My bridge sways precariously in the wind, the distant shore I’m trying to reach still a seemingly insurmountable distance away. Back on the near shore is its precarious perch, which, although I’ve done my best to anchor it firmly against the ground and provide a counterweight against the swaying main arm, the slightest move or the next gust of wind threatens to bring the whole thing colliding into the spinning wheel of doom that will start a chain reaction of collapse. And meanwhile, both I and my construction material are quite literally staring into the abyss.
Before Minecraft, before smartphones, and before the mass acceptance and dominance of Steam, World of Goo appeared on the internet as a prototype: a product of a student project from Carnage Mellon’s Experimental Gameplay workshop, which two years later would go on to a full release in 2008 and become one of the first high-profile indie success stories of the decade. The obvious superlatives have since been well documented over the course of its history (and multiple ports): charming art style, excellent soundtrack, quirky and unique game mechanic, multiple IGF and other awards. To repeat them would be superfluous. This is not a review of the game, but the first in what I hope will be a series (if an irregular one) of analysis from the view of a student, aspiring to be a Famous Game Designer, trying to tease out some of the specific elements that make the game work so well; to find traces of the magic “something” that makes the game tick. That I like the game in question should generally be pre-supposed: while there is value in playing and learning from other’s mistakes, that subject is well trodden by other writers (case in point, the well established “Bad Designer, No Twinkie!” series) and nailing down the concrete things that make great gameplay is a lot harder that pointing out what makes it bad.
So here we go.
Physics-based gameplay is nothing new. At the time World of Goo came out, Half-Life 2 had lately made “physics puzzles” an industry buzzword. Even the market for physics-based gameplay featuring amorphous goo balls had been previously blazed, with 2005’s indie platformer, Gish. And of course, everyone and their publisher in the big budget side of the industry was touting the flash and show side of physics: whose ragdolls could collapse the most convincingly and whose engines did the best job of hurling explosion debris in a realistic, real-time calculated manner.
The significance of World of Goo’s approach to physics gameplay didn’t really hit me until recently. I was replaying the game on a tablet (thanks in part to the Humble Bundle, I’m one of those people who’s effectively bought the game twice). The actual platform is irrelevant but the fact it was a repeat play-through was important. It had been a few years since I last touched it on the PC, where I played it twice. For most games, I’ve had my fill after beating them once. Yet, it struck me that upon seeing a familiar level, I could remember the basics of what sort of structure I had to make to complete the level, yet that foreknowledge didn’t make the level predictable or easy. Manipulating the constantly shifting goo towers was a constant, chaotic balancing act. Having foreknowledge of the challenge only made it less likely that I would make especially dumb mistakes or more likely that with repeated effort, I might be able to fulfill the level’s “Obsessive Completion Distinction” criteria. Instead of a puzzle with a fixed outcome (put enough weight on one end of the see-saw and it will hold you up so you can access the next area), the number of variations on the solution approaches infinity, simply because the designers chose to implement physics the way they did.
But wait, you might say. Isn’t physics by its nature deterministic? Well yes...but while the computer program can calculate the angle of a bouncing ball to a very fine degree, the human player can only make a very coarse estimate. Anyone who’s ever watched a ball bounce down through a pachinko machine knows the slim odds of being able to predict exactly where the ball is going to add up. Now consider that the current trajectory of that swaying goo tower is being recalculated multiple times per second.
The end result is dynamism in spades. No play-through of World of Goo will be exactly like any other. Every level solution is “of the moment”: shoring up a weak spot here, adding a counterbalance there, quick, move the balloon goo further out to redistribute the weight away from the spikes of doom, oh no, that one is drooping too close to the fire.
Other features I found worth noticing: the game offers a limited “undo” function which will inevitably be called on to avert a total and fatal collapse, and it is probably necessary; having to do a full reset as often as you would otherwise, in the face of considerably unpredictability, would be particularly frustrating. Many puzzle designers only dream of puzzles that open ended. To mix things up, World of Goo offers an evolving set of both obstacles and unique goo balls with different physical properties, while most of the time reinforcing the core mechanic.. There are a couple levels where the game even switches away from goo balls to stacking blocks. While they work well enough, in the same way that Jenga works as game, discarding the elasticity of the goo balls looses something, I think, in the transition. Figuring out the balance point of rigid objects with considerable weight and inertia is noticeably easier, I guess.
Far be it from me to say that this is a “best” or “only” way to use physics as gameplay, but it is a very interesting data point to compare other games to. Since we already mentioned Half Life, let’s talk about it. Half-Life 2’s physics gameplay worked because it shifted the player’s focus to working with and manipulating the environment in a vaguely realistic way, which complemented the design of its protagonist as an “action survivor”. If I use the words “gravity gun” and “Ravenholm” in the same sentence, anyone who played the game will probably remember a bit of what I’m talking about. But while the application plays out differently, Ravenholm is also an example of dynamism at work in physics gameplay: balanced to be short on traditional ammo, the player has to think on the fly to decide which of the Havok-enabled physics objects rattling about the level will be the best deterrent to the headcrab zombie horde. Even at the very root of the video game’s origins, you have Pong...the physics are basic enough that Pong clones are a very suitable student exercise, but the human player will be kept continually adapting to the momentum of the bouncing ball.
Is there anything wrong with including physics simulation just for the sake of showing off? No. Nor am I myself particularly likely to design a game around physics simulation; I’m much more of a worldbuilding fan. Do I find it fascinating? Absolutely. Will these observations be of use to someone else looking to find the secrets of the craft? I certainly hope so. Considering physics in this light opens up a whole realm of possibilities for a designer and it would be a shame to dismiss them unthinkingly.