Saturday 8 December 2012


I don't much enjoy solving puzzles.

I like uncertainty and randomness in games. I like strategy; I like complex open-ended decisions that have a chance of going horribly wrong, and doing my best to deal with unexpected situations that arise; I like taking risks, winning with a sub-optimal but surprising approach. I don't like digging through a limited state-space to find a particular solution, I find it menial to do work that I could just get a computer to do. I've written a program to solve Sudoku puzzles, so why would I ever want to solve one by hand? (I'm not a big fan of Chess for the same reason; although it has deep strategy, a lot of the time you can gain an advantage just by looking more moves ahead: computer work.)

There are some things that look a bit like puzzles that I do enjoy. I love solving mathematical problems, designing and balancing games, writing algorithms. These are very open-ended, and sometimes you're not guaranteed that a solution exists - in mathematics you might be able to prove that one definitely doesn't. There's more ambiguity; you might not want just any solution, but the most beautiful or efficient one, or the one that makes the ideas behind it clearest; and if no perfect solution exists, you need to judge which criteria to weaken to get an approximation. I actually did enjoy puzzles when I was small, but having tasted higher things the lower no longer hold my interest. (Some mathematicians really enjoy solving Sudoku and the like - to me this seems crazy; how can that tedious elimination hold any joy when you've been out proving real theorems? - but different people are different.)

I make games for myself to play. So this doesn't obviously lead to making puzzles - not only am I not so much into them in the first place, but I'd already know the solutions so I'd enjoy them even less - instead I focus on highly replayable types of game. But sometimes I've ended up with puzzles by accident; you can never be sure when you start following an idea where it will end up taking you. The Sense of Connectedness started as a strategy war game (although this is more about experimenting to try to figure out what the rules and objectives are; once you know those actually solving the puzzle is pretty basic). Game Title / Lost Levels started as a kind of dungeon crawl. Vertex Dispenser's puzzle levels.

Because I can't really appreciate these puzzles I've made, I kind of forget about them, I don't think of them as significant. I was talking with Terry Cavanagh a couple of months back, and he said The Sense of Connectedness and Lost Levels are his favourite things I've made, which kind of surprised me - for me they're quite low on the list, overshadowed by things I still enjoy playing. But it reminded me they exist, and got me thinking about how I actually really enjoyed the process of making them. Lost Levels especially was a wonderfully difficult problem, one of the best times I've had making games, rocking back and forth muttering to myself trying to fit everything together. While solving puzzles is quite closed, making them is the kind of open-ended thing I love. So maybe it's worth me trying to make more puzzles - if others appreciate them and I enjoy making them, it doesn't matter so much if they're no use for me to play.

So I've been working on a new puzzle on purpose. It's optimised to be difficult to make - everything is interconnected in complicated ways, such that shifting something by one tile on one screen could entirely break something somewhere completely different. I hope it proves interesting for others to solve, but my main goal has been completely selfish and hedonistic: looking at the pleasure I got from making those other puzzles and trying to maximise that. And it's succeeded: it's been extremely challenging and frustrating. Negative mushrooms, myriad ways to get stuck, skeletons, impossible configurations from visiting rooms in a particular order. There's a tricky balance between individual sections and the holistic magic of it; some test subjects were getting stuck in an early room and not getting to the really interesting part, so I opened up the early areas letting you choose what order to solve things in and possibly skip a few you're stuck on. But this has created a new problem, where if you've left a room in a state from which it cannot be solved and go somewhere else without undoing it first, you might end up having to undo any progress you make elsewhere in order to solve it. And unfortunately because of how everything is interconnected it doesn't seem feasible to reset rooms independently of each other. The whole thing's been like this: every problem solved in one place creates a new one like a bubble popping back up elsewhere. I think it's pretty good?


  1. I spent quite some time with The Sense of Connectedness but ended up stuck at a certain point - I could never quite keep the combo of "things happening" if that makes any sense.

    Picked it up via RPS and didn't realise it was my Michael the Brough until this year.

    Nonetheless - interested!

  2. I think your distaste for logic puzzles might be based on only doing the ones written by machines. Hand-crafted puzzles feel much less like boring deduction, and much more like finding the new, surprising logical steps that the creator wanted you to find.

    (I would argue that most good puzzles are like this: the constructor finds something neat, and tries to trick you into discovering it yourself. For Lost Levels II, it's about what happens immediately after you leave a screen; In Boxgame, it's about the combined effect of three translations in S^3; in a good crossword, it's some new pun that the constructor found; in a good logic puzzle, it's a new combinatorial lemma that tears through an otherwise tedious solve.)

    I'd strongly recommend the puzzle blogs of Thomas Snyder, Grant Fikes, or Palmer Mebane. Maybe steer away from Sudoku: my favorite puzzles are the ones that start with chaos and end with order, whereas a big jumble of numbers like you get at the end of a latin square puzzle is rarely satisfying to look at.

    1. hello, from the future.

      i think "your distaste .. based on only doing the ones written by machines" is giving way too much credit to designers of 'bad' puzzles, wherein the (human) creator is failing to illuminate neat things or find them in the first place.

      or, well, when what the creator considers neat isn't up to your standards, or you're not up to theirs.


      michael, i don't know if i ever read this blog post. i like it. i liked reading this immediately pre-corrypt blog post.