Rules. Another cute idea compromised by uniqueness.Hmph.
Heyawake 3: Compromised


Prasanna Seshadri said...

You could just have a 1 in R1C3 to make it unique, unless there is a theme I am missing beyond the break-in. Nice puzzle till that point.

Edderiofer said...

I believe it's unique even without the clue you mention, Prasanna. I think what Antonis is referring to is the possibility of exploiting uniqueness logic.

rob said...

I liked it, too (and found hard logic easier to find than uniqueness arguments). Beautiful break-in, and not much wrong with the puzzle that I can see. Were you hoping to have no numbered clues aside from the centre?

Δόλιχος said...

Thanks for all the attention, let me explain. Once you solve the central 5x5, the region with the 5 clue is ambiguous but guaranteed to block connectivity at the top even if you ignore everything else. I wanted to preserve that ambiguity till the remaining solution progressed clockwise to reach it from the left. Turned out I couldn't quite do it without growing the puzzle from the left / using non-rectangular regions / breaking the break-in / using a substandard left side.

Using a 1 in R1C3 instead of 2 in R1C4 of course works, and is arguably slightly more tricky. But it still short circuits my goal, and requires another clue to fix the top right, so I went spare.

I feel, and I'd welcome disagreement on the issue, that Heyawake without very dense numerical or border constraints are almost always going to be susceptible to uniqueness. I'm generally happy if uniqueness doesn't allow someone to finish in a fifth of the forward time - does this one break down completely some place I missed?