malvasia bianca

Cocoon is an extremely well made game, maybe one of the best designed puzzle games I’ve ever played? It’s a puzzle world where the puzzles themselves are artificial but embedded in the world, in the sense that the puzzles involve stuff like carrying objects from place to place, putting objects in receptacles, flipping switches, etc. And all of that is in one continuous world. (Well, a few different continuous worlds, but they’re still pretty large, and they’re interconnected.)

Which raises some game design questions. How do you separate puzzles from each other within a single world? If the elements that make up puzzles have non-local effects, how do you contrain the search space for the player within the level, and do you even want to? Usually the scope of a puzzle in Cocoon is pretty obvious: there’s a gate that you want to get through, say, and some switches to flip and items that you’re carrying around or are in the area; but, if the player doesn’t figure out quickly what to do with that, how can you reassure them that no, everything they need is right here, they don’t need to wander all over the level.

And I liked Cocoon’s solutions to that issue. The game isn’t particularly subtle in how it deals with the problem: fairly early on, I went through a gate, it closed behind me, and so I could only move within an area that was about a screenful in size. And that’s one way that the game tells you that everything you need to solve the next puzzle is right here within this section of the world. They don’t put in gates like that all the time, but I ended up actually finding those gates reassuring instead of frustratingly confining, because they’re the game telling me that things are okay.

This is easy to describe, but I suspect it’s hard to design the world in a way to enable that. You carry around orbs in the game, and they unlock different powers: so if you need a specific orb to solve a puzzle in a given region, then it would be a problem if you weren’t carrying the orb when a gate closed! Or if the physics were sufficiently free-form, then maybe I could bypass that gate either on the way in or the way out of that area; potentially cool for speed runners, but not so great if it can lead to players stuck because they got somewhere they shouldn’t without tools to get out. (I suspect this is part of the reason why the first half of Portal consists of a bunch of discrete levels: it limits the scope of problems that you can get into.)

And the problem must get even harder to design with as the game goes on. One aspect of the game is that you can enter the orbs that you’re carrying around, with each orb representing a separate world: so the gates in the world have to take into account the fact that you might be dipping into different worlds! You can’t do that at arbitrary locations, but of course the game wants to let you do that decently often in order to increase the potential scope and richness of the puzzles; so they really do have to grapple with this issue.

But somehow they manage to do that in a way that doesn’t feel particularly heavy-handed, even as the puzzles get more elaborate. I remember one puzzle in particular, towards the end of the game: they were testing your knowledge of multiple different orbs, and what sorts of interactions were enabled by going into and out of them. And I remember thinking “hmm, in the previous puzzle, there was one switch that I could flip that transported me in a way to enable some particular solution component – do I have to use that for this particular puzzle too?” It felt like that was far enough away that it would be a bit mean / out of character for the game to ask me to incorporate that into the solution of this puzzle: the closer parts of the puzzle were complex enough already, after all. But I couldn’t think of anything else to try right then, so I walked back there: and I was actually glad to see that the game had disabled that switch, I appreciated getting a clear signal that “no, this isn’t how you’re supposed to solve the puzzle, go back to the core of the puzzle and play around some more”. And, sure enough, I had the solution to that puzzle a few minutes later.

So: really good world design supporting the puzzle design, and also really good puzzle design. The game does a very good job of teaching you skills, having you combine the skills, and of knowing when to move on from some specific class of puzzles because you’ve shown that you know how to do that kind of puzzle, you don’t need to keep on proving yourself. There were one or two puzzles where I was stuck for longer / had to think a little more laterally than I would like, but it’s hard to avoid that sort of thing happening, and those examples didn’t seriously interfere with my enjoyment or progress through the game.

And they added other little niceties, too. For example, when you’re working on a complex puzzle, and when you’ve got things arranged correctly but haven’t actually done whatever the last step is to solve the puzzle, the game plays a subtle chime. Took me quite a while to even notice that the game was doing that, but it does, and it feels good.

The game isn’t 100% puzzles: there are also a handful of combat-based boss battles in the game. Those did interfere with my enjoyment of the game: I’m pretty comfortable in saying that the boss battles were just a bad idea, that the game would be better if they were removed. But they weren’t particularly long and it didn’t take me too many tries to make it through any one of them, so ultimately it was fine: a bad choice but not one that overshadowed the many strong parts of the game.

A really good game. A nice set of puzzle primitives that combined in interesting ways, a well-judged difficulty curve of puzzles to keep the game interesting, a good job of embedding those puzzles in the world, and a game that didn’t overstay its welcome, it was fun while it lasted and didn’t push on past that fun. And I’ll just pretend that the boss battles didn’t exist.