Electron Dance

Puzzleworks 1: Puddle Knights

It's high time I wrote about some of the puzzle games I've been tearing through recently. Although "tearing through" is probably an exaggeration. It's more like syrup dribbling through a heap of used coffee filters. Anyway: the first is Puddle Knights.

Nutshell: Clever mechanic that tickles same neurons as a Sokoban game or Snakebird. I cheated on the final level.

I was told I could have a review key, if I so desired, for Puddle Knights (Lockpickle, 2020). These days I'm inundated with such mails - I can barely keep up with the puzzle PR pushed through my letterbox and considered nailing it shut on more than one occasion. The name also sounded uninviting. "Puddle Knights" was thematically on point but far too mundane to quicken the heart. It sounded more like a throwaway casual game than an intense mind-busting affair. Dung Masters. Dirt Quest. Bird Brains. Puddle Knights.

However, something about the premise intrigued and so I took a leap of faith... into muddy puddles. In Puddle Knights, nobles must get to their shiny-glowy-teleporty-portal without walking across muddy patches. To do this, your knights must lay down their capes to cover the puddles and allow the nobles clean passage. This perfect puzzle idea is full of little nuances that make it much more interesting.

As you might expect, there'll be more mud than cape, so you'll need to reuse a knight after a noble has passed over one puddle. But once a noble is standing on a cape, the knight cannot move and tug at the noble's feet, so you must place your capes carefully so that nobles don't fix your knights in place unnecessarily. However, if another knight is standing on a cape, the knight can move causing the fixed portion of cape to tear off; this torn section becomes a permanent covering on the ground. These basic concepts conjure more complexity than you might think.

Although you're not pushing blocks around, Puddle Knights still requires Sokoban thinking. You are either juggling actors in a tiny amount of space or dealing with capes that seem just too short to do what you need to do. And, occasionally, too long. Having multiple knights with capes trailing behind them evokes Snakebird although none of your Snakebird skills are applicable here: it's an entirely different challenge. However, there are specific manuevers that prove to be useful over multiple levels like you find in Snakebird; some of the Puddle Knights levels attempt to teach them. (A notable example is the crucial pirouette move in World 6.)

The theme is brilliant: I can't think of any other story that describes the mechanics so efficiently and I suspect that the theme, of putting down capes for people to cross muddy puddles, was the original genesis of the concept. The graphics are crisp and readable although in the later levels I did miss a few open pathways - thus causing a great ten-minute-long confusion - simply because the ground and boundaries are both a mix of grey and brown. You can rotate each level as you wish but that can cause your brain to jam up and interpret it as a different puzzle instead of the one you've already been working on for half an hour.

Puddle Knights does not stop with capes. Raised platforms will give your knights a headache and drawbridges, effectively one-way passages which can only be operated by knights and kept open by their capes, are maddening. There are 100 levels stretched over 8 worlds, many of which are optional to make progress. Optional levels usually challenge you with shiny metal capes which don't tear.

The most difficult puzzles are reserved for the end where the capes have holes; a noble will not walk on damaged cape. I guess I have a trace of trypophobia (honestly, don't Google this in image search) in my psychology because I find these holed capes a little disturbing to look at. Now while most Puddle Knights challenges require investigation to find "the problem", I found there were a few which needed a little too much investigation to find the solution - that is, more Monte Carlo than smarts. I found this issue worsened with the holed cape levels.

In fact, it got so bad, the final core level was beyond me. After spending hours on it and establishing what I thought was the correct midpoint of the solution, I had regressed to guessing. I couldn't reverse engineer what I wanted from the ending as certain moves seemed to be impossible and so I surrendered to a walkthrough. It was Nova 7 all over again.

Still, do not worry. If you complete every level including the optional ones, World 8 opens up and offers a small set of fascinating edge case bonus levels. I thoroughly enjoyed these even thought they were not as difficult as many of those that came before. And this made for a much better ending than the official last level. I'm a fan of a fun endings rather than a super-charged pain-challenge ending - final boss levels marred my feeling of closure in both Cosmic Express and Snakebird.

So, overall, I loved Puddle Knights and found it compulsive. Puddle Knights is currently available for Windows and Mac from Steam. I received my copy for free.

Next Puzzlework: Tametsi

Further Watching

You can also watch me explain how Puddle Knights plays in the video below.

Download my FREE eBook on the collapse of indie game prices an accessible and comprehensive explanation of what has happened to the market.

Sign up for the monthly Electron Dance Newsletter and follow on Twitter!

Electron Dance Highlights

Comments (10) Trackbacks (0)
  1. Also Puddle Knights has a 30% discount this week.

  2. Also, also, on the Electron Dance Richter Scale of Challenge, I’d say it was hard.

  3. Always nice to see notable puzzle games pop up from new devs, though I’m trying to not start anything new ATM as I’m hard stuck in both Matchstick Elegy and Vertebrae… Looking forward to more puzzle game writings, would be curious to see more in depth thoughts on Sokobond as I played that recently and found it interesting – lots of potential for concise mindbenders but I found a lot of the levels up until the finale were bloated and too prone to solve themselves.

  4. Hello Johan! This was originally going to be a write-up of four different games, when I realised writing up one was long enough, I expanded it into four articles. Then I realised why stop? So Puzzleworks will continue indefinitely. I’ll probably only write about games I enjoyed as opposed to those that gave me stomach ache. I did play Sokobond many years ago, before Ouroboros, but got stuck about halfway through, lots of puzzles I couldn’t quite figure out how to solve. I was surprised at just how many levels it contained.

    Again, this is a work earlier in Alan Hazelden’s career. His previous commercial work was These Robotic Hearts of Mine which I did not really like at all (as I said to camera many years ago).

    Not heard of either of those other games you cited, though?

  5. Sokobond does have a lot of levels, though I think the top notch level select helps. Since the levels use a lot of obstacles and have complicated end states, they’re hard to read until you master the rules, leading to flipping between Monte Carlo and frustration depending on how easily they unravel. I’ve only played Alan’s puzzlescript games besides Sokobond so not sure how it compares.

    Matchstick Elegy is from the guy who made Permanence and Shackle, I think the pacing and puzzle design isn’t great but it has some cool mechanics. Similar to The Flames, but with objects burning on a local timer rather than global. I got stuck at level 12, and still can’t figure it out after learning new tricks for the later levels, but besides that one it’s been a middle of the tree kinda deal. https://steven-miller.itch.io/

    Vertebrae on the other hand is fantastic. Rivals Snakebird in how versatile the moveset is and the puzzle design doesn’t slack with every single one being super hard and with a delightful solution. At least up until level 11, which seems impossible, so I can’t speak for the last 3/4ths of the game yet. https://www.puzzlescript.net/play.html?p=ce2474f62432e2a703bba3fb65f5b01f


    I should do another run-through of Sokobond. Right now the only thing I remember about it is that there’s one bonus level “Open Space” which is

    one hour later

    as I have said absolutely impossible for me to hold the solution in my head OKAY so I went and played through the first three areas and yeah, I did find myself Monte Carloing a lot. There are some techniques–don’t know if they merit the name–like filling up the bonds on something so you can use it as a pusher, and assembling a hook to pull things back when the you atom can’t get close enough to a wall, and some ways of maneuvering some things around splitters. But there was also one level (Chandelier) which was all about eventually realizing there was one way to start that I hadn’t thought of. And the non-splitting levels were so much about becoming the right shape that it was like playing a game of Snakebird where everything was Tetris drops.

    A lot of the time here I’m working from my knowledge/memory that Hazelden likes schmuck bait puzzles–there’s often a simple way to start, which you try even though you haven’t shown that that’s the only way to go, and it’s wrong. You have to consider all the possible opening moves, even non-obvious ones, and try them out. So when I’m stuck I know it’s often that I need to try something different at the very beginning.

    Hazelden’s later non-Puzzlescript stuff is not like this for the most part–the levels in A Good Snowman Is Hard To Build are very small and Cosmic Express has a big sweeping line-drawing interface so it’s not so costly to experiment for the most part, plus the end states are much more restricted. Cosmic Express is also not Sokbanny at all. Though wow, if you got 11 levels into Vertebrae awkward shapes don’t seem like they should be much of an issue. I have tapped out at level 3.

  7. This week on puzzlectron dancepuzzle .sokoban

    my inferiority complex deepens

    I’m not playing the puzzle games you see

    does anyone want to hear about my MMR?

  8. Thanks Matt for doing a quick run at Sokobond for me; it sounds like I might end up in the same situation I was in when I first tried it out. I wonder, sometimes, whether a lot of people who “like puzzles” are actually Monte Carlo players and that’s how they think puzzles are solved. In other news, I’m currently stuck on level 7 of Vertebrae and, Jesus, this is a real headspinner. Still a lot of Monte Carlo – it would have been nice if it helped you build your skills better because it seems like a massive leap from zero to hero in just a couple of levels.

    Sorry CA, I was planning to the write about Something Else this week but that Miracle Sudoku video prompted me to write about Tametsi instead. The Puzzleworks are not supposed to be weekly…

  9. Well I should say that I don’t know what I meant about that one level, or rather I do think I know what I meant but it was wrong. “Chandelier” is a level that makes perfect sense with one exception, and combines a few strategies/techniques.

    One is figuring out the directionality of the bond you need–if there’s an H on the left wall, you know it has to be bonded from the right or vertically, and that if you’re bonding it vertically the atom it bonds to can’t have anything to its left already. Often this will help you cut down your solution space combinatorically (there are two Hs on the left wall, there’s only one available right bond, that means you have to bond one of them vertically and then pick up the other rightward–that doesn’t apply in Chandelier to be clear.)

    There’s also the aforementioned bond-something-to-make-a-pusher technique, and making things the right shape so you can maneuver them, which is more easily foreseeable here than in some levels, and working your way around atoms that you don’t want to bond too soon.

    The one thing that makes me a bit twitchy here is an analogue of a movement order thing. Sometimes you make a move that could lead to more bonds than are actually available–imagining moving an H next to an O and an N, where the H could bond with either one but it can’t bond with both because it has only one free electron. Which does it bond to? The rules for this aren’t ever specifically disclosed; I think that it’s that the atom you control bonds first, then atoms with higher numbers of bonds go (so the N would bond before the O), then I have no idea. I had some solutions that depended on this where I think it was possible to avoid this, but in this level I’m pretty sure you have to count on this.

    I’m almost positive I meant the level Rosie. But when I went back to try it again, it made more sense to me–if you think about directionality of bonds, what you need to avoid, and what shape you need to be so you can move around (and when you can afford to lose maneuverability), it makes sense.

    Anyway, NOT that I want to badger you back into Sokobond, but I feel like I got more insight into some puzzle types. One is the pure flash of insight, where there aren’t any technical obstacles at all and it seems like you can’t do anything until you see the thing you need to do. There are a lot of Baba Is You levels like this (complicated by the fact that you can’t tell what some of the rule behaviors will be until you make that rule). To spoil something from a game nobody will probably play, Flash games being dead and all that, in one of the Blocks With Letters On games where the goal is to move block with letters on them to form a word in a goal area, there’s a level where you start with “BELOW” and there is a blank space below the third square in the goal area, and the L and W don’t float so would fall if placed in the third square. There’s a lot of machinery in the right side of the level that always leaves you one step short of being able to solve the problem… but the solution is to form “ELBOW” instead. (This is a blatant troll.)

    Another thing is the careful slow boring of hard boards kind of problem. The last of the initial group of Snakebird pushable-block levels, which you absolutely hated, is like that. (Level 40, “Head of a Pin,” where you have to make a staircase of two two-by-one frames on the top of the level to reach a distant exit.) For me this was about gradually working the snakebirds through the frames, with a little trick about weaving a snakebird between the frames when it can’t reach the top, without letting things drop. There’s a lot of trial and error involved here but it wasn’t Monte Carlo for me because I could see when I was making progress, so it was about inching forward. Sudokus seem to have a lot of this (in the Miracle Sudoku puzzle where he’s cleaning up all the 3s, for instance). And a lot of people find it boring, but it’s meditative exercise for others.

    Then for me pure Monte Carlo is when I’m trying things just to try them without seeing where they could possibly lead–then after banging stuff around for a while I look up and that thing is to the left of the other thing when it was to the right before, and that’s what I needed. Which isn’t satisfying.

    But puzzles also often involve a sort of constrained Monte Carlo where you can see–or preferable, figure out–that some conditions are fail states you need to avoid, and that you need to accomplish this, and then you can experiment with what you can do to tiptoe around the fail states. Which is more satisfying to me than straight up Monte Carlo because I at least know some of what I’m doing. That can also get into the meditative exercise zone, even if it’s not always repeatable.

    This is distinct from a kind of Matryoshka where you start “How can I get out of this seemingly impossible state?” and then you see how to do it but there’s still a lot of puzzle with how to get to the solution. Sometimes this is a sequence of insights (or “linchpins” as DRODders apparently call them), sometimes there’s more work involved. What these have in common with constrained Monte Carlo is that you are avoiding a fail state but once you’ve avoided it you still don’t see how to win–the difference is that in Matryoshkas you can see how you’ve made progress, and in constrained Monte Carlos you may never be sure that your experiment is working until you’re in sight of the goal. And there can be Matryoshkas that work by a sequence of constrained Monte Carlos, I suppose.

    I am going to consider it an exercise of my willpower never to try Vertebrae again. Just ran straight into a difficulty cliff there.

  10. I mean, you, remind of that fine line between true Monte Carlo guesswork and experimentation-with-mindfulness. Sometimes you do have to throw the dice occasionally but if you’re paying attention, you’re learning something from these experiments. However: there are times where you’re throwing the dice a lot and your experiments are giving up nothing. I had this with the last level of Puddle Knights. Now I like those feelings of insight, that you’ve seen something important, but is it enough? That creeping doubt, though, that maybe this is a red herring, a freaky strategy that leads nowhere…

    (Still, I don’t think I’m going back to Sokobond at this late stage.)

Leave a comment

Notify me of followup comments via e-mail.

No trackbacks yet.