Showing posts with label variant. Show all posts
Showing posts with label variant. Show all posts

Monday, January 27, 2025

Puzzles #738 - #746 - Santa's Choco Banana

For the Secret Solver puzzle exchange on Puzzle'rs Club this past December, I got assigned Menderbug, who had previously written for me back in 2022. My first idea was to do something with Rail Pool, but before I started working on that idea I had a thought - what if I took the same idea that was made for me - a series of Symmetry Area puzzles on increasingly less regular grids - and applied that to Menderbug's signup comment of  "I like Choco Banana"? Once I had that idea, I had to go for it.

For all of these puzzles, all shaded areas must be convex and all unshaded areas must be concave. The entire pack can be found here.

My original plan was to make a small and large puzzle for each grid shape, and so I tried to focus on interesting logic for the smaller ones and that + a good layout for the large ones. Time pressure ended up pushing back some parts of this plan but the modular nature of the plan meant that I could easily scale back and have something that felt complete.


Menderbug is also the person who has pioneered triangular Choco Banana among the most, including making a few for GAPP. I definitely felt some pressure to try to deliver something new and interesting here, and while I'm not sure I succeeded I'm quite happy with what I ended up making.

Tetrakis Choco Banana has also been done before, though only with rectangle/not rectangle. I made so many mistaken deductions while trying to make these two puzzles... it's a weird geometry to get to grips with.


The last 3 puzzles are where the rules really start to fall apart a bit, as some of these geometries are really only good for a one-off puzzle. Rhombi worked okay but were prone to paired cell ambiguities, Snub Square actually worked very well and my biggest regret is only having time to make one of those, while Penrose... well, it's much like Rhombi except weirder. Definitely a one-off but a one-off worth exploring.

Overall I'm quite pleased with this set, and while Menderbug immediately thanked me specifically for it (not so secret) I knew that I was revealing my identity with the plan. He also said "I was going to say which puzzle/tiling I liked best, but I don't know, there was something really cool in each of them." after solving, which is very high praise indeed.

Sunday, January 26, 2025

Puzzles #735 - #737 - We Are So Back (Double Back Variants)

Continuing my backlog posting of old puzzles I haven't shared, here are the three puzzles I made for Logic Showcase 66. The prompt was to make a puzzle with a Double Back rule, and I took that prompt in three different directions.

The first puzzle I made was a Norinori where the unshaded cells had to be solved as a Double Back. In other words, draw a loop passing through all but 2 cells in every region, that enters and exits every region exactly twice. Cells unused by the loop should be shaded, and each shaded cell should have exactly 1 shaded neighbor. I had a lot of trouble finessing some parts of this to be unique near the end, which was a bit annoying given how nicely the start worked out.

The second puzzle I made was an Aquapelago where each region has to have exactly two unshaded components. Originally I dabbled with a fully clueless grid for this but decided a few clues would be an important addition. The first version of this puzzle was also broken, unfortunately, but I was able to fix it.

Finally, I knew I wanted to make a proper loop puzzle - not just a Yajilin-style mashup. So I combined Double Back with All or Nothing - solve as an All or Nothing where every region is visited either 0 or 2 times.

Tuesday, January 21, 2025

Puzzles #702 - #734 - Mini Threes Pack

So I've still had a lot of things going on in real life that have generally prevented me from having any time to work on puzzles, including having almost my entire holiday vacation suddenly taken up. Even as I write this I'm recovering from being sick, but being too sick to work but well enough to actually do things is somehow the sweet spot of being able to assemble things I've already previously made, as long as it was just assembly.

A few months ago I made a single puzzle using only a couple threes, and that set me down the path of trying to make a varied, interesting set of 3-only puzzles. Yesterday I put them all in a PDF and, fittingly (but annoyingly) took 3 tries to get everything in I wanted (forgot links once - there are links now!) and had 3 non-unique puzzles (all fixed now - I originally intended to re-test everything, but I was rushed and still recovering and that's a lesson learned for the future). So yeah. Here's a big pack of puzzles and hopefully the final version, thanks to CJK for informing me of the problems with the first versions. I think in the future I'll just use a post like this as a hub for any packs I put out - that way I can at least fix in-place easily.

Mini Threes Pack (PDF)


Puzzle links follow, I will not be adding 33 images to this post nor will I be compositing them all into one image to include. You'll have to check the PDF.

Nurikabe      Canal View     Shugaku      Snake Egg
Cave      Akichiwake      Nurimisaki      New Tren
Kurotto      Mochinyoro      Choco Banana      Shakashaka
Slitherlink      Vertex Slitherlink      Hotaru Beam      Hashi
Fourcells      Nawabari      Aho      Compass
Fillomino (Borders)      Nanro      Sukoro      San-Anko

Thursday, October 3, 2024

Puzzle #691 - Variant Masyu Day!

It's Masyu da- I already made this joke. About two and a half months ago, a couple people in Puzzlers Club were making meme Masyu puzzles with the variant rule that circles didn't have to be visited, but any circle that was visited had to be satisfied. These were memes because the solution would frequently avoid touching any circles, sometimes even just being a simple 2x2.

So I made a much harder puzzle to show that this variant did not have to just be a meme. It was a bit of a struggle, though, since I decided to start making this on my phone and quickly regretted that decision.

Wednesday, September 25, 2024

Puzzles #673 - #678 - One to Eight Collection

So, I've been sitting on this one for a while. As I write this, entries are open for Logic Showcase 66. I constructed this set of 6 puzzles for Logic Showcase 56, with the task of constructing 1-6 puzzles with 37 or fewer cells. I immediately thought of revisiting dividing a 6x6 grid into eight regions of sizes 1 to 8 each. Some of these clue types were obvious - Palisade and Checkered Fillomino - some took a bit more thinking - Symmarea and Kropki Pairs - and some were much more experimental - Thermo and Difference. If you only do one of these, do the Thermo, it's absolutely my favorite. The Symmarea was the biggest struggle to construct - turns out even finding a solution with rotationally symmetric 1-8 regions is hard. The most disappointing is the Difference since I needed the "cheater" even clues to avoid killing the puzzle extremely prematurely.

Rules for the puzzle are in the first link, and reproduced at the bottom of this post.

Divide each grid into 8 regions, each having exactly one region of size 1 to 8 each. Each grid has its own set of rules.
Checkered Fillomino: Numbers indicate the size of the region they are contained in. Regions must be two-colorable.
Palisade: Numbers indicate the number of edges drawn around the cell containing the clue. The edge of the grid counts as a drawn edge.
Symmarea: Numbers indicate the size of the region they are contained in. All regions must be 180 degree rotationally symmetric.
Thermo: Starting from the bulb, each thermometer cell must be part of a region with a strictly larger area than the previous.
Kropki Pairs: Regions separated by a white dot must have areas that differ by exactly 1. Regions separated by a black dot must have areas with a ratio of exactly 2 - one is twice as large as the other.
Difference: Regions separated by a gray dot must have areas that differ by exactly the number on the dot. An E stands for an unknown, positive even number.

Monday, September 23, 2024

Puzzles #668, #669 & #670 - 24HPC Zodiac (Little Killer Latin Square, Antiknight Tomtom)

I also contributed several puzzles to the Puzzlers Club rounds at the 2024 24 Hour Puzzle Championship this year. We authored two rounds, each with 24 puzzles. The first round had one puzzle themed on each Greek letter, while the second round was a double zodiac (signs and animals).

The other half of this round was themed on the Chinese zodiac animals. Rat went unclaimed for a long, long time, so I elected to make a puzzle using the Little Killer sudoku variant rule. Rather than make a sudoku and deal with regions and larger grids or irregular regions, I went with just a Latin Square. Here's a quick example puzzle I also made for the instructions booklet: the rules are pretty straightforward. Place a number from 1 to N (where N is the side length of the grid) into every cell, so that there are no repeats in a row or column. Arrows outside the grid give the sum of all the numbers inside the grid that they point to.

And the actual competition puzzle, worth 20 points. I tried to make it really tricky, but ended up missing a case so the logic is really messy and honestly the cleanest way is just a bifurcation, unfortunately. It's still unique and still a puzzle I made, so here it is.

I also claimed the spot for Horse with an Anti-Knight Tomtom; like with the IPC Killer Doppelblock, I didn't know which genre I wanted to use beyond a latin square type, but Tomtom quickly proved its worth and potential for some crazy theming. I ran into so many dead ends from only a few cages in this construction and it was really painful, even with using f-puzzles to help speed through "easy" anti-knight steps. The final puzzle is actually unique without the 21 cage, but since it's worth 110 points with the extra clue... yeah I think it's necessary for a good puzzle. Reminder everyone, harder/minimal cluing is not always better.

Wednesday, September 18, 2024

Puzzles #658 & #659 - 24HPC Domino Kurodoko

I also contributed several puzzles to the Puzzlers Club rounds at the 2024 24 Hour Puzzle Championship this year. We authored two rounds, each with 24 puzzles. The first round had one puzzle themed on each Greek letter, while the second round was a double zodiac (signs and animals).

Very early on I grabbed Beta and made a variant Kurodoko where instead of shading single cells, you shade dominoes instead. This variant does result in a bit of struggle for uniquely specifying the orientation, and can also sometimes have long-term propagations that break the puzzle way in advance. I'm pretty happy with how this 40 point puzzle turned out. I also wrote an example for the type, since it wasn't a classic type.

Tuesday, September 17, 2024

Puzzles #655, #656 & #657 - IPC Extras (Yajilin, Masyu All Cells)

Like last year, I helped contribute to the Indian Puzzle Championship this year. If you're interested in seeing the rest of the set, including bonus puzzles, check the link for details on how.

I had a couple extra puzzles I made that weren't used for competition - two alternate Yajilin options I didn't offer (since I liked the one that was used to much for the slot), and an All Cells Masyu meant for round 2 - logical optimizers. It didn't fit the round theme that well and there was already a puzzle with a similar vibe, so instead you get it as just an All Cells Masyu puzzle. Yay!



Monday, September 16, 2024

Puzzles #652, #653 & #654 - IPC Round 3 (Killer Doppelblock, Clone Snake)

Like last year, I helped contribute to the Indian Puzzle Championship this year. If you're interested in seeing the rest of the set, including bonus puzzles, check the link for details on how.

Round 3 was themed on taking sudoku variants and applying them to non-sudoku puzzles. I put myself down for number placement, wanting to make a puzzle of a type that has arithmetic normally combined with killer cages. I thought about a few types before settling on Doppelblock. I also signed up for clone Snake since I knew the combination had potential from old, old Islands of Insight rule prototyping. In fact, now that that game isn't getting new content, I know a lot of these old unused puzzles will stay unused, so retain all the rights to them. I've gone through my archive and have about 80 puzzles in this category I think are good enough to share, so at some point I'll have to set up solving interfaces/write rules/share them.

Anyway, Killer Doppelblock! Standard Doppelblock rules apply. Additionally, all the numbers inside a cage add up to the number in the corner of the cage. This is a pretty cool combination, where the different ways of getting partial sums interact in some really cool ways. This was worth 65 points.

Clone Snake: Standard Snake rules apply, with given endpoints and an unknown length. Two cages of the same shape and orientation must have the same arrangement of snake and not snake cells. The first puzzle here is actually one of those old IOI prototype puzzles, since I added it to the organization sheet after a couple failed constructions, to at least have a puzzle in the slot. I put a note that this was from a couple years ago and I didn't think anyone at the IPC would have seen it before. A few hours later, I finally turned up a satisfactory second puzzle, and added that to the sheet to provide an option for which puzzle to use. Prasanna chose both, worth 30 and 50 points.

Saturday, February 17, 2024

Puzzles #636 & #637 - Monkey Madness (Compass Galaxies)

Before WPC 2023, I ran a Puzzle Getchu!, a contest themed on Ape Escape. I had intended to queue these puzzles for posting before WPC, but didn't have the time. Better late than never!

Monkey Madness is the final level of Ape Escape, taking place first in an amusement park, and ending up in a castle floating in space. The level has 24 monkeys - 11 in the first (bottom) half, and 13 in the second (top) half. It's also huge, taking up 20% of a speedrun, so for the contest puzzle I had to go big. 14x14 was where I started, but was actually too small, so I bumped it up to 16x16. Oh right, the genre choice: I really liked the "one X per region" rule from the Pentominous in So Many Bees, so I brought that back with monkeys instead of bees. It ended up at 40 points, which on finishing pace would be 16 minutes for this puzzle. Interestingly, Ken Endo got here with more time than that and did not solve in time, similar to his performance in Puzzle Zodiac - also arriving at the final puzzle with plenty of time and coming up short. My contests always seem to have a "boss puzzle" at the end that really kicks people around, but it's always a standout and this was no exception: in the feedback I got afterwards, this was the most mentioned puzzle by far.

Rules: Divide the grid into regions along cell boundaries so that each region contains exactly one monkey (shown as a monkey inside a hexagon) and exactly one other clue. A white circle (Spiral Galaxies clue) indicates that the region has 180 degree rotational symmetry centered on the clue. A number in an X (Compass) clue, if given, indicates the number of cells inside that region that are further than the clue in the direction of the number.

Oh right I also had to make an example puzzle. So I did - it's pretty uninspired, but meh. I needed an example and this was the last one I made, and I knocked out all the examples over 2 days to rush the contest out before WPC. But also, I didn't want to re-use ideas from the main puzzle!

Friday, February 16, 2024

Puzzles #634 & #635 - TV Tower (Multi-Skyscrapers)

Before WPC 2023, I ran a Puzzle Getchu!, a contest themed on Ape Escape. I had intended to queue these puzzles for posting before WPC, but didn't have the time. Better late than never!

The penultimate level, TV Tower, is a skyscraper. But I didn't just want to go with a big Skyscrapers puzzle since I felt an 8x8 or larger would be less interesting. I also didn't think I could pull it off well! So I added the multi rule: Circled cells act as clues for all adjacent grids and will have the same value for all grids.

The example had a lot to show and I'm kind of surprised this grid layout even has any solutions, let alone working with such a sparse antisymmetric cluing. I almost like the example more than the contest puzzle.

The 20 point contest puzzle had to be multiple stacked grids, of course. My first attempt was a pair of 7x7 grids, but after a few hours I found I had no way to resolve the middle. So I scrapped that start and briefly considered three 5x5 grids before discarding that idea - WSPC 2021 did a 3x3 array of these and I had nothing new to add. I then set up three 6x6 grids and thought it looked way too large, so I nerfed it down to a pair of 6x6 grids and went for it. I used the same break-in as the aborted 7x7 attempt since I'd already discovered it, and there was no point not using the things I learned from that failed attempt.

Monday, January 29, 2024

Puzzle #600 - Interconnected Secret

This year for Secret Solver I was assigned feadoor, someone I know is quite a good solver and constructor across a variety of puzzle types. I wanted to try to lean into as much of those types as I could, remembering the Skyscrapers and Sudoku from a GP round, Yajilin from Yajilin and Variations, and Heyawake from Heyawake Hike. The rest of the genres were picked to cooperate with the connections between grids, and lead to a final secret. I did a lot of planning before any actual construction here!

Anyway, here's all the rules. The short version is that standard rules apply for each individual puzzle, and that cages marked with the same letter must have the same pattern of shaded cells. Country Road requires all regions, and unused cells are considered shaded. Nanro (Signpost), the numbered cells are considered shaded. Skyscraper Sudoku is 1-6 in each row, column and region. Domino Yajilin counts the number of shaded cells.


I really should have taken my own advice and not made an interconnected puzzle - it's very stressful when something starts not working, or seems like it might not resolve. I also ended up doing most of the writing over 2 days, with a few scattered bits of progress before that point. I even brought a notebook with transcriptions of the starting points (determined links) for 3 puzzles to a relatives' for Christmas Eve, and finished those puzzles there!

To think my original idea involved having to figure out the cage matching as part of solving... that would have been impossible within the time I ended up having, and especially with the other constraints.