dailysudoku.com

sdq_pete · Joined: 30 Apr 2007 Posts: 119 Location: Rotterdam, NL

The X-wing on 7's leaves a triplet in C5 followed by "XY 178 R3C9".

Peter

Marty R. · Posted: Thu Jan 17, 2008 6:04 pm Post subject:

This puzzle offered an excellent opportunity for a technique I learned just recently and which I had used just once before now, namely, Aligned Pair Exclusion (APE).

cgordon · Joined: 04 May 2007 Posts: 769 Location: ontario, canada

That Ape thing is cool. Sure beats looking for xy-wings. I'd have difficulty spotting the xy-wing if there were only 3 unsolved cells left. I used the Forum Offered method this morning. So let's Go Ape eh - A pox on wings.

storm_norm · Joined: 18 Oct 2007 Posts: 1741

put me down for the xy-wing {1,7,8} r2c8, r3c9, r3c4

takes out the 1's in r2c45

tlanglet · Posted: Thu Jan 17, 2008 11:16 pm Post subject:

eddieg · Joined: 12 Jan 2006 Posts: 47 Location: San Diego, CA USA

I enjoyed today's puzzle. I was stuck too until I found the same chain you found. My problem, is once I find these types of moves, and make the first elimination, I forget to see if there is a second or third elimination based on the same move.

Marty R. · Posted: Fri Jan 18, 2008 1:32 am Post subject:

cgordon · Joined: 04 May 2007 Posts: 769 Location: ontario, canada

That APE method reminds me a bit of the Sue de Coq method - which I actually learned a year ago - but have never since found an opportunity to apply. Sue de Coq is about finding two cells in a box with 5 numbers that see two pairs with 4 of those numbers. Thus - by relatively easy logic - the 5th or "unpaired" number is included in one of the cells.

Just thought I'd mention that.

(edited)

Asellus · Posted: Fri Jan 18, 2008 10:46 pm Post subject:

Craig,

Your description doesn't match my understanding of Sue de Coq, but does seem to describe a sort of APE.

As I understand it, Sue de Coq involves N digits in N cells arranged in such a way that they form various locked sets amongst various houses. This produces eliminations in cells outside of those N cells. It is similar to, say, a naked triple in a single house in that regard. (I.e. three digits in three cells that eliminate those digits from other cells in that house.) And, it is not limited to 5 digits. I've seen Sue de Coq situations involving 7 or 8 digits.

I find that Sue de Coq situations are difficult to spot and so have only rarely used the technique myself.

The two methods are closely related in that they both employ locked set logic.

cgordon · Joined: 04 May 2007 Posts: 769 Location: ontario, canada

Hi Asellus: I gotta have pictures. I don't think I described the Sue de Coq correctly - but my graphic understanding is shown below.

The three cells in Box1 C2 must contain three nos.of 12389 - but can't contain a 1 AND 9 or a 2 AND 3 - otherwise the linked pairs 19 and 23 disappear. So the three cells must contain an 8. So we can remove any linked 8 from Box 1 or C2.

Hope I got that right. Anyway I have looked for months for a chance to apply this method - but as you say they are very rare. Pity, I could have looked like an expert.

Asellus · Posted: Sat Jan 19, 2008 7:18 am Post subject:

Craig,

Yes, it was your description that wasn't doing the thing justice. What you have illustrated is Sue de Coq. There are 5 cells involved: r1239c2 and r3c3. And, these 5 cells contain exactly 5 digits: {12389}. And finally, they are arranged properly: {19} is restricted to c2 in the 5 cells; {23} is restricted to b1 in the 5 cells; and the 2 <8>s are forced to be strongly linked. And, there are more eliminations possible than you are saying (this is the potentially powerful aspect of Sue de Coq): <1> and <9> are removed from r45678c2; <2> and <3> are removed from r123c1 and r12c3; and <8>s are removed from r123c1, r12c3 and r45678c2. (In the particular example you saw perhaps only <8>s were present for removal. However, without knowing anything else about the grid, all of those eliminations are possible.)