Puzzle Description:

Lighthouse very roughly based on the March picture in my 2007 calendar

#1: Zapman (Zapman) on Nov 4, 2007

Am I missing something here?

Can this one actually even be started just by logical thinking?

This puzzle is logically solvable, although I tried to make it a challenge. The key to starting is the 25 row towards the bottom. If you start with it positioned far right, then you can start narrowing down this row's position from the row of 4 1's above it and the row of greater than 1's below it.

I removed one block since your attempt to solve it that makes the start even more difficult - I thought it would make it unsolvable, but was still able to work it out logically.

The puzzle is fairly easy to guess, so I'll leave it up to everyone whether to sweat it out or go for the quick solution. This will probably fall into Jan's "trial and error" category, but hopefully solvers who like a challenge will enjoy it.

Well, I was able to confirm that it does have only one possible solution, but though the approach you suggest does get me some distance into the puzzle, I had to do some pretty deep trial and error, investigating possibilities pretty far along before finding a contradiction so I could "undo" or "revert" my way back up again. I'm afraid I would have to classify it as a trial and error puzzle, though I'm not necessarily the cleverest solver on the block.

So, on the whole, I guess the puzzle does exactly what you want it to do, and I recognize that getting it all to work out and still have a decent picture at the end must have taken a lot of work. Pretty impressive.

As you say, it is pretty easy to find the solution by guessing.

I've just discovered that Im really good at logical guessing :)

A real challenge !
First time (months ago) i did it by guessing.
Today , I began by the 25 row as Robert hints, so that I could do all the bottom lines.
Then I went for the top which can be solve without too much trial and error, then the sides goes smoothly downard.

Thanks for this small challenge ;)

I really don't understand why this one is marked "may required guessing"

This puzzle has come up on my random puzzle click hundreds of times. Today, I decided to dive in. May not REQUIRE guessing, but it sure does help! :D

This puzzle is one of the ones that I used as a sample puzzle in my Survey of Paint-by-Number Puzzle Solvers as an example of a really difficult puzzle that still has a unique solution. It stumps about half of the better solvers, so Robert really succeeded at his goal of designing a really hard puzzle.

I don't think Robert comes around here much these days, but if he does, I wonder if he would be willing to give permission to allow this puzzle to be freely redistributed as part of a standard test set for developers of new solving programs.

Jan - feel free to use anything I've done for whatever you need. I don't design puzzles for profit, just a change of pace from solving, and have no problem with whatever you want to do with them.

Thanks

I was able to get R42C38-R42C50 as dots using what Robert described above. This led to some line logic.

Next I did edge logic starting at R45C19. There are many cells in this area that will make row 41 invalid. R44C19-R45C22 were made dots by this. More line logic.

Next I looked at the 6 clue in row 43 starting at the left. R43C1-R43C7 can be made dots as a 1 clue would be forced above or below this row which would be invalid.

The same technique can be done on row 44 to make R44C1-R44C3 and R44C47-R44C50 dots.

Now you can also make R44C30-R44C40 dots because no matter how those cells would be filled in, they would make either above or below the row invalid. More line logic.

Using the same logic on row 43, R43C30-R43C36 are made dots. More line logic.

Next I looked at the 11 clue in row 12. If you try to start it at R12C18, this would force 5 cells total either above or below this row, but only 4 cells can be filled in. So R12C18 is a dot. I then used edge logic on the 7 clue in column 18 to dot R2C18-R19C18. The same logic can be done on column 19 and 20 to make R2C19-R14C19 and R2C20-R5C20 dots.

I'm not sure where else to look at this point. I can make more dots by looking 4 moves ahead, but that would not be considered logically solvable.

OK, here is some more. If you assume that both the 6 and 5 clues in row 43 are on the right side, then R43C38-R43C42 would have to be black. Looking one move ahead, this causes row 44 to be invalid. So we know the 6 clue has to be on the left side. This lets us complete the bottom 8 rows and a lot of the columns now.

Now we can do some internal edge logic with the 11 clue in row 12. If it extends to the right, then the two 6 clues will cause row 13 to be invalid. So R12C33 and R12C34 must be dots. The rest solves with line logic.

Found to be logically solvable by gator.

difficult to start, but obvious when you read the title that it will be a light bulb or a light house. I started at the top down because it is easier to find the spots where the two & three spots are at the very top.

After determining the definite blanks on line 4 from the top, it was obvious the top was symmetrical and the rest was easy.

Not fun.

i apologize to anyone else i gave a 1 star review to. this is at the bottom - total waste of time.

Hey, it's karl!

Since I solved this puzzle so long ago and this may have been before Deep Lookahead was an option, I verified that this is Moderate Lookahead. Rather difficult to find the ways forward, but still within the bounds. I also feel like I am not nearly as intelligent now as I was back then. :(

Goto next topic

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