03.09.2019, 11:15

Some design comments, spoilers for the puzzles if you haven't solved them

The 8 Standard puzzles are in the order in which I constructed them. Number 1 has a "Clockwise Countdown" theme, with each ring of clues consisting of 3210 and a clockwise arrow sequence in the outer two rings. With number 3 I wanted to construct a puzzle with very few clue cells. Puzzles 5 and 6 were inspired by the many Yajilin puzzles you can find on twitter by japanese authors. With puzzle 7, I tried to make the solving path diverse while sticking to one kind of clue only. Puzzle 8 can be solved by careful marking of empty cells, but in the top left corner it is faster to note that one entrance is in column 1 or 2, so column 4 cannot have 2 entrances, which immediately resolves the 2-up clue.

Regional Yajilin is one of my favourite variants, and I constructed some before for this year's Puzzle GP. The second one (with a "double digit" theme) is quite hard, but logically solvable, with a path that starts at the 5 and 4 and then proceeds anticlockwise. The end uses a difficult parity step, where you can use the fact that a loop always has an even number of cells to determine that there must be an even number of black cells, which helps to resolve the puzzle logically.

Myopia was fun to construct, I wanted to use a symmetrical layout with as few clues as possible. Getting a 10x10 puzzle with 6 symmetric clues to work was quite challenging.

I'm not sure if the star battle hybrid has been done before, but I wouldn't be surprised if someone has thought of it. The second one was challenging to make, because you can really only place two different clues, and it is very easy to get either zero or too many solutions, or to need a lot of extra empty

clue cells to get it unique.

Total False is a variation that I've seen on gmpuzzles.com. It's quite easy to miss possible deductions or mistakes while making them, and in fact the second one was broken in an earlier version which I didn't realize until months after constructing...

Cipher Yajilin is in fact the first Yajilin type I constructed and published on the portal. The first one is an easy one, but I'm happy that this layout works.

I wanted to do something different with the Outside puzzles, so I used a global break-in for the first one which I hadn't seen in Yajilin before.

The second one has a rather unconventional path: An Outside puzzle with an empty top will always have at least two solutions because you can flip the puzzle along a vertical axis. The adjacent rows of threes have only two options, and the are mirror images of each other. The intended path is to realize that the 1 clue on the top is there to disambiguate the "flipped" solutions, and you can pick one of the options for the 3s by going with your gut feeling (aka "guessing"). You can then solve downwards using only the row clues, and see if the 1 clue works. If it does, you picked the right orientation. If it doesn't, you don't need to erase everything, simply place the 1 clue in column 8 and solve the "flipped" version! The answer key is chosen so that it is not affected by this.

The Castle Wall puzzles were fun to make, the rules may have been slightly confusing, so I gave them a little more points. I would like to make some more of these in the future.

Double Yajilin is another type from gmpuzzles. I hope the second "clueless" puzzle was a surprise for some of you.

Unequal lengths as a loop variation first appeared on a Puzzle GP round this year. I think it is an interesting rule that can change things quite dramatically. Getting it to work in Yajilin was a huge challenge, as you get contradictions all the time because the type is so constrained. I'm very happy with the two puzzles, but I don't think I'll make more of them in the near future, because the second one in particular really exhausted my patience. Luckily, tweaking the grid to 10x11 resolved the puzzle without using much extra clues, and I'm still amazed that this actually works (Also, I noted that the two puzzles use a 1 clue, a 2 clue, a 3 clue, and 4 empty clues, which was unintentional).

If you have comments or questions about the puzzles, you are welcome to post them here.

The 8 Standard puzzles are in the order in which I constructed them. Number 1 has a "Clockwise Countdown" theme, with each ring of clues consisting of 3210 and a clockwise arrow sequence in the outer two rings. With number 3 I wanted to construct a puzzle with very few clue cells. Puzzles 5 and 6 were inspired by the many Yajilin puzzles you can find on twitter by japanese authors. With puzzle 7, I tried to make the solving path diverse while sticking to one kind of clue only. Puzzle 8 can be solved by careful marking of empty cells, but in the top left corner it is faster to note that one entrance is in column 1 or 2, so column 4 cannot have 2 entrances, which immediately resolves the 2-up clue.

Regional Yajilin is one of my favourite variants, and I constructed some before for this year's Puzzle GP. The second one (with a "double digit" theme) is quite hard, but logically solvable, with a path that starts at the 5 and 4 and then proceeds anticlockwise. The end uses a difficult parity step, where you can use the fact that a loop always has an even number of cells to determine that there must be an even number of black cells, which helps to resolve the puzzle logically.

Myopia was fun to construct, I wanted to use a symmetrical layout with as few clues as possible. Getting a 10x10 puzzle with 6 symmetric clues to work was quite challenging.

I'm not sure if the star battle hybrid has been done before, but I wouldn't be surprised if someone has thought of it. The second one was challenging to make, because you can really only place two different clues, and it is very easy to get either zero or too many solutions, or to need a lot of extra empty

clue cells to get it unique.

Total False is a variation that I've seen on gmpuzzles.com. It's quite easy to miss possible deductions or mistakes while making them, and in fact the second one was broken in an earlier version which I didn't realize until months after constructing...

Cipher Yajilin is in fact the first Yajilin type I constructed and published on the portal. The first one is an easy one, but I'm happy that this layout works.

I wanted to do something different with the Outside puzzles, so I used a global break-in for the first one which I hadn't seen in Yajilin before.

The second one has a rather unconventional path: An Outside puzzle with an empty top will always have at least two solutions because you can flip the puzzle along a vertical axis. The adjacent rows of threes have only two options, and the are mirror images of each other. The intended path is to realize that the 1 clue on the top is there to disambiguate the "flipped" solutions, and you can pick one of the options for the 3s by going with your gut feeling (aka "guessing"). You can then solve downwards using only the row clues, and see if the 1 clue works. If it does, you picked the right orientation. If it doesn't, you don't need to erase everything, simply place the 1 clue in column 8 and solve the "flipped" version! The answer key is chosen so that it is not affected by this.

The Castle Wall puzzles were fun to make, the rules may have been slightly confusing, so I gave them a little more points. I would like to make some more of these in the future.

Double Yajilin is another type from gmpuzzles. I hope the second "clueless" puzzle was a surprise for some of you.

Unequal lengths as a loop variation first appeared on a Puzzle GP round this year. I think it is an interesting rule that can change things quite dramatically. Getting it to work in Yajilin was a huge challenge, as you get contradictions all the time because the type is so constrained. I'm very happy with the two puzzles, but I don't think I'll make more of them in the near future, because the second one in particular really exhausted my patience. Luckily, tweaking the grid to 10x11 resolved the puzzle without using much extra clues, and I'm still amazed that this actually works (Also, I noted that the two puzzles use a 1 clue, a 2 clue, a 3 clue, and 4 empty clues, which was unintentional).

If you have comments or questions about the puzzles, you are welcome to post them here.