Design Language: The Portal Paradoxes

[Starting a new design column for Gamasutra, Sinistar co-designer and LucasArts veteran Noah Falstein presents a comprehensive design critique of Valve's acclaimed Portal - from intro to 'Still Alive'.]

Welcome to my new regular feature, Design Language. I've been the design columnist for Game Developer magazine for six years straight -- an eternity in game industry time -- and it's time for a change.

This ongoing feature will cover a range of game design oriented topics. I will continue to discuss principles and rules of good game design, but also interview noted game designers and writers, analyze and discuss games that exemplify interesting aspects of design, and talk about what game developers can learn from them.

And for this first foray, where better to start than with 2007's Game Developers Choice Awards Game of the Year, Portal!

There already many articles in praise of Portal, online and in print. But this game is all the more interesting because at first glance, its success seems based in part on business and development strategies that are risky, or at least contrary to common sense.

Although many aspects of Portal's design are built on solid principles that any designer would be wise to emulate, in many other ways it is the proverbial "exception that proves the rule", and if the games industry is to learn from its success we need to be careful to look beneath the surface. To do that it will of course be necessary to discuss many of the details revealed in the course of play, but if you haven't already played the game, what are you waiting for?

One of the tougher responsibilities of a game designer is to learn to evaluate games taking into account your own biases, but not being overwhelmed by them. To come clean on that score, I do like puzzle games and although I'm not a hard-core physics game junkie, I have always appreciated the realism that good use of physics adds to games.

As an old and jaded game designer, I'm definitely in favor of interesting new game mechanics. I heard about Portal through the YouTube video of Narbacular Drop, the game that was essentially its working prototype, and was immediately intrigued, both by the gameplay shown and by the "bright team of developers makes good" story behind it.

Perhaps most importantly, I really like good writing in games; I loved the Old Man Murray columns and was a huge fan of the writing in Psychonauts, so I was pleased to hear Erik Wolpaw was involved.

They did such a good job making fun of the game industry tendency to destroy untold numbers of innocent crates, that it's only fitting he had a hand in making the player feel remorse over one particular cube in Portal.

So all in all I was probably biased slightly in favor of Portal before I even played it, hoping to be favorably impressed.

But I never expected to be completely blown away.

I finished Portal several months ago, and I still find myself humming the song and smiling from time to time. I can't hear the word "cake" without experience a mild sense of betrayal. And I find myself talking earnestly to the metal base of my desk lamp because it is cube-shaped and heavy enough to merit the term "weighted". OK, I made that last bit up, but I did take the unprecedented (for me) step of replacing my ring tone on my phone with an excerpt from the Portal song. But I get ahead of myself.

A Confluence of Delight

The first, and perhaps least controversial lesson we can learn from Portal is that when several really good elements are smoothly integrated in one game, the result can exceed the sum of its parts. Portal has a very simple high concept that is quite captivating, particularly if you see a short visual example - as I did on YouTube.

It seems likely the basic concept of a teleportation Portal gun, and a game based on using it both as a way to move through environments and to fight enemies, is immediately intriguing and would have piqued publishers' interest even without the "students to superstars" Cinderella story aspect of its origin.

The gameplay concept by itself would have been enough for a fun game. But the Valve team that made Portal has also layered in the most minimalist and yet emotionally engaging story I've seen in years in a game -- perhaps ever.

Certainly I'll go out on a limb and say that the character GlaDOS is the best AI character I've ever encountered -- more convincingly psychotic than HAL, with a more emotionally engaging death than Floyd, and funnier than C3PO and R2D2.

GlaDOS aside, the way the story elements are presented is masterful as well. You can play through the game with absolutely minimal attention to the story elements -- I would venture to guess that if you played with the sound off you'd still have an excellent chance of being able to figure out what you are supposed to do and could enjoy the gameplay alone.

But her character -- as a guide, a friend, and ultimately a great enemy -- adds immensely to my enjoyment. The writing itself is excellent. It would have been typical and easy to play it straight, like HAL in 2001, and have a computer whose machine-like dedication to its mission (testing the portal gun) means that the insignificant test subject human (you) are expendable, but to only reveal that slowly.

To do so with a wonderful sense of dark humor -- and to have the humor be consistent -- and furthermore to evoke an amazingly convincing echo of some sort of dysfunctional romantic relationship gone bad, is nothing short of amazing.

And consider that the character you play, although fleshed out a bit in the backstory, is essentially enough of a cipher that it's very easy to identify with her plight and situation without having to pull out the hoary chestnuts of amnesia or being plucked from a normal life and thrown into a different world, or being an elite super-soldier -- all of which are still being used with depressing frequency in other games.

I mention back story -- that's part of why I loved the way the story was told. There's not much in the game itself, which is beautifully appropriate. But there's plenty online to get more background and detail for those who want it without getting in the way of those who don't with endless exposition. You are dropped right into the game and the storyline just unfolds.

I've seen a lot of heated discussion online about Black Mesa versus Aperture Science and how they fit together. But as someone who just doesn't care, I love not having to sit through a long lecture about it early in the game -- the typical delivery of backstory in most games these days. Through the story, the game reassured me early on that I was on the right track and could trust the design of the game not to jerk me around if I just did what seemed right and kept going.

So the primary, overall reason Portal is so good is that it is made out of elements -- core gameplay, story, humor, even a final song -- that work together harmoniously.

The Play's the Thing

I'll get back to the story elements, but games are first and foremost about interactivity and in this case the core gameplay is fun and engrossing. The basic gameplay elements are also quite minimal, a true mark of excellent game design. I often quote Einstein, who said, "Everything should be made as simple as possible, but no simpler" as guidance in game design, yet it is rare that a game follows that precept as well as Portal. The initial levels are perfectly tuned to just fly by, giving the player a sense of mastery and competence.

In fact, when I listened to the level design commentary after completing levels I was struck by how nearly all the changes and tuning seemed to involve adjusting the learning curve of the game; making things clear, introducing some elements more gradually in order to let players grasp them fully, reminding them with a simple challenge of game mechanisms introduced earlier on just before they are needed again.

It's also worth noting how the simple icons at the start and at key points of each level give you the hints necessary to understand what you need to do without dialog or even text labels -- yet another minimalist touch. Likewise, the type of enemies is severely limited compared to most first person shooters. I love the fact that they resisted the opportunity to add various types of guns and tools and powerups or complex double-jumps or climbing mechanisms that clutter so many examples of the genre.

Ultimately, one clearly positive lesson they followed is one I've heard echoed by such luminaries as Sid Meier and Will Wright; to iterate the design frequently and test it with fresh players over and over again.

This is an important way to keep from tweaking a game to be gradually harder and harder as the same testers grow bored and keep requesting tougher challenges. Project Leader Kim Swift has mentioned how they constantly adjusted the game to make it understandable to beginners.

This is a supremely accessible game -- you may feel challenged at times (I know as a fairly basic action player I had trouble mastering some of the timing required at higher levels) but it was always clear what I had to do, and it never felt far out of my reach. Conversely, it felt tough enough that I had a very strong sense of accomplishment and victory -- what Nicole Lazzaro refers to as Fiero -- when I came up with ways to get through the levels.

In fact, one of my favorite moments in the game was at the intended end of the test sequence when, after carefully training me in how to use the portal gun to avoid dangers, I am led to "certain death" by incineration. This was amply foreshadowed so by that point I was ready for it, and even though the use of the portal gun at that point requires much less precision in timing or aim than immediately previous challenges, I still felt incredibly elated to have dodged it the first time I tried.

The gameplay to that point trained me very well, and the story made it not only obvious that I should expect an even more direct attempt on my life, but also made it supremely satisfying to survive it. I can't remember ever experiencing the same sense of accomplishment from a movie or book.

I see Portal as an example of how games are wonderful at evoking triumphant emotion. It represents a culmination of many aspects of game development -- clean, fun and original game elements, the maturation of storytelling and sophisticated use of humor, and use of physics and fast 3D rendering that just wouldn't have been possible even a few years ago. The game just makes me proud to be a game developer myself.

It may take decades before the games industry enjoys the same level of general respect as an entertainment and art form that film has now, but when that time comes people will point to Portal as, if not our Citizen Kane or Birth of a Nation, certainly our Wizard of Oz.

I've mentioned that the basic gameplay is great, and that its emotional impact is strongly boosted by the great writing and character development of GlaDOS. But I would be remiss if I didn't also touch on the additional "force multiplier" element of the ending credits song, "Still Alive". Ending the game with a song was a brilliant touch, and one that I think other games may well benefit from -- but only if the song is as well crafted on its own and fits as well into the game as Jonathan Coulton's did.

I knew there was a song at the end before I got to it, from some online comments I'd read. Still, I was blown away by how much the song added to the whole experience. It validated my success, made me realize how much GlaDOS's personality both resembled an obsessive human while growing logically out of the basic modules that we learn are part of her programming gestalt.

Just as with the game, it was not only the content of the song that was great, but the presentation. The voice acting, excellent throughout the game, reached a literally pitch-perfect level in the song, making me almost feel sorry for GlaDOS. Almost. And the running text commentary on the song kept it light and funny. Even the in-jokes were perfect. The reference to being out of Beta and releasing on time could logically have referred to the portal gun, but it was a great laugh aimed at fellow game developers.

The sweet, psychotic love song aspects were right on target too. Having finished the game, I had to admit that I did break her heart and kill her -- and tore her to pieces -- and threw every piece into a fire. And I'd do it again, you crazy sadist! I loved how the song gave closure of sorts -- some people debate whether you are dead at the end of the game, but the song makes it clear that not only do you survive, but GlaDOS seems likely to return in a sequel eventually.

For that matter, I expect to read a good psychology master's thesis on GlaDOS's condition someday soon. And on top of all that, it was a catchy tune, it rhymed here and there, and it scanned well -- a minor detail that even most songwriters can't handle even without the burden of tying their song to a game. This was a triumph!

Lessons for Developers

So what about those other risky conclusions I warned about at the start of this article? One of the tenuous assumptions one can make from Portal's success is that the best way to make it big in the game industry is to attend a game development school and create a school project that will instantly transport you (as if through a trans-dimensional portal?) to fame and success.

But consider how many things had to go right for the Narbacular Drop student project to get picked up by Valve. It had to be presented to someone with the clout to hire the team on the spot, at a point where the game was far enough along to show its promise, but before the team had graduated and moved on to possible separate destinies.

It certainly helped that the company was in the same town as the team already, and was already known for excellent first-person shooter games with unusual weapons.

And perhaps most unusual was the fact that Valve has not only committed to online distribution, but also was willing to place this small and atypically short game in their Orange Box, where it could benefit greatly from being packaged with other more proven AAA properties. This happy set of coincidences is not going to happen very often. I don't want to discourage school teams from dreaming big, but counting on success like this is about as reliable as playing the lottery.

Another happy accident was matching Erik Wolpaw with the Portal team. They didn't work with him specifically to add humor to the game -- at GDC 2008 Wolpaw said that the humor more or less crept in because that was what he was comfortable with.

It's fairly easy to imagine a Portal without humor, with just the standard resolute main character forging ahead silently and relentlessly, but the infusion of ironic humor makes Portal the opposite of relentless. Good humorous writing is very hard and there are very few practitioners with a track record out there -- it's great that in this case the resource was applied just right.

Writing is Not a Dessert Topping

And the final surprising irony of the production of Portal may be that after years of countless writers and designers clamoring to get the writing expertise into the game early to avoid the superficial and/or confusing kludge that so many game stories can become, Portal managed to add the writing relatively late in the process, after much of the gameplay was well along, and do incredibly well with it. I fear producers and publishers everywhere will say, "See, we can just add the writing in during the last few months! It'll be fine."

But looking more closely, I think that the circumstances of Portal once again show being in the right place at the right time. First, there is only really one substantial character in Portal, without a huge amount of dialog.

A typical action-adventure game character might have dozens of times as many total lines and will often interact with many other characters, and an RPG will typically have many times even that number of dialog lines -- you just can't cram that all in late in production and have it work well.

Also, the simplicity of Portal overall lent itself well to a sparse approach. But perhaps most fortuitously, the fairly sterile and sparse lab-rat's-eye-view of the world in Portal very much needed the lighter touch of the dialog to turn what would otherwise been a foreboding and cold experience into something joyful. I'd say it was like adding frosting to a cake -- but, of course, the cake is a lie.

The bottom line is that this is a game that should be enjoyed, and then carefully studied by any serious student of game design. But don't be too swift to generalize any lessons learned without considering the circumstances that made Portal so unique a production story.

Like many masterpieces, it may stand alone -- although I hope it inspires many others to create games that are equally as fresh, fun, and exciting. Or to paraphrase GLaDOS herself, look at me still talking when there's game design to do!