The New Sudoku Players' Forum

[CathyW]

In this post from the Help with Particular Puzzles forum there's some discussion about the validity of exclusions from colouring. Can anyone please help me to clarify whether my colouring efforts on the puzzle in question were valid or not, and if not why. Thanks.

[Carcul]

Hi CathyW.

Regarding coloring on "2" I cannot see where are the conjugate links that can be used to eliminate "2" from r2c7.
Regarding coloring on "3", same thing: only the elimination from r1c7 seems ok.
Regarding coloring on "4" and "5", I also cannot see the conjugate links.
Regarding coloring on "7", the eliminations r2c6,r2c7<>7 are correct but only using grouped coloring (grouped X-cycles) and not simple or multiple coloring (if I understand correctly the definitions of the various types of coloring). The eliminations that follow from r9c37 are also correct. The elimination from r5c8, however, seems obscure.
Regarding coloring on "8", I also cannot see how you can do the two eliminations.
The coloring on "9" is also incorrect: the correct elimination with colors is from r1c2 due to the conjugate links r3c3=r3c4 and r4c2=r4c4. Then, due to locked candidates on box 1 you are able to eliminate the "9" from r9c3.

Regards, Carcul

[ronk]

Can anyone please help me to clarify whether my colouring efforts on the puzzle in question were valid or not, and if not why.

Perhaps it would be more constructive for you to pick a specific example ... e.g., one where Carcul said your deduction is incorrect ... and explain why you think your deduction is valid.

Ron

[CathyW]

Thanks Carcul and Ron for your replies. I did try and explain further down the post on the help forum about the eliminations I'd made. Hanyou Hottie's post on the same referring to his message about colouring allowed me to think they were all OK:

Personally I find coloring to be much easier to spot, because all you have to look for are conjugate pairs. You should check out the post I made a while back where I try to explain the more advanced aspects of coloring, and how to use it easily:
http://forum.enjoysudoku.com/viewtopic.php?p=14558#p14558
as well as
http://forum.enjoysudoku.com/viewtopic.php?p=14695#p14695

but then doubts set in again after Tracy's further post.

I'm going round in circles - the eliminations based on the intersection cells of separate conjugate pairs of the same candidate are not "invalid" because Simple Sudoku would have told me otherwise. Yet I'm being informed there is no other logical reason to make some of the eliminations.

[TKiel]

CathyW,

I totally agree with you about colouring being easy to spot. I also think it's pretty useful, especially when one can find two different conjugate chains. I've been reading the thread about 'skyscraper' & 'two-string kite' and a lot of that looked just like colouring to me. (Not all though.) I've been trying like heck to spot a 'turbot fish' in a puzzle though I haven't yet, partly because I usually try to find conjugate chains before I look for something like that and I bet a lot of the time the colouring makes the elimination that a 'turbot fish' would make and thus I never see it. One of the problems with colouring is that is does rely on conjugate pairs and some puzzles never get to the point where there are enough conjugate pairs or they don't relate to each other in the necessary manner, so it definitely has it's limitations.

One thing from your post in this thread caught my attention but I'm not sure that I interpreted it in the same way as you meant it:

the eliminations based on the intersection cells of separate conjugate pairs of the same candidate are not "invalid" because Simple Sudoku would have told me otherwise.

Simple Sudoku only says an exclusion is invalid when you've excluded the candidate that is the 'answer' for that cell. It implies nothing about the validity of the argument used to make that exclusion. As I said in my other post, it does seem unlikely that none of your exclusions were ruled invalid, unless there is indeed a logic behind them but that in itself cannot be taken as proof that it was a logical exclusion.

Tracy

[CathyW]

Simple Sudoku only says an exclusion is invalid when you've excluded the candidate that is the 'answer' for that cell. It implies nothing about the validity of the argument used to make that exclusion.

Tracy - thank you! I think I was being blinded by being 'able' to make the exclusions without really understanding why they were 'possible'. I'm going back to first base on colouring and will edit the other post.