The Stack

Zen Bound 2 is a sequel to a cell phone game, and a pretty convincing argument that cell phones are where the real innovation in gaming is these days. It’s a game about wrapping things in rope.

Each level gives you a small wooden sculpture with a taut rope tied to a nail in it. Purely by rotating the object, you reel the rope around it to bind as much as it can. Parts of the surface within a certain distance of contact with rope get coated in paint, which comes off if you unwind — clearly it’s not so much paint in the realistic sense as a means of visualizing your progress. Once you reach a minimum of 70% of the surface painted, you can tie the rope off at a second nail to end the level, or you can keep going for more paint and greater glory. Challenge is created by limiting the length of the rope, and through hard-to-reach concavities. Remember, you can’t just thread the rope through things; it’s always taut, and only goes into grooves if the groove is where its tension leads it. On a couple of occasions, I’ve had difficulty just reaching the final tie-off nail because of this. There are a couple of embellishments beyond this — paint bombs that cover a larger area, attached to either additional nails or to your rope — but this is a game of simple design, and getting much beyond the basics would be inelegant.

The “zen” part of the title, apart from supplying a rationale for Japanese garden imagery in the menus, seems to mostly just mean that there’s no time pressure, or indeed pressure of any kind. Getting to 100% coverage on every level would be a feat (and provides an Achievement), but just getting through the minimum seems like it just requires idle fiddling. I don’t think I’ve ever seen a game more deserving of being called “casual”. It’s a simulation of toying with knick-knacks.

The technology necessary for this game to exist, though, is pretty advanced. It needs a way to apply the paint texture to arbitrary portions of a surface. It needs a good physics model — usually in games that mostly means collisions, but here it means tension and friction, making the rope slide along surfaces in convincing ways. And it needs a really good UI for rotating 3D objects. Apparently the iOS version uses tilt and multitouch for this, but that’s not an option on PC. I’ve had to implement rotating things in 3D with a mouse before, and it’s surprisingly hard to get it feeling right — mainly because 3D rotations have three degrees of freedom, but your mouse has only two. Usually, as a player, I can get used to whatever scheme has been implemented, but in this game, absolute control is crucial. So it comes up with a compromise: dragging with the left mouse button pressed rotates things one way, dragging with the right button rotates things another way, and between the two schemes, everything is covered. I’ve played the game enough to internalize this, but I couldn’t tell you exactly how it works geometrically, except that the right button lets you rotate about the direction the camera is facing.

Cogs is essentially a series of sliding-block puzzles of the sort that have bored puzzle-gamers for over a hundred years (ever since the “fifteen puzzle” craze of 1880 died down), except that instead of putting together a picture, you’re putting together a machine. There are tiles with gears on them, capable of transmitting motion, and tiles with segments of piping, capable of transmitting steam. Some levels have the one, some have the other, and a few have both.

Gears are basically simpler than pipes, because you don’t have to worry about orientation or leakage (some pipes branch), but there are levels that complicate things by putting gears at different heights on a single shaft. In fact, some of the tiles (both gear and pipe) are two-sided, so you have to flip the board around to coordinate the needs of two different sub-systems. This takes advantage of something that sliding-block puzzles based on pictures or ordering numbers generally don’t have: the interchangeability of identical parts. The game doesn’t ask you to put particular pieces in particular places. It just asks you to put together something that works, and if a piece has more than one purpose, it’s possible for it to work for one purpose and not the other.

Actually, sometimes a specific piece does need to be in a specific place. For example, if you have exactly one pipe bend, and have to connect an inlet and an outlet that are perpendicular to each other, you have to put that bend at the place where straight lines from the inlet and outlet intersect. But in such cases, the exact placement of the tile isn’t one of the puzzle’s constraints, but rather, something derived from those constraints. The first step in any puzzle here is figuring out what the derived constraints are — not “How do I get the tiles where I need them?”, but “Where do I need the tiles?” The act of getting them there is a relatively trivial matter for the experienced puzzler, and exists mainly to take up time and build up anticipation, like all the walking around in a typical adventure or RPG.

Being able to enact your designs quickly and efficiently is rewarded, however. Each puzzle you solve rates you on time and number of moves taken, and gives you a number of “stars” on that basis. Stars contribute towards unlocking more puzzles, or, once all 50 puzzles are unlocked, go towards certain Achievements. Also, there are two “challenge” modes that put hard constraints on the time spent or moves taken. To me, move challenges are an interesting extension of the puzzle, and I may well spend some time on them, but time challenges don’t interest me at all. Since there’s no randomization in this game, the obvious way to pass a time challenge is to simply memorize a solution, and that’s not how I want to play a puzzle game. Even going through the puzzles in non-challenge mode, I usually maxed out the stars for moves but not for time.

I haven’t mentioned the game’s most obvious feature from screenshots and videos: the three-dimensionality, the fact that some puzzles are on multiple sides of objects that you can rotate freely. That’s because it’s not all that important. The 3D puzzles are pretty, but the tiles are still bound to specific faces; there are linkages that transmit rotary motion from face to face, but you can’t move them. So all the 3D does is chop the puzzle up into smaller and easier sub-puzzles. There’s one puzzle that looks at first like it’s a single tile-grid spread over the surface of a cylinder, but in fact there’s a vertical seam, so it’s really just a rectangle that you can’t see all at once. I suppose the UI isn’t set up to support actual loops — you can push multiple tiles in a line at once, so if a puzzle were truly cylindrical, there would be ambiguity when you clicked on a tile about which direction you wanted to push it.

It’s possible that I’m not the target audience here. There are two puzzles that really are just traditional picture-assembly (with the excuse that you’re piecing together large gears composed of more than one tile), and the first and easier of the two warns you that it will involve extra planning ahead, as if you’re not familiar with this sort of thing. Likewise, there’s the length: solving all 50 puzzles took me less than four hours, and I expect it’ll be similar for most experienced puzzlers, provided they don’t give up.

Still, this is a game worth looking at for what it does: it takes a tired puzzle-type and comes up with ways to make it interesting again. I wonder if anyone’s tried something similar with the Towers of Hanoi? I suppose that’s kind of what The Bryant Collection does, but only in a sort of proof-of-concept way.

To my surprise, I managed to clear all of Puzzle Dimension‘s Achievements on finishing the game. You may recall my doubts that I’d reach a total score of 50000 points, as one of the Achievements requires. Going into the final level, I was more than 2000 points short of that, about four times as much as I expect to get in a typical level. But the final level wasn’t typical. It was one of those puzzle-game final levels that recapitulates everything you’ve learned, and that means it was large, and that means it wasn’t hard to get my bonus multiplier up to the maximum of 32x just by roaming around before I started trying to solve it.

I have to say, finishing this game produces the sense of epic accomplishment a lot better than anything else I’ve been playing lately. At least it does for me. I’m not sure what it is. There’s a mental phenomenon that I think most hardcore gamers are familiar with: the sense of waking up from a game. (Probably the best descriptions I’ve seen of this are those in the novel The Player of Games by Iain Banks.) After an absorbing multi-hour session, your mind is in an altered state; if you’ve just beat the end boss, that state is probably at its most intense. And that’s exactly the moment when you stop playing, so the contrast, as you stand up and become aware of reality again, is quite noticeable. I feel like I’ve talked about otherworlds in games a lot lately, things that take the place of the spirit world of the shamans. But viewed from outside, games themselves are otherworlds, and single-minded focus on playing is a kind of trance.

Anyway, emerging from the end of Puzzle Dimension produces something of that effect, which is a little strange, considering that I haven’t been playing it in intensive multi-hour sessions. Possibly it’s just the unfamiliarity of the form. I can play a text adventure or FPS and experience it as just another text adventure or FPS. But Puzzle Dimension asks us to flex mental muscles that don’t get a lot of use — specifically, keeping three-dimensional objects straight in your head without a consistent sense of gravity to orient yourself by.

Bob Came in Pieces is a charming indie physics puzzler, 2D but with 3D decorations, short enough to play to its conclusion in a single lazy Saturday. The basic idea is that you control a crash-landed alien spacecraft with Lunar Lander-like controls, and you have to find and reattach the bits that fell off in the crash so you can use them to solve various mechanical puzzles. When I first heard about this, my first thought was that it must work something like Knytt, each piece you recover representing a power-up that gives you a different special ability. But in fact it’s simpler and more complex than that. The majority of the pieces you collect are purely structural, like legos or tinkertoys: various shapes of socketed strut that let you reattach your rocket engines in different ways. Or sometimes not even that: sometimes all you really need is a poking-stick to give a tap to something past a narrow gap.

And once the tap is delivered, you go back to the building-place and reshape your ship into something sensible, because off-balance structures make you list or even spin in midair. This is where the physics comes in. You can have rockets pointed in any of the eight cardinal directions, and control them independently or in any combination you please — the game lets you assign keys to specific rockets arbitrarily — but how much of their force propels you and how much makes you spin depends on how you place them. Aside from rockets, there are two other sorts of similarly-controlled device: a puller and a pusher, both of which produce beams of directional force. The puller attracts moveable objects to your ship, but also, in accordance with Newton’s Third Law, attracts your ship to them, which can be hard to control until you’ve found enough rockets to counterbalance it. The pusher does the same in reverse, creating a repulsion between objects and your ship. But wait — isn’t that exactly how rockets work? What distinguishes the pusher from a rocket? Well, the pusher only pushes objects that are flagged as moveable. (You can tell them by their green shine.) If you’re not near one, it produces no effect. Rockets always affect the motion of your ship and nothing else. (At least, nothing else directly. Your ship can push stuff.) So this isn’t quite real-world physics we’re dealing with here, but we still have force and weight and momentum both linear and angular.

You spend a substantial fraction of the game in the ship-building interface, a rather well-put-together UI. It provides the option to save configurations to disk to reload them later, but I was up to level 11 (out of 14) before I took advantage of that. Before that point, most of my redesigns were single-use, designed around a specific puzzle — and besides, I kept getting more useful pieces. But towards the end, additional stuff seemed superfluous, and I spent most of my time with two basic designs. There was the Tiny, a minimal ship for small spaces, and there was the Lifter, which had a wide row of symmetrically-arranged pullers and rockets on its underside. It’s actually a bit of a letdown when you’ve collected so many pieces that you can easily make whatever configuration you want. In the earlier stages, you have to come up with ingenious linkages to use the parts you have, and then hope that the result is balanced enough.

The puzzles are all about getting access to places. The levels are all essentially systems of tunnels, even if they don’t look like it — in some places there are rows of trees, with their canopies acting as a ceiling. Some of the later levels provide some genuine wide-open spaces, and by that point, you have some high-powered rockets and can go zooming and swooping around like the UFO you are, with a satisfying sense of heart-plunge when you shut off the rockets and go into freefall. (Not that there’s any real danger; there’s no death in this game. Even in the one level that throws fireballs at you, all they do is smack into you and send you backward.) But the puzzles are on the periphery of such large spaces, in tunnels, so that’s where I spent most of my time. The build-it-yourself approach gives the player a great deal of leeway in the details of how the puzzles are solved, but unlike, say, World of Goo, the overall solutions are generally designed rather than emergent. You might, for example, have a platform suspended from a chain with a counterweight, and the counterweight is blocking a passage, so you have to put something heavy onto the platform before you can get past. Occasionally there are alternate routes, sometimes involving hidden passages, which may be the key to passing the time trials (something I have no intention of even attempting). And occasionally it’s even possible to bypass the intended solutions by doing something fiddly and inefficient, but that’s hardly satisfying.

I do like this game overall, but I can’t help but wonder what a different design sensibility could have done with the concept — maybe something that would make me create more than two reusable ship designs. I suppose this means it’s a good candidate for mission packs, fan levels, and/or sequels.

I’m definitely in the home stretch on Puzzle Dimension. One more session should be enough to cover the remaining three puzzles, if my progress through the rest of the cluster is any indication. It really seems like the puzzles have been getting easier toward the end, but this could be illusion — perhaps I’ve simply cracked this game’s code, learned the right way to think. Or maybe the level designers are just running out of tricks, as happened in The Humans. If so, it’s a good thing it’s ending before it gets too repetitive.

Since I’m paying so much attention to Achievements lately, let’s look look at what I have left here. There are four I’ve yet to get. There’s one for completing the last cluster, and one for solving every puzzle in the game. Obviously I’m going to get both of these together, but that’s not necessarily the case for everyone — you can complete the last cluster while there are still puzzles remaining in earlier clusters. There’s one for unpixelating every tile in the game, which I’ll probably also get at the same time, but only because I’ve been diligent: many puzzles don’t require you to go everywhere. None of them make it particularly hard, though; if you can hit all the flowers, you can get within one tile of everything else. You just have to remember to do so. The level-selection UI displays levels you’ve completely unpexelated in a different color, so it was easy to go back and fix things up when I noticed that this achievement existed.

That leaves just one Achievement: the one for getting a total score of 50000. As I write this, I stand at 45912. That’s far enough away that I almost certainly won’t make it within the next three puzzles, but close enough that I could plausibly pull it off by replaying a few of the earlier ones. When you come down to it, all it asks is that you get an average score of 500 per puzzle, and 500 points is usually a reasonable goal, provided you can get your multiplier up to x8. But then, on some levels, particularly the early ones, it’s just plain impossible. I don’t doubt that I could reach 50k with a bit of effort, but why go to that effort? The pleasure of a puzzle game is in solving the puzzles; once I do that, this game is meaningfully exhausted. I suppose it all comes down to how much it pains me for just one checkbox to remain unchecked. I’ll go some distance for true completion, but I have my limits.

If there’s one thing Puzzle Dimension really hits home, it’s the dimensionality. This is an experience that’s more fully three-dimensional than most of what we do in real life. I mean, in real life, we’re generally bound to a single surface, with a seemingly consistent normal defining the direction of down. PD lets us explore any surface, regardless of orientation. Let me just share a few illustrative screenshots:

This level is made mostly of two partial shells, made of pathways along a convex surface, joined by some stepping-stones. Because of the way the shell bends, paths that are partially parallel cross and diverge. The implied hull has a definite inside and outside, with flowers on both, so the player has to seek ways to cross between the two — the hook-shape on the right side is a way to transfer from inside to out, by rolling off the end of the hook and dropping to the other side. Dropping from outside to inside is easier, and can be done in multiple places, but the falls necessary are longer, and more dangerous-feeling, because it’s hard to feel completely sure that things are lined up right over a longer distance.

There are several levels with this sort of barrel shape. Again, we have a distinct inside and outside (and yes, it is possible to make continuous nonorientable surfaces in this system, although I don’t have a screenshot of one handy), but this time, the design works against maintaining any other sense of orientation: the paths, especially the ones around the curve of the barrel, wind too tightly to easily maintain a sense of absolute direction. Note that the blue tiles are slippery ice; only on the brown tiles can you stop moving. The tentacles here are similar to the hook in the preceding picture, and you can switch from inside to outside by dropping off any of them, but if you’re not careful, you’ll land on an ice tile and go flying off the edge of the structure.

This level has a shape that I wouldn’t have thought possible given the game’s constraints: using only 45-degree bends, it produces an elongated three-sided spike. It manages this by constructing three planes that are perpendicular to each other, like the sides of a cube, but whose grid lines are diagonal to the cube’s sides. Before seeing this level, I kind of intuitively assumed that the tiles, regardless of orientation, were part of an overall global 3D grid. But of course this is inconsistent with the observed facts: all tiles are clearly square, so anything situated at a 45-degree angle would break the lines of such a grid. Nonetheless, most levels before this one are arranged more or less rectilinearly.

Given the extreme three-dimensionality, it’s especially interesting that the designers decided to make an entire cluster of levels that are flat or mostly flat. I suppose it’s their way of showing that they’re not entirely dependent on disorienting the player, that they can make interesting puzzles even when deprived of their main tool.

Well, I’ve just opened up the final cluster. I haven’t delved into it much yet, but the first puzzle in the cluster is significantly easier than the puzzles I had to solve to make it available. In fact, the remaining ones, the ones I haven’t managed to solve yet, are difficult enough that it could take me several days to finish them.

Since we’ve got multiple posts ahead of us, let me take a moment to just talk about how polished it all is. The graphics certainly take advantage of current technology, all bump-mapped and shiny with the appropriate sheen for ceramic tiling or painted wood or whatever is appropriate for your current selected theme. There are four graphical themes, unlocked one by one over the course of play: a yellow-and-brown one with desert-like connotations, a cool whitish one, a wooden one on a green background, and an underwater one. At first I didn’t think much much of them — sure, it’s nice to get some visual variety, but switching themes means I have to get used to how the various tile types look all over again. But then I realized that the background art in the underwater theme — which includes a 3D sea floor, a shimmering surface above, some sea life and rusting barrels and Atlantis pillars — provided cues for remembering my current orientation, in contrast to the more uniform backgrounds in the other themes. It’s hard to say this for sure, but I think this helps even when I’m not consciously paying attention to it.

There isn’t a lot of physics in this game — there’s basically no moving objects other than the ball, and it moves in a simplified, rule-based way, always falling straight down. But there’s one bit of object interaction worth noting: if you brush past a flower — say, because you’re falling through the space it occupies, perpendicular to its orientation — it bends appropriately. It’s a little touch, but a nice one, especially considering that it’s the only sort of interaction that needs such special treatment. What does the game do if the flower is still pixelated? Well, first of all, if you’re close enough to brush by it, you’re close enough to unpixelate it. So it does that, and then the flower bends.

The sounds in the game are mostly arcade-like beeps and boops, but there’s a nice touch in the way the gameplay affects the background music. As in, for example, Peggle, changes in the music reflect your progress. You start off each level with chiptune background music, all square waves and drums made of white noise, and as you unpixelate the level, the music, in effect, unpixelates as well. It’s a subtle effect, because the music remains heavily electronic and synthesized even when it’s unpacked; the thing that really made me start noticing it was the sudden shifts back to NES-quality when you restart a level (something I didn’t do as often in the earlier levels). I suppose it’s an instance of a general theme, that this world, the puzzle dimension, becomes realer as you touch it, more detailed and particular as a result of your experience of it. Even the unlockable themes progress from the abstract to the concrete.

Now, I’ve said that Puzzle Dimension is essentially turn-based. And this is true. But its scoring mechanism is extremely realtime.

I don’t have a good handle on the details of the scoring system — it’s definitely tracked per-puzzle so that you can try to beat your score on any particular puzzle on revisiting it, and it seems like you get points for each flower collected, for finishing the level, for unpixelating tiles (that is, moving onto or near them for the first time), and even, if my observations are correct, for making breakable tiles collapse. But the game doesn’t tell you outright what specific actions are worth, and I haven’t bothered figuring it out. The one thing that I have figured out is the multiplier. It starts at x1, then proceeds to x2, x4, and so on exponentially up to x32 at least (although this can only be reached on some levels). So you can see that getting the multiplier up is crucial to maximizing your score. The way you increase the multiplier is by unpixelating tiles. Behind the score on the screen there is a progress bar showing how close you are to increasing the multiplier. When you move into new territory, it goes up. But over time, it decays. You don’t lose your current multiplier to decay, but you do lose your progress toward the next.

Now, at this point, you might be saying “Why do you care? This is a puzzle game. The goal is to solve the puzzles. If you do that, you’ve won, regardless of the score.” And this is certainly how I usually feel about puzzle games with a scoring element that depends on speed. But, well, this game has Achievements, and some of them are linked to scoring, and that makes me care a little. There’s one Achievement, for getting a total score of over 50000, that I probably won’t get — Steam tells me that it’s the least-achieved Achievement in the game, achieved by only 0.4% of players. But I intend to make an honest try at it.

(Tangentially, I like what they’ve done with the Achievement icons in this game: they’re on backgrounds that are color-coded by type. So all of the “complete all the puzzles in cluster X” Achievements are on blue, all the “get a multiplier of Y” achievements are on green, etc., making it easy to spot trends in a list ordered by other criteria.)

Anyway, playing for points affects how I approach the levels. The first goal in any level is to get the multiplier up quickly by hitting as much of the board as I can without consequences. Once I run out of places I can go without thinking about it, it’s time to put the game into Camera mode, which pauses the game, including the multiplier decay, so I can actually take a look at the whole thing and make plans — plans which may involve more multiplier-grabbing in places that weren’t easily accessible. So the overall pattern becomes one of alternate rushing and contemplation.

Eventually you reach the point where there’s nothing left to grab — either you’ve unpixelated everything (which turns the progress bar green), or the only thing left is on a one-way path to the exit portal. The exit portal isn’t necessarily open at this point. Just because you’ve been at or adjacent to every tile doesn’t mean you’ve figured out how to get to all of the flowers — maybe you’ve only been to the flip-side of some of the flower tiles, maybe there’s ice or sand interfering with your ability to move directly between adjacent stuff. But I’m finding that even when things are in this state, I’m still acting as if I needed to rush everywhere and then pause the game. Habits come quickly in games.

Like many games, Puzzle Dimension introduces new game elements one by one over the course of play. Except it kind of introduces them in pairs. The levels are grouped into “clusters” of ten, and all of the earlier clusters introduce two elements. If you play the levels within a cluster in order, you get the new elements one at a time, but you don’t have to play them in order; unlocking a cluster makes all of its levels available.

Most of the clusters have titles that play on the things they introduce — for example, the first cluster, “Broken Ice”, introduces first breakable tiles, which crumble after one use, and then ice tiles, which force you to keep moving in a straight line. It’s a pairing with a great deal of puzzle potential, as the ice tiles make it important which direction you move onto them from and the breakable tiles limit your opportunities to approach them as you like, but there’s still something peculiar about it. Because you can still jump when sliding on ice, and indeed sometimes have to in order to solve a level, ice tiles represent the only realtime element I’ve seen so far. Breakable tiles, meanwhile, really emphasize that this is at heart a turn-based game: you can sit on them for as long as you want, and during the entire time that they bear your weight, they play a wobbling, crumbling, just-about-to-collapse animation. But they don’t actually collapse until you move on.

The second cluster, “Jump in the Fire”, gives us springboard tiles, which catapult you forward three spaces (one space farther than you can jump), and then fire tiles, which start off dormant, but become deadly after you use them once. This is a similar pairing to the previous cluster: you get one thing that lessens your control over where you go and requires you to approach them from the right direction if you don’t want to go flying off the edge, and one thing that eliminates tiles from use. In fact, in practical terms, fire tiles and breakable tiles are usually equivalent. There are differences, but they’re subtle ones. Occasionally, you might want to drop through the space left by crumbled tile to a floor below. Also, remember that some levels loop around so that you can reach the flipside of your tiles. The reverse of a lit fire can be a normal tile; the reverse of a hole where a breakable tile used to be can’t.

Cluster 3, “Toggled Blocks”, is sort of an exception to the two-new-things-per-cluster pattern: the one new concept introduced here is switch tiles that turn things on and off. (Each togglable element can start off in either state, and if there are multiple switches, each switch toggles everything.) It’s still treated as introducing two elements, though, both of which are toggled by the switches: tiles that appear and disappear, and tiles with spikes that retract and extend. As with the breakable and fire tiles, these are mostly equivalent, just two different types of obstacle, except on levels where you need to pass through a disappearing tile space. But then, it seems like the levels here employ that trick a lot more. One notable thing about the switches: they’re the first element such that touching them can alter the board in a good way. Sometimes you roll onto a switch and immediately want to press it again to get things back into the state they were in before. It took me a while to realize that you didn’t need to roll off the switch and back on to accomplish this: you can do it by jumping in place.

After that, we get “Shifting Sands”: teleporters and sand tiles. Teleporters come in pairs, each sending you to the other — in fact, I’m not sure it’s possible for a level to have more than two. At any rate, they don’t seem like a very interesting trick to me. They’re not so much a thing that creates new puzzle opportunities as a convenience for the level designer, a way to avoid figuring out how to create a physical connection that doesn’t interfere with the puzzle. Sand tiles, now, that’s a thing to build puzzles from. The metaphor here seems to be drawn from sand traps in golf, which can only be got out of with great force. Likewise, you can’t roll out of a sand tile here: you can only jump out of them, leaping over the tile between. A grid composed entirely out of sand tiles would thus be broken into four mutually-unreachable sub-grids. Throw in a few ordinary tiles to permit crossing over and you have a very unintuitive sort of maze. Navigating sand feels a bit like a Knight’s Tour or something, forced by the rules to skip over the places you really want to go.

The fifth cluster, “Hidden Blocks”, I’ve only glanced at. It introduces tiles that are invisible until you get close to them, and then tiles that are visible from a distance but disappear close up (until you’re actually sitting on them). Both are rendered, when visible, in a way that suggests the reflections of light on glass. I can’t say I really like this addition to the game’s repertoire, adding hidden information to a puzzle type that was getting along without it before, and was otherwise quite good about letting me view the entire structure of each level freely. It may make for difficult puzzles, but it’s a cheap difficulty, and doesn’t make for better or more interesting puzzles.

That covers the first half of the game. After that, the introduction of new elements seems to stop, and the remaining clusters focus on different ways of applying what we’ve already seen. Based on what I’ve said above, I think I’m coming to the conclusion that the elements aren’t really all that varied. Not that this necessarily matters — sometimes a designer can do a lot with combinations of a minimal set of stuff. But it kind of seems like the designers here want to make it seem more varied than it is, which suggests a lack of confidence in their game elements.

Steam had one of its big sales over Thanksgiving, including a different five-dollar “indie [adjective] pack” with multiple games each day for five days. I wound up buying four of them. I really should close this loophole in the Oath. Anyway, now that I have all these games, I feel like I should at least give some of them a try. Finishing Bioshock can wait for the weekend, when I have the concentrated attention to spare. Weekday nights, I’m coming to believe, are for little indie puzzle games with self-contained levels — things where I don’t have to track a lot of state.

Which brings us to Puzzle Dimension. This is a puzzle-platformer about rolling a ball around on a grid, skirting obstacles and collecting flowers; once you have all the flowers on a level, an exit portal opens. Now, I say “rolling a ball”, but that’s only skin-deep. There’s nothing about the mechanics that suggests ball-rolling. What you’re really doing is moving an avatar in discrete steps in cardinal directions, and sometimes jumping over tiles. The gameplay seems designed not for keyboard and mouse, not even for a modern gamepad, but for an Atari joystick, a four-direction controller and one button. (Even the menus don’t recognize the mouse.)

I suppose that’s not quite true. It does support an additional button to toggle “camera mode”, which lets you get an overview of the playfield and rotate it freely to view it from any angle (again using four-direction digital controls to do this). This is important because most levels are intensely three-dimensional, and just grasping the geometry can be the key to solving them. The surface you’re on can go through 45-degree bends; sometimes it’s possible to wind up on the opposite side from where you started. The world is always presented so that you’re upright, but it partakes a little of the same gravity-reversal theme as VVVVVV. In particular, if you roll or jump off the edge of a pathway, you fall straight down — relative to your current orientation. Some levels require you to exploit this.

Also of interest is the use of pixelation. Every tile, and everything on a tile (flowers, teleporters, springboards, etc.), starts off rendered in blocky voxels. They become fully-rendered smooth objects when you move onto or adjacent to them, as if your presence is finishing something rough-hewn. This lets you visually keep track of where you have and haven’t been, and also apparently works into the scoring system (which I haven’t paid much attention to yet). Now, obviously there are a lot of games out there that use deliberate pixelation as a stylistic thing, but I think this is the first time I’ve seen it used to denote a marked state, like italics.