Puzzle Description:

japanese character for shadow

#1: Claudia (clau_bolson) on Apr 29, 2020 [HINT]

I had to guess the entire right part

when testing it i found that it was possible, though hard, to determine it through logic. it may have helped that i already knew what it would look like though.

I had just a few dots and a black on the right side before I was feeling like it was trial and error.
So I counted up the numbers on the rows for the 3 sections:
24
23
22
And I counted up the numbers on the top. Because the remaining clues in the rows are only 1 number each row with no 1s it seems like everything has to be a solid connected line, so either a / or a \. To make the column numbers add up to each of the 3 sections' rows numbers, it means each row of column numbers lines up with its corresponding row sections, with the right-most 1 having to be in the middle section.
So then I just made each of the 3 sections into individual puzzles, which all worked with line logic apart from the bottom one, which only worked because I again presume that it will not work unless it is either a / or a \, which allows me to put in enough dots to make it work with line logic.

Just extrapolating the "/ or \" nature of it seems like very deep lookahead to me but I could be wrong. Either way the difficulty difference between the left and right halves is very large.

Don't think it matters, kedo, watashi mo kono kanji mou wakattita.

Nice. Did have to guess twice.

I used Emily's observations, but I did find some details I think we need to check for a rigorous proof/non-guessing solution.

But her intuition gets us most of the way there & I've seen people use the summing-dots approach for when there's just one box & I wouldn't blame someone for making that assumption and proceeding. It's a fun/interesting enough challenge then.

Here here we need to make sure that we can't have 2 or 0 of any column from 18-25 in any of R3-8, R11-16, R19-25.

This can be verified with moderate lookahead, nothing too many steps ahead, but there are a few columns to check!

In R3-8C18-25 it's not hard to show with moderate lookahead that we can't have 2 lines from the same column in C18-25.

That means the maximum possible is 25 if you count the columns (including C26) and 24 if you count the rows. But if one column of 18-25 has nothing in R3-8 then you have a maximum of 23 even with C26 being 1.

So there's exactly one column and row of dots in the upper right box. Also, counting dots, C26 R3-8 must be dots so columns = rows = 24. This lets us fill in the upper square with line logic.

Then I chipped away at C18 in the middle square. R11-12-13 are dots by edge logic. Then I said, what if the 2's in C18 are in R19-25? This got a quick contradiction with the upper 2, as you get a case where the 3 in C19 creates a 1 in a row.

This places a black dot at C18R20. Line logic for the middle box.

Edge logic on R19 and R25 places dots in C22--the 3x3 boxes quickly give contradictions wherever the 3's are in R19 and R25.

Then edge logic makes C18R20 a dot. It it is not, then R23C21/24 are black dots, a contradiction.

Line logic.

Found to be solvable with moderate lookahead by blurglecruncheon.

Goto next topic

You must register and log in to be able to participate in this discussion.