
+------------------+----------------------+----------------+
| 12457  259  8    | 1257   12579   125   | 2479  6   3    |
| 123457 2359 3479 | 123567 1235679 12356 | 24789 78  249  |
| 237    6    379  | 4      2379    8     | 5     1   29   |
+------------------+----------------------+----------------+

Play this puzzle online at the Daily Sudoku site

This looks very difficult (as it should - it's a 'Super hard +', rating 906). But there is a magic bullet - easy to follow, hard to spot (for me anyway).

(I used two moves in the bottom right, and eventually found a third that starts there.)
Back to top
View user's profile Send private message
daj95376



Joined: 23 Aug 2008
Posts: 3854

PostPosted: Sun Feb 27, 2011 5:56 pm    Post subject: Reply with quote

Every puzzle should have a solution. I don't know about JV's "magic bullet", but ...

I don't recall my solver's network logic previously leading me to a 5-cell ALS.

Code:
 (4)r3c9 = r3c7|r8c9 - r7c7 = r7c1 - (4=5)r54693c1  =>  r3c9<>5
 +-----------------------------------------------------------------------------------------+
 |  8        345      1        |  235      2345     9        |  6        235      7        |
 |  9        345      2        |  3567     34567    356      |  1        35       8        |
 | e356      7        36       |  8        1235     1235     | b234      9       a24-5     |
 |-----------------------------+-----------------------------+-----------------------------|
 | e2367     1        5        |  236      236      4        |  23789    78       269      |
 | e2346     239      3469     |  12356    8        7        |  23       235      1256     |
 | e2367     8        367      |  9        12356    12356    |  237      4        1256     |
 |-----------------------------+-----------------------------+-----------------------------|
 | d12457    259      8        |  1257     12579    125      | c2479     6        3        |
 |  123457   2359     3479     |  123567   1235679  12356    |  24789    78      b249      |
 | e237      6        379      |  4        2379     8        |  5        1        29       |
 +-----------------------------------------------------------------------------------------+
 # 143 eliminations remain

Basics complete the puzzle.
Back to top
View user's profile Send private message
tlanglet



Joined: 17 Oct 2007
Posts: 2468
Location: Northern California Foothills

PostPosted: Sun Feb 27, 2011 6:12 pm    Post subject: Reply with quote

What a move Exclamation

I really doubt that a person would ever search for such an als but life is always got a surprise for us.

Ted
Back to top
View user's profile Send private message
peterj



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

PostPosted: Sun Feb 27, 2011 6:29 pm    Post subject: Reply with quote

Danny, nice!
I don't understand why you need to worry about r8c9?

Some might see it easier from the other angle...

Code:
(5)r3c1=hp(15)r78c1 - (4)r7c1=r7c7 - r3c7=r3c9
Back to top
View user's profile Send private message
daj95376



Joined: 23 Aug 2008
Posts: 3854

PostPosted: Sun Feb 27, 2011 8:17 pm    Post subject: Reply with quote

peterj wrote:
Danny, nice!
I don't understand why you need to worry about r8c9?

Some might see it easier from the other angle...

Code:
(5)r3c1=hp(15)r78c1 - (4)r7c1=r7c7 - r3c7=r3c9


I just missed the HP() when trying to unravel my solver's network.

As for r8c9, I just threw that in to show that an alternate path was available from r3c9 to r7c7.

Thanks for making the obvious obvious.

Regards, Danny
Back to top
View user's profile Send private message
JV



Joined: 09 Jan 2011
Posts: 24
Location: Devon, England

PostPosted: Sun Feb 27, 2011 8:22 pm    Post subject: Reply with quote

Thanks Danny & Peter. Amazing Eurekas! By 'magic bullet' I did indeed mean deleting the 5 in r3c9.

I don't know how other folk see things: I mean do you think in Eureka? (I do for an ordinary chain, but something like this is too complex for me.) I simply saw, as you no doubt did, that r8c9 = 4 => r7c1 = 4 => r8c1 = 1 => r3c1 = 5 => r3c9 <> 5.

Another test if you'd like one, how one expresses an APE in EUreka: there's one in r78c7. (I'd already removed the 7s in r47c7 because of the UR 78, but I don't think that makes any difference.)
Back to top
View user's profile Send private message
peterj



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

PostPosted: Sun Feb 27, 2011 9:43 pm    Post subject: Reply with quote

JV, I have never spent the time to get to grips with APE. Partly because I never see anyone use them (other than some computer solvers) and my understanding is that they can usually be represented by xyz+-wings and als-xz techniques...

There is an wxyz-wing (I think)... which is hard to represent in Eureka other than a kraken presentation...
Code:
APE/wxyz-wing(2479) r7c7 ; r8c7<>2
(2)r7c7
||
(9)r7c7 - (9=2)r9c9
||
(4)r7c7 - als(4=23)r35c7
||
(7)r7c7 - als(7=23)r56c7

In this case you can see it as an "almost naked quad" move also...
Code:
anq(2347=9)r3567 - (9=2)r9c9 ; r8c7<>2

I think if you consider all the possible combinations of values in the "aligned pairs" r7c7 and r8c7 you will find that all the ones with (2)r8c7 are not valid - I haven't tried this! Life is too short...
Back to top
View user's profile Send private message
Marty R.



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

PostPosted: Sun Feb 27, 2011 10:20 pm    Post subject: Reply with quote

Peter, three years ago we started discussing APE here and this post by Keith probably put an end to it:

"All,

I did a search on APE, and was surprised to find my footsteps all over the discussion.

http://www.sudoku.com/boards/viewtopic.php?t=3882

As I now recall, Ruud proposed that APE was a missing method from the usual arsenal.

I did an analysis that showed APE is another (less general) form of XY, XYZ, and WXYZ-wings. In other words, APE makes one exclusion at a time, XY.. wings make multiple exclusions.

APE is a monkey that you should not worry about.

I welcome any example of an APE exclusion that is not an XY... wing.

Keith"
Back to top
View user's profile Send private message
JV



Joined: 09 Jan 2011
Posts: 24
Location: Devon, England

PostPosted: Sun Feb 27, 2011 10:38 pm    Post subject: Reply with quote

Thanks Peter, but I think I'm sorry I asked!
Right-hand boxes with 7s gone (because of the UR78):
Code:

+---------------+
| 6    235 7    |
| 1    35  8    |
| 234  9   245  |
+---------------+
| 2389 78  269  |
| 23   235 1256 |
| 237  4   1256 |
+---------------+
| 2479 6   3    |
| 2489 78  249  |
| 5    1   29   |
+---------------+


Clearly, (7)r7c7 = (8)r8c7 . (It doesn't matter but they're in fact strictly conjugate , since their only other occurence in the box is as a naked pair.

If you have 7 in r7c7 then there's a locked set 234,23, 237, 7 in the column: no 2 or 4 in r8c7.
Or, r8c7= 8; no 2 or 4.

I know it's not very common, but it's also very easy. (The combination of 7 & 8 in box 9 hint's at the existence of an APE, since that's at least one pair disallowed.)
Back to top
View user's profile Send private message
keith



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

PostPosted: Sun Feb 27, 2011 10:50 pm    Post subject: Reply with quote

Here is the corrected link:

http://forum.enjoysudoku.com/aligned-pair-exclusion-ape-no-dj-t3882.html

Keith

Marty R. wrote:
Peter, three years ago we started discussing APE here and this post by Keith probably put an end to it:

"All,

I did a search on APE, and was surprised to find my footsteps all over the discussion.

http://www.sudoku.com/boards/viewtopic.php?t=3882

As I now recall, Ruud proposed that APE was a missing method from the usual arsenal.

I did an analysis that showed APE is another (less general) form of XY, XYZ, and WXYZ-wings. In other words, APE makes one exclusion at a time, XY.. wings make multiple exclusions.

APE is a monkey that you should not worry about.

I welcome any example of an APE exclusion that is not an XY... wing.

Keith"
Back to top
View user's profile Send private message
JV



Joined: 09 Jan 2011
Posts: 24
Location: Devon, England

PostPosted: Sun Feb 27, 2011 10:54 pm    Post subject: Reply with quote

Marty, I've just read your post. I'm sure you're right, that APEs are a subset of more general methods. (And as I've said they're probably not common.) But . . . they're easy for some of us to spot where more complex methods might not be.

Isn't it the case with many common methods that they're special cases, used because they're easy to find? We like skyscrapers, but they're just one special type of single digit chain. Or W-wings, or whatever.

In this puzzle I was scratching round for something to do. I saw the APE simply because I find them fairly easy to see. I don't have the ability to see in any other way the two deletions I made.
Back to top
View user's profile Send private message
JV



Joined: 09 Jan 2011
Posts: 24
Location: Devon, England

PostPosted: Sun Feb 27, 2011 11:02 pm    Post subject: Reply with quote

Now I've read Keith's post, including the link to the APE discussion, but that hasn't left me much wiser. I can see vaguely that the deletions I made are essentially because of an intersection of ALS in c7 and box 9. And if I were pressed I just might be able to formulate that. But so what? - I'd never have been able to see this in any other way.
Back to top
View user's profile Send private message
keith



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

PostPosted: Sun Feb 27, 2011 11:17 pm    Post subject: Reply with quote

JV wrote:
Now I've read Keith's post, including the link to the APE discussion, but that hasn't left me much wiser. I can see vaguely that the deletions I made are essentially because of an intersection of ALS in c7 and box 9. And if I were pressed I just might be able to formulate that. But so what? - I'd never have been able to see this in any other way.

JV, can you be more explicit about what you are calling an APE? Which cells, what eliminations?

Thank you,

Keith
Back to top
View user's profile Send private message
keith



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

PostPosted: Mon Feb 28, 2011 2:42 am    Post subject: Reply with quote

I have started an APE thread in the "Solving Techniques" forum. Please post further discussion and examples of Aligned Pair Exclusions there.

Keith
Back to top
View user's profile Send private message
Asellus



Joined: 05 Jun 2007
Posts: 865
Location: Sonoma County, CA, USA

PostPosted: Mon Feb 28, 2011 3:20 am    Post subject: Reply with quote

tlanglet wrote:
I really doubt that a person would ever search for such an als but life is always got a surprise for us.

I probably wouldn't have noticed its usefulness in this case, but it's not so hard to spot such an ALS. Whenever you have a conjugate pair within a house, the other unsolved cells of the house form an ALS. This should be evident. You have n candidates distributed in n cells. There exists one candidate that is limited to two of those cells. The other n-2 cells thus contain n-1 candidates and form an ALS. (I'm assuming that the grid has been properly reduced and that there are no hidden sets.)

In this case, c1 had 7 unsolved cells with conjugate <1>s in r78c1. The remaining 5 cells, r34569c1, form the ALS with the 6 remaining candidates, 234567.

peterj's wing is another example. It uses the 5-digit ALS (almost naked quad, which is just an ALS) formed by excluding the conjugate <8>s in c7. His notation is fine, though I would notate it as:
ALS[(2)r3567c7=(9)r7c7] - (9=2)r9c9 ; r8c7<>2
It's just another way to write the same thing.

ALS are definitely a powerful solving technique and very useful in AICs. Spotting them is the trick.
Back to top
View user's profile Send private message Visit poster's website
JV



Joined: 09 Jan 2011
Posts: 24
Location: Devon, England

PostPosted: Mon Feb 28, 2011 10:00 am    Post subject: Reply with quote

Thanks Asellus.

Keith, I do see that it's not necessary to use an APE, but that's how I saw it. If you write down the possible pairs for r78c7 you'll find that none of them includes a 2 or a 4 in r78c7, thus deleting them.

And I have to admit that Asellus is right, that it's not too diffcult to spot once you know to look. I suppose the question is how to know where to look for something.
Back to top
View user's profile Send private message
tlanglet



Joined: 17 Oct 2007
Posts: 2468
Location: Northern California Foothills

PostPosted: Mon Feb 28, 2011 2:34 pm    Post subject: Reply with quote

Asellus,

Once again you have pointed out another insight into the sudoku world. I, for one, have not considered the simple, obvious perspective for spotting an ALS that you presented.

Thanks...........

Ted
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