It’s a path puzzle. You draw a path that enters the grid at the top and exits at the bottom. Along the way, there are 50 sections, each following a different ruleset. Consecutive sections overlap by a few rows, creating short hybrid sections in between.
I kicked off the puzzle with the first section at the top and then passed it on to the other authors. Each author could only see the section right before theirs (and the list of all rulesets so far, to avoid duplicates). I’m assuming you’re here to solve a puzzle though, so I will write more about the process in a separate post.
There’s a few options:
All of these include self-contained versions that come with all the rules and example puzzles. But if you want those separately, or want solvable links for the examples, please refer to:
| Rules (PDF) |
|---|
| A4 (2.5M) |
| Letter (2.5M) |
The printable versions include all the rules (and the Letter format one also had enough room for shape banks on the side) and are split into three sections per page. Each page ends in an overlap section that is included again at the top of the next page, so you can transfer progress over.
| Printable puzzle (PDF) |
|---|
| A4 (11.4M) |
| Letter (11.3M) |
As mentioned above, the Penpa+ version is split into ten parts, made up of five sections each. You’ll be able to see a greyed out version of an additional section above and below each part. In general, answer check should pop up once you’ve determined the entire path before the greyed out section. This part of the solution will already be given in the next Penpa link. In a couple of cases you might have to solve a little bit of the grey section to complete the path above, in which case you’ll have to transfer that progress over manually.
The intended experience here are the Penpa links that include all the rules and examples, but certain input modes might still cause some amount of lag, depending on your setup. If this is an issue, you can try using the version without rules (and refer to the rules document above), as that greatly reduces the underlying grid size.
| With rules | Without rules |
|---|---|
| Part 1 | Part 1 |
| Part 2 | Part 2 |
| Part 3 | Part 3 |
| Part 4 | Part 4 |
| Part 5 | Part 5 |
| Part 6 | Part 6 |
| Part 7 | Part 7 |
| Part 8 | Part 8 |
| Part 9 | Part 9 |
| Part 10 | Part 10 |
Note that these are using my own fork of Penpa+, which contains a number of improvements that were necessary to make this puzzle feasible in Penpa at all. They will be changed to links to the main fork once my changes have been accepted there and I am not committing to keeping my fork available or stable beyond that point. So if you want to share the puzzle, please don’t link to these Penpas directly and send people to this blog post instead.
Finally, you download the entire puzzle as a single image, either as a PNG or SVG. There are three versions here: (a) the full version including rules, examples and shape banks, (b) a version that drops the shape banks, (c) a version containing only the puzzle grid itself.
| Everything | No shape banks | Puzzle only |
|---|---|---|
| PNG (13.0M) | PNG (12.6M) | PNG (5.2M) |
| SVG (5.0M) | SVG (5.1M) | SVG (3.6M) |
I don’t want you to get stuck! This puzzle is full of wonderful ideas all the way to the end, and I don’t want people to miss out on those because they hit a roadblock along the way. So I’ve set up a hint page with hints for every section, which hides each hint until you click on it. You can also download PDF versions, but they’re just plaintext, so you might accidentally spoil yourself on later sections if you’re not careful:
| Hints |
|---|
| HTML (recommended) |
| PDF, A4 (2.5M) |
| PDF, Letter (2.5M) |
If these hints aren’t working for you, and you can’t get more bespoke hints on Discord (for example), I encourage you to peek at the solution to get you unstuck and see more of the puzzle:
| Solution |
|---|
| PNG (5.8M) |
| SVG (4.5M) |
Before I get to all the wonderful folks who wrote a part of this puzzle, I need to shout out a few people who have done a lot more than that:
As for the authors of the puzzles themselves, they are listed next to their section in the puzzle grid and also in the rules document, but here is the complete list (in the order of puzzle sections):
Martin Ender (Menderbug), boboquack, Lavaloid, Andy Tockman (tckmn), Michael Tang, Botaku, Eric Fox, Ammar Fathin Sabili, DireKrow, yosh, ft029, Anonymus25, Blaž Urban Gracar, Walker, dohz, Sam Cappleman-Lynes, phenomist, au voleur!, Zachary Barbanell, Teal, Jonah Ostroff, scor, Deusovi, bakpao, wen, kays, Christian König (CJK), lemononmars, Tonta, moeve, David “Taco Dave” Millar, Tom Coward, Rubrica, Craig Kasper, Joseph Howard, Wessel Strijkstra, Emerson Golladay (Rook), Barbitos, Seren✩, djmathman, Elyot Grant, muhorka, IHNN, Stef, Prasanna Seshadri, Jamie Hargrove, jovi_al, wormsofcan, Rever, William Hu (TheGreatEscaper)
]]>The example puzzle below is the one we also used in the instructionless context, so feel free to try and deduce the rules from that, instead of reading them below.
I don’t have a ton to say about the genre itself. We needed an object placement ruleset and I was looking at the various inputs Penpa supports for answer check and wanted to do something novel with them. I landed on the half-triangles which are normally used for Shakashaka and thought there might be something interesting one can do with placing diamonds.
As a side note, I have tried reusing this structure as a Bramble variant (replacing the dominoes with diamonds). The first attempt looked promising, so I’ll be exploring that eventually.
]]>Since I made Kiki Loop a couple of years ago, the terms Bouba and Kiki have found some wider adoption to describe loop variants that require obtuse or acute angles, but I’ve never revisited the idea myself. Also, to my knowledge, no one has applied these variants to region division genres yet. So I made these puzzles to rectify both of those things.
Both puzzles are quite tough, but I hope they convey a lot of interesting insights about the structures resulting from these variants.
]]>Round 14: Choco Banana (Planet)
Round 15: Slitherlink (Singular)
]]>There was no particular occasion this time, we just felt like doing this again. We were initially going to make a Pentominous/Square Jam hybrid and while that seemed like a promising ruleset, it felt very unsuitable for this process, so we started over.
This one took a lot longer to put together than the previous one (we started working on this in late May), and you will probably see why when you’re solving it.
]]>As usual, you can find PDFs for all rounds of the championship on the Sfinks website. Please check out the other rounds as well.
I realised last year that I really love domino variants, so I made an entire round of them. As with last year’s Superposition round, I tried to cover the five main genre categories. I also wanted to pick classics as the base genres, so the learning curve of an all-variants round would be offset by familiarity with the core rules. Like last year’s Guide Arrow round, I wanted to do two puzzles in each genre, but this time I didn’t try to make one easy and one hard. I figured this would just create too big of a gap in point values in each pair, where weak solvers would just have to skip all the big pointers and stronger solvers who can’t quite finish rounds would be best off skipping the small puzzles. I was hoping that there would naturally be some variation in difficulty which worked out well.
Instead, I settled on some theming constraints throughout the round. So each puzzle only has clues that come in pairs (it’s a domino round after all), and one puzzle has equal values in each pair and the other has unequal values.
Pentominous (Dominoes) — Example
]]>I should’ve seen this coming when I made Water Walk and Fire Walk, but those have led to a proliferation of Walk genres on Discord (especially in the Cracking the Cryptic server) and even an entire set of original rulesets for EnigMarch by TwoHoleStraw. One of the rulesets from the CtC server was an Arbor Walk by kinoseidon, which inspired me to make this Forest Walk. The hope was to end up with solutions that are reminiscent of a map of branching forest trails, which I think mostly succeeded.
I was afraid that the forest patches would be huge uniqueness nightmares, but these puzzles have been surprisingly forgiving in their construction. There are some interesting parity constraints on the branches, and the way neighbouring branches interact tends to give just enough structure to resolve them without too much trouble (usually).
The first puzzle below is an example with solution, the other three are the original prototype (which serves as a nice introduction) followed by two new puzzles which are a bit trickier.
]]>I don’t have much else to add to this puzzle. It’s pretty tough, but it’s got some nice logic.
]]>I believe the idea of generalising rectangular areas to convex areas on other grids was first used by Discord user blotwell. It seems a lot more interesting than just allowing hexagons or triangles for example. So I wanted to explore that variant for a while, and this contest gave me a good excuse to do so.
]]>The prompt was to make puzzles with clues whose exact values are unknown but somehow restricted, such as Knapp Daneben or Even/Odd variants. Initially I wanted to pick a genre with non-numerical clues and made a Guide Arrow variant and this puzzle.
I believe Bryce Herdt originally came up with this Pentopia variant a long time ago and when GMPuzzles finally published it earlier last year they asked me to make the warmup puzzle for it. I really liked the ruleset and thought this would be a good opportunity to explore some trickier logic on a larger grid.
]]>
