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 

simple forcing loop/ xy-loop/multivalue x-wing

 
Post new topic   Reply to topic    dailysudoku.com Forum Index -> Solving techniques, and terminology
View previous topic :: View next topic  
Author Message
storm_norm



Joined: 18 Oct 2007
Posts: 1741

PostPosted: Wed Apr 30, 2008 6:28 am    Post subject: simple forcing loop/ xy-loop/multivalue x-wing Reply with quote

the simple forcing loop, A.K.A xy-loop




very very importantly... not all xy-loops are going to be nice and rectangular like the one in the image above. xy-loops can take on different shapes and involve many cells. I like to call them "pools" or "ponds" since most big patterns involved fish names. Smile

in the image is a puzzle that was discussed in the forum.

Normally, the circled candidates would represent your average, run of the mill, xy-chain. However, this particular xy-chain has a potency for damage beyond what a normal xy-chain can do.

#1... a particular xy-chain which has its pincer ends sharing the same house can be considered a simple forcing loop, xy-loop.

#2... the green arrows show that the direction of the chain is indeed valid in both directions. what this means is that regardless of the cell you start in and which ever candidate you substitute in for that cell, the loop stays true in either direction you travel. keep this in mind...

#3...Imagine this as a multi-sided polygon. each two cells represent a "side" of the polygon. the row, column, or box that the "side" is in is where the eliminations take place. the eliminations come from the candidate that the "side" has in common.

in other words, each side contains two cells. Those two cells automatically become pincer cells (using the rules of a xy-chain). The cells that form the "side" of the loop share one candidate and can therefore eliminate any other candidates they see.
(this will sound sort of like describing a xy-chain)

example...
consider the cells, r6c2{8,9} and r6c4 {4,8}, these cells share the number 8. the rules of a xy-chain says that the 8 either goes in r6c2 or r6c4. since the 8 has to go in either of the two ends of the chain, you can eliminate 8 from r6c3.
This applies to all sides of the loop and makes eliminations as shown by the red boxes in the image.
it might be best to get a handle on what a xy-chain is first.
---------------------

this next example is alternative logic to the forcing loop. BUT!! it is highly specialized. how specialized you ask??

#1... it can only be a 4 cell xy-loop
#2... it only involves eliminations on two sides of the xy-loop
#3... the "roof" and "floor" candidates must be conjugate to the row, column, or box they reside in.
I'd say that is pretty special.



now before I go any further, I have to site the source of my research.
http://www.scanraid.com/Multivalue_X_Wing_Strategy
in the link is two examples, and both are exactly 4 cell xy-loops.

in the image above is an example of a simple forcing loop. it eliminates the 4 in r6c3.

this particular and very specialized pattern is what Andrew Stuart calls a multivalue x-wing. consider rows 3 and 5... candidates 5 and 8 are conjugate to those rows respectively. focus on column 5. what if both the 5 and the 8 are true? this would force 4 into two cells in column 3. Can't have that.
It also elminates the 7's in col 5. can't have that either.
this forces a 4 to exist either in r3c3 or r5c3 and eliminates the 4 in r6c3

edit: to change some of the wording. thank you to Steve R who has suggested changing some of the wording in the second example.


Last edited by storm_norm on Wed Apr 30, 2008 10:46 pm; edited 1 time in total
Back to top
View user's profile Send private message
storm_norm



Joined: 18 Oct 2007
Posts: 1741

PostPosted: Wed Apr 30, 2008 7:40 pm    Post subject: Reply with quote

I welcome and encourage any thoughts on how to clean up the wording.
I probably tried a little too hard in making the connection between multivalue x-wings and xy-loops and made some wording errors that, I hope, can be stated clear enough for the readers of the forum to understand.
Back to top
View user's profile Send private message
Marty R.



Joined: 12 Feb 2006
Posts: 5064
Location: Rochester, NY, USA

PostPosted: Thu May 01, 2008 12:21 am    Post subject: Reply with quote

Do I understand correctly that those eliminations of 8 in rows 68 only occur in an XY-Loop? I'm no theoretician, but if the answer is yes, then why? Why wouldn't they occur in an ordinary XY-Chain where the pincer ends aren't in the same house? For example, if the 34 were in r6c7 rather than r6c8, the 8s are still there.
Back to top
View user's profile Send private message
storm_norm



Joined: 18 Oct 2007
Posts: 1741

PostPosted: Thu May 01, 2008 12:56 am    Post subject: Reply with quote

Marty, hello there...

ignore for a moment that we are discussing xy-loops and form the xy-chain from r8c2 going clockwise around to r8c8. this chain says you can eliminate the 8's in row 8.
However, if you move the (3,4) cell to r6c7, this breaks that chain. now the (3,4) cell is no long connected to the (3,Cool cell in r8c8 and then doesn't eliminate the 8's in row 8.

the same can be said about the 8 in r6c3

except this time you are starting your xy-chain in r6c4 and going around connecting to r8c2... again, moving the (3,7) cell would break the ordinary xy-chain that this forms.

a side of a loop is comprised of two adjacent cells of the loop. any two adjacent cells of the loop can act as pincer cells.
Back to top
View user's profile Send private message
Victor



Joined: 29 Sep 2005
Posts: 207
Location: NI

PostPosted: Thu May 01, 2008 11:02 am    Post subject: Reply with quote

Marty wrote:
Quote:
Why wouldn't they occur in an ordinary XY-Chain where the pincer ends aren't in the same house?

I too pondered this question when I first met XY-loops. My answer to myself was as follows:
Take an open chain (3,4)- (4,8) - (8,9) - (9,8) - (8,3). I.e. where the 3s aren't in the same house. Now ordinary XY-chain logic essentially says: "If either 3 is NOT true, then the other must BE true." What it DOESN'T say anything about is the possibility that BOTH 3s are true - and that wouldn't tell us anything about the intermediate numbers.
However, in an XY-loop, this doesn't apply: no two successive numbers can both be true. So we can guarantee that all round the loop one is true & one is false, which kills any other occurrence of that number in that house. (I particularly like it when successive cells occur diagonally within a box, so that you can do an elimination(s) within the box.)

(Perhaps another way to look at it is to say that an XY-loop is a locked set: find one number and you've found them all. But finding one number in an ordinary chain doesn't solve them all. E.g. if you found the first (3,4) cell to be 3, well, that doesn't say that the next is 4.)
Back to top
View user's profile Send private message
Marty R.



Joined: 12 Feb 2006
Posts: 5064
Location: Rochester, NY, USA

PostPosted: Thu May 01, 2008 3:36 pm    Post subject: Reply with quote

I guess I just assumed the 8s would still be true since the XY-Chain is still valid if one of the pincer cells is moved, but after looking at Norm's answer and studying it, I realize I was wrong.
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    dailysudoku.com Forum Index -> Solving techniques, and terminology 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