Puzzle Description:

Hiya, this is my first attempt at making one of these puzzles. My SO loves them, so for her birthday, I made this one as a gift. I'm too new to be able to solve this myself, so thank you so much for your input!

#1: David Bouldin (dbouldin) on Oct 14, 2022

97% done and trying VERY hard to stick to "moderate" lookahead. Really liking the image so far!

Same situation here David.

Got to 98% :)
I have so little left that it seems like it should either be multiple solutions or we should be able to find a moderate lookahead...and the checker says unique, so I'm still trying :P

For a long time I was thinking I was looking at a close of Daffy Duck - until other pixels came into the right place, and now I see something entirely different

Very cool pic. I couldn't do it without guessing, but I ain't the sharpest tool in the shed, so take that with a few grains of salt.

okay so I finished it, and can't quite make out who they are. Ghost?

Found to be solvable with moderate lookahead by dbouldin.

Found my break-through...wish I had tracked it. I did a LOT of moderate lookahead, but nothing I'd consider deep. Might do a resolve while tracking depending on how the comments go. Lovely picture. I'm going to guess a self-portrait with s/o

When I was left with the lady's features, I employed some lip and nose logic, feeling pretty sure that the leftover 4s would go in those spots. Not exactly kosher, but it allowed me to finish. :)

I used what I call "contradiction logic" on the middle 4 in C20 and later the 2 in C14. That's where I say "what if it starts HERE" and solve until I reach a point where no matter what there's a red ball somewhere. Then I revert and put a dot/white square wherever HERE was. Do that until you can use regular line logic some more, rinse & repeat as needed.

Thank you! I really like this one!

After LL leaves only a few blocks of unsolved in the left face, (95% done)

EL 3 c14 = r22 white


If we assume the 4 in r20 were in c18-21 then:
> the 4 in c20 would have to be in r19-22,
> which would trigger the 2 in r22
= which would cause a conflict in either direction
SO the assumed 4 cannot be there and r20c21 must be white

LL

If we assume the 4 in c20 were in r19-22, then:
> it would cause the rest of r19 to be white
> forcing the 3 in c17 to go into r22
= causing a second block in r22 where there's only one clue
SO the assumed 4 cannot be there and c20r19-22 must be white

LL (98% done)

Arguably just-barely not deep lookahead:
Three-way logic on the 2 in r22: (it can only be in c18-19, c19-20, or c20-21)
> if it were in either c19-20 or c20-21, it would trigger the 4 in c20 and so c20r19-22 would be black
> if it were in c18-19, it would trigger the 2 in c19
..> which would trigger the 2 in r21
..> which would trigger the 4 in c20 and so c20r19-21 would have to be black.
= In all three cases c20r19-21 must be black

LL to finish

I'd like a better (closer to vanilla moderate lookahead) step than that last one to definitively call this moderate lookahead, though

Recreated to your 98% point (I think) and then ignored your clue and tried to find a path. I probably messed up and have too much done or something because this one move and then LL finished it off for me:

- no matter where the 2 in C14 goes, C15R12 is black

Probably shouldn't post this until I double-check, but it's late and I just want to put this out there.

Excellent first attempt. I liked this and it was hard!

You got it, David. I remember looking at the 2 earlier in the solve and it was too long of a chain. And then I forgot to look at it later when it was not. I assumed it still was, I guess.

I do that a lot! Happy to help!

Good job. Excellent puzzle.

Found to be solvable with moderate lookahead by infrapinklizzard.

It became votable again

That's not good! I haven't noticed that happening since I fixed that one bug that was causing it. I'll look into the database logs this evening to see if they give any clues about what happened.

Joe: I traced it through the database logs. On October 16th, David set the puzzle's Logically Solvable setting to Definitely Moderate Lookahead. Then on October 17th, Slavmi Losmi, the original author of the puzzle, edited the puzzle. I'm not sure that Slavmi changed anything about the puzzle (I don't think it could have been changed, or the software would have given it a new version number), but the action of saving the puzzle also resets the Logically Solvable setting back to unknown, which makes it votable again.

I'd say it's debatable whether this is a bug or a good idea. Ideally, if the puzzle content and size haven't changed, then the software should be smart enough not to erase this setting.

Anyway, I'm glad to have a clear explanation for why the setting got erased!

Maybe tying the resetting of the solvability to the changing of the version number would work?

I dug into the code. The code changes the version number when the author changes any square in the picture. In the case where someone blotted some clues but left the picture the same, the version number would not change, but I think the solvability could change then. So I added a line of code that says "if this edit did not change the clues, then don't change the Solvability setting." That should fix this bug.

Makes sense. Good catch Valerie!

Goto next topic

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