The New Sudoku Players' Forum

[StrmCkr]

Code: Select all

Yes, you did, thanks for that. Did you notice that it doesn't describe a loop and uses only three strong links (unlike the pattern you showed earlier)? So, I have much less of a problem accepting it as a "wing". It's also described as the inverse of XY-Wing, and I think there's a name for that... what was it again? Oh yes, L-Wing! :)

yes indeed it does break down to 2, L wings pretty nicely. with 1 elimination
instead of 8 digits in 1 cell, and "n" digit in "2" spots

to consider it use full think of the skyscraper technique 4 eliminations compared to the 2 x 2 minimal finned X-wings combined.

whats your thoughts on these

http://forum.enjoysudoku.com/post261728.html#p261728

odd balls i noticed when cross checking my w-wing code to another solver.

[SpAce]

yes indeed it does break down to 2, L wings pretty nicely. with 1 elimination
instead of 8 digits in 1 cell, and "n" digit in "2" spots

to consider it use full think of the skyscraper technique 4 eliminations compared to the 2 x 2 minimal finned X-wings combined.

That's a pretty good analogy. However, it seems to me that this is closer to Dual ERs in that you only need to find one half of the pattern (and its elimination) to get the guaranteed placement and thus the other elimination (among others) trivially. It's one step either way. Skyscraper and other Siamese Fishes don't guarantee that. If their sub-patterns are used separately you may have to use them both to get all eliminations (two steps), so it makes more sense to use the compound pattern in the first place to get them all at once (one step).

whats your thoughts on these

http://forum.enjoysudoku.com/post261728.html#p261728

odd balls i noticed when cross checking my w-wing code to another solver.

They don't look like any sort of W-Wings to me, not even extended or degenerate cases. The only visible W-Wing feature is the two identical bivalue cells, but nothing else seems to match, including and most critically the elimination logic. Pretty weird if XSudo sees them as such. Would you accept them as W-Wings?

[StrmCkr]

Two bivavle that see a bilocal strong ink or a grouped strong link.
Chain works the same
.rules dosent say any where that the bivavle.s can't be connected to the same cells on strong link.

[SpAce]

Two bivavle that see a bilocal strong ink or a grouped strong link.
Chain works the same
.rules dosent say any where that the bivavle.s can't be connected to the same cells on strong link.

Not quite sure what you're saying, but I just don't see any W-Wing there. To me it's obvious that the identical bivalues should be the chain ends in a W-Wing, and thus any eliminations must see both of them. That doesn't happen here with either the (49) or (59) pairs. It's not a VLV-chain, using my parlance. All I see is a VVL, or more simply just LL, and they don't count as W-Wings in my books. Or, maybe I just don't see it right.

On another note, what do you think of the unlisted 2-digit H-Wing (VLL) possibility I mentioned (as useless) earlier:

(x=y)A - yB = yC - yD = yE => -x BE, -y A [A and E see each other]

Like that it's of course unnecessary, because you get the same result with a simpler Turbot Fish (of some kind), but don't you agree that it's theoretically a valid H-Wing type (using my definition which only restricts the link count and types, but not digits or cells)? Also, if you replace the first bivalue with an ALS (again possible with my definitions), then you'd get this:

(x=vwy)AAA - yB = yC - yD = yE => -x E, -y A

In this case the Turbot would give you only the A-cell elimination but since it won't (necessarily) result in a placement there, you don't always get the E-cell elimination without a separate chain. The ALS-H-Wing would give you both at once. So maybe it's not totally useless conceptually?