peek at solution solve puzzle
quality: difficulty:
solvability: moderate lookahead
Puzzle Description Suppressed:Click below to view spoilers
#1: Adam Nielson (monkeyboy) on Mar 11, 2009 [SPOILER]
Comment Suppressed:Click below to view spoilers#2: Web Paint-By-Number Robot (webpbn) on Mar 11, 2009
Found to require some guessing by jan.#3: Twillis (twillis) on Mar 12, 2009
I also ended up guessing on the green, but I may have just been too tired to make edge logic work.#4: Petra Lassen (Stjarna) on Mar 12, 2009
I used edge logic too... but I guessed where to start to count - and I guessed right.#5: Jan Wolter (jan) on Mar 12, 2009
I was unable to find any way to solve this with edge logic. I'd be interested in what edge you found that worked.#6: Adam Nielson (monkeyboy) on Mar 12, 2009 [HINT]
I started with the 1 in the 8th column, placing it in the top left corner and went down (I think).#7: Petra Lassen (stjarna) on Mar 12, 2009 [SPOILER]
Maybe that is more trial and error, I don't know. I don't really want to debate the semantics and terminology of edge logic vs. guessing, vs trial and error. I am probably wrong about calling what I did edge logic, then.
Comment Suppressed:Click below to view spoilers#8: Jane Doe (telly) on Mar 13, 2009
I did what Petra did and got it...but like Adam said, it might just be guessing, or whatever.#9: Gator (Gator) on Jul 30, 2009 [HINT]
I found this puzzle to fit it's title well...no purpose in doing it.
If you try to fill in R9C13, this would cause R5C10 and R7C10 to be filled in, which makes column 10 invalid. So R9C13 is a dot. Next consider the 3 clue in row 5. If you try to place it in columns 8 - 10, the rest of columns 8 and 10 will be dotted, causing the 3 clue in row 7 and the 2 clue in row 9 to fill in what is left which makes column 12 invalid. So R5C8 is a dot. The remainder will solve normally.#10: Jan Wolter (jan) on Jul 30, 2009 [HINT]
Hmmm... at this point we are solving just the green, so it's essentially a 6x6 puzzle. In your "next consider the 3 clue" step, you looked at row 10, columns 8 and 10, row 7 and 9, and then found a contradiction in column 12. That's an awful lot of look-ahead, involving half the rows and half the columns of the puzzle. I guess you can do it in you head, but only because the puzzle is so tiny.#11: Gator (Gator) on Jul 30, 2009 [HINT]
I think that goes a bit beyond the limit of what I'd call logical solving and drifts into trial and error search. It's a fuzzy line, but I'm going to let the "some guessing" ruling stand on this one.
Ingenious though.
I agree as you have to look one more step ahead than I would consider valid for contradiction logic. The size just makes it easier to do in your head. It was fun to work out this contradiction. :)#12: Kristen Vognild (kristen) on Jul 6, 2014
Perhaps this could be reclassified as Deep Lookahead?#13: Web Paint-By-Number Robot (webpbn) on Jul 30, 2014
Found to be solvable with deep lookahead by gator.#14: Gator (gator) on Jul 30, 2014
Agreed.#15: RB (rb2013) on Apr 27, 2020 [HINT]
Solution for the green part:#16: Web Paint-By-Number Robot (webpbn) on Sep 23, 2020
Check out Gator's puzzle 'Some call them skeleton' and ripkoops's puzzle 'Buckley' for some similar interesting logical reasoning.
Here, look at the sole 1 clue in C10 and the 3s in R5 and R7. One of the 3s have to go in C8-10, otherwise R5C10 and R7C10 (the contradiction ressembles to the one in #9) both have to be black, invalidating C10. It means that either R5C12 or R7C12 has to be filled, so the only 1 clue in C12 cannot be put in other rows. So R4,6,8-9 in C12 are dots.
The same logic can be applied to the 1 clue in R6 and the 3s in C8 and C10, so either R8C9 or R8C11 has to be black -> R8 C8,10,12-13 are dots (R8C12 is has already been dotted). LL gets an extra dot in R9C13.
Now consider the case when R9C10 is filled. Because of there's only an 1 clue in C10, R5C10 and R7C10 are dots. These imply that R5C13 and R7C13 have to be black (because of the 3 clues), which contradicts the condition in C13. Therefore, R9C10 is also a dot. LL solves the puzzle easily from this point.
These reasonings require 2 steps of lookahead, which may be appropriate for a moderate lookahead solution.
Found to be solvable with moderate lookahead by gator.#17: Gator (gator) on Sep 23, 2020 [HINT]
I agree that we would consider this more moderate lookahead. Looking at it again, it is clearly 2 moves ahead.
Show: Spoilers
You must register and log in to be able to participate in this discussion.