Analysis: Game Design Accessibility Matters
[In this in-depth design exploration, Canabalt creator Adam Saltsman lays out some essential, practical principles for making video games more accessible -- without dumbing them down.]
On the weekends my wife and I go visit our friends to hang out and play games. Sadly, this very rarely means video games. Usually it means board games -- stuff like Pandemic or Settlers, not Candyland. They're geeky, complex, exciting, difficult games.
So how come we almost never play video games, if our friends are into geeky, complex, exciting stuff? I think it might be because all video games are unnecessarily intimidating and cater to a very highly developed set of tastes and dialects. This is not a radical idea by any means, but it got me thinking about specific ways to modulate or mediate these exclusive, discouraging aspects of games.
When I talk about accessibility I am talking about removing the sense of frustration that lots of people feel when we try to introduce them to these things that we love: our Castlevanias, our Quakes, even our Shadows of the Colossus. To most of humanity, these games are unplayably hard, but I feel a lot of noise has been made in the past about content being the source of the problem -- that these games require math, or that they feature geeky themes, or that they're juvenile and misogynistic.
I like elves, space marines, and breasts as much as the next guy, but I don't think those are the root of the problem. These things exist in books and film and are devoured by people that shun games. What I hope to do with this article is point out some things that I think are at the root of the problem.
But why bother making accessible games? People have been making these "inaccessible" games for a long time and our industry is thriving. Lately, though, it seems as though games are starting to flirt with some fascinating ideas and themes -- and, as Hal Barwood puts it, "Shakespeare wrote hits." If we are going to make art, why do we keep making it for the same tiny subsection of humanity? Have you met those guys?
Even if you're not a passionate artist trying to express yourself through this new medium, there are some obvious financial benefits to seeking out a new audience (see Nintendo and PopCap). Accessibility is important.
I'm not interested in making games easier, or dumber, or more boring. While Nintendo and PopCap have seen incredible financial success courting casual gamers, much of what they offer is repellent, condescending, boring, insipid, and unfair. We are frequently presented with a choice: you can either play a casual game (and I use the term pejoratively) with your non-gamer friends, or you can play a really rad game by yourself, or you can play board games. I want to find that middle ground where multiplayer video games are as inclusive, fun, and complex as board games.
This isn't a misguided anti-hardcore crusade. Games like Street Fighter IV, God Hand, and Ikaruga are spectacular, inspirational works that could be worsened by taking any of the unsolicited advice I am about to dispense. But they are also really specifically designed for a very small group of people, and the barrier to entry for these games is incredibly high. How many people dismiss Street Fighter as a button-masher because they can't get over that initial hump?
Let's find the middle ground. Let's make cool games that we can play with our friends, even if they don't know what a Halo or a FADC is.
Gravity & Complex Simulations
I've written a little bit in the past about how I am suspicious that gravity and parabolic motion has a tendency to intimidate non-gamers. Games my mom and wife play (Katamari Damacy, Flower, Secret of Mana, Dr. Mario) don't include real-time gravity as an important part of gameplay. Some phenomenally well-designed games have managed to include gravity and still draw in a huge non-gamer audience (Super Mario Bros.) but they are the exceptions, not the rule.
I suspect it is not gravity specifically, but rather complex nonlinear physics simulation generally that is problematic. It requires a certain amount of predictive visualization to interact with it, even in a simple way, and this is an acquired skill, not something at which we naturally excel. A lot of games combine these systems with strict punishments like death, which can be frustrating and intimidating.
Complex simulations are a cornerstone of game design, and we cannot nor should not abandon them. However, they can be vicious when exacerbated by overall game speed or strict punishments. Super Mario Galaxy and Offroad Velociraptor Safari are games that use complex 3D physics in different ways, but neither game really punishes the player for bad predictions.
If you are reading essays on Gamasutra, you are probably not intimidated or confused by the D-pad and double face button combination. Most of us have moved on to a combination of a D-pad, two sticks, four face buttons, four shoulder buttons, three center buttons, and an accelerometer. This is a problem, since the 20-year-old D-pad and two face buttons setup is beyond the comfort zone of most humans.
Witness the popularity of Rock Band, Wii Sports, and Wii Fit if you need immediate, obvious, physical evidence. It's not weird or rare for non-gamers to have to look down and make sure they're hitting the correct arrow or the correct face button. Like all the things I am going to bring up in this article, this is not necessarily a fatal flaw, but coupled with some of the other things on this list it can be a real problem.
Simplify input, bank on context, and don't be afraid to provide subtle or optional visual reminders in-game. See Batman: Arkham Asylum or any recent 3D Legend of Zelda offering for examples.
Speed & Reflexes
Like simulations, relatively fast real-time inputs and reactions make video games what they are. They are an essential ingredient. Interaction is a pretty special thing in our art form. But requiring really fast reactions for basic interactions automatically excludes a huge group of people.
I think one reason board games are so much more accessible than video games is that they're turn-based, not real-time. This may be common knowledge, but a basic way of actuating the difficulty of your game without changing damage ratios or level design is to simply change how fast the enemies and projectiles move.
Another facet of reflexes is "telegraphing," the tendency or ability of the game to warn you about what is coming up. For example, a boss might blink or strike a specific pose before attacking the player. Shinji Mikami (Resident Evil, Godhand) is an expert at tuning and manipulating telegraphing.
If this sounds too weird or separate to thinking about speed, it may help to think of telegraphing as a modifier for speed. The farther you can see ahead, the faster you can safely travel. This is why cars have headlights. It is a really simple, fundamental thing that is frequently ignored in game design.
The speed of objects and simulations has a tendency to dramatically exacerbate other accessibility problems like simulations and input. For example, if the player doesn't have the controls memorized, and doesn't have time to look down at the buttons, he or she will not enjoy the play experience.
I think it's interesting that movies, plays, operas, TV shows, popular board games, and card games all take roughly one to three hours to consume, but any video game under 10 hours is maligned and considered less substantial than its longer counterparts. This may be less true than it once was; Spelunky and Captain Forever are exploring shorter, replay-based adventures in a really positive way.
That's not to say there is no place for long-form adventures, but the idea that a 30- or 40- hour experience is a standard to strive for strikes me as a little ambitious.
While it may be more correlation than causation, I think it's interesting that a lot of mainstream media is designed to fill up a free slot in an evening, rather than devour an entire weekend (or series of weekends).
This gets a little technical, but I've seen it crop up too often to ignore it. Bounding rectangles or other shapes are how game systems view game objects. To us, Mario looks like a fat dude in overalls, but to the Nintendo system, he is a box. When he crouches, he is a shorter box. This is how the game can figure out quickly and easily whether or not you are hitting a wall or not.
It's really easy and intuitive to just set your bounding box to the maximums of your graphic or mesh. For example, if your character is 20 units wide by 40 units tall, you might set the bounding box to 20x40. Makes sense, right? Unfortunately this leaves very, very little room for error or flexibility. If your graphic doesn't extend all the way to the corners of that box, it can also look like your character or object is actually colliding with nothing, which can be really frustrating when you die from hitting spikes that you didn't even touch.
Shrinking bounding boxes is not a pre-made solution, though. Environmental content especially benefits from a really authentic bounding volume. Projectiles fired at the player are more forgiving when they have a small bounding volume, while projectiles fired at enemies are more forgiving if they have a larger bounding box though. Look no further than the Japanese shmup scene for constant, gleeful, and expert manipulation of bounding boxes.
Carefully tuned bounding boxes lead to more close calls, more player satisfaction, fewer "cheap" deaths, less frustration, and more accessibility.
Positive permanence was really popular back in the 1980s and has been totally lost, much to my regret. That's my own made-up phrase for something two of my favorite games, Metroid and The Legend of Zelda, employed to great effect.
Rather than providing save points or quick-save buttons, or auto-saving after boss fights, lots of state changes in Metroid and Zelda are permanent. Every time you unlock a door in either game, it is unlocked forever. Even if you die, it stays unlocked. Picked up a rad item but died at the boss? No problem, you still have the item.
When you die, you are placed back at the entrance of the region -- not all progress is permanent. And enemies constantly respawn in almost every area. But the big things -- the important things -- are not thrown out when you die from some unforeseen danger. You are never placed in a situation where you think, "Oh nice, I got the bombs!" and die a minute later, screaming, "FUCK YOU WHERE ARE MY BOMBS?!"
A nice side effect of this system is it helps modulate the difficulty of game areas. If you're a total rock star, you can Rambo through the dungeon in one go. But if you are not up to it, or have slower reflexes, you can run suicide missions into the area -- unlocking a door here, collecting an item there -- until you've learned enough and powered up enough to handle that region. It serves the same essential function as grinding and experience points, but it is skill-based, not time-based. Grinding can work, but it's a less elegant solution to the problem of modulating the difficulty of your game areas.
Don't take away people's accomplishments and victories! It creates short-term frustration and long-term accessibility problems. Modern grind-and-save systems serve the same purpose but they are clumsy and painful.
The Guide Player
Most board games serve two players or more, and without fail one or two of end up spearheading the learning process by reading the rules and helping explain the gameplay to other players, who may not be as bold.
This is something many of your favorite co-op games (Halo, New Super Mario Bros. Wii, Left 4 Dead) already do, only in the most basic way: They have respawn systems that are oriented around regenerating the bad players near the good player. This is a very practical and common compromise for controlling player proximity, but it has the nice side effect of orienting overall progress around whatever player happens to survive longest at any given time.
I would love to see this concept pushed further. You should be able to help your friends get to areas that are hard for them to access, share items with each other, and even temporarily switch roles.
Assistance systems are a fundamental cornerstone of real-time multiplayer gaming and they are largely ignored and unexplored. Skillful players need more stuff to do so they don't get bored, and less skilled players could definitely use the help.
This isn't so much a specific tip about game design as it is an overarching philosophy of accessible design. Layering is the idea that you can present your audience with both a simple interface and a hard interface at the same time. Layering is how you let skilled players advance more quickly than unskilled players without ruining the game for anyone. In many ways, the ideas of having a guide player and ideas about positive permanence are just different ways of implementing layering.
The easiest way to think about layering in game design is input. Imagine a simple scenario where a game character has two attacks. One is quite weak but very simple to input, and the other does a lot of damage but is hard to do. This is layering! When you first pick up the game and it is new to you, you might only interact with it in a simple way, but as you grow in physical input skill you can interact with it in more complex ways and advance through the game more quickly.
Obviously, the wonderful thing about layering is that you're not excluding anyone. The guide player can help move things along, noobs can get their learn on, and neither will feel the game is too intimidating or too easy. Of course, this isn't nearly as easy to implement as it is to talk about. Too bad! It's absolutely critical to creating games that can appeal to both loyal players and new blood.
The old ways of doing this (difficulty levels, experience points, grinding) require a lot of boring, grunt design labor and testing, and ultimately comprise merely a slightly finer mesh for dividing your players.
Layering is absolutely not about making players unlock better moves or more advanced gear. This is exactly the opposite of layering; this is serial, not parallel. You can only embrace a wide audience by giving them an implicit choice all the time. If everyone has to use the same crappy moves on the first level, then you don't get a guide player, you don't reward people who like the challenge of skill-building, and you are losing your players.
Always provide an alternate easy way to play, even if it is less stylish.
I am suspicious about catch-all solutions and magical cures for game design problems, since game design is such a nebulous and unexplored art form. I really can't make any spectacular claims about how taking my advice will somehow make your project a bestseller or, better yet, a dearly-loved cult hit.
But if it can at least spark a discussion about the worth of accessibility, or get us to take a second look at our work, maybe we can start finding a new, larger audience for these things we're making. When our Shakespeare arrives, maybe he will be able to write a hit.
[Adam “Atomic” Saltsman is an independent game designer, artist, programmer, and entrepreneur whose past projects include Canabalt, Cave Story Wii, Gravity Hook, Paper Moon, and Owl Country, as well as the Flixel Flash game framework. He is also the co-founder of Semi Secret Software (wurdle) and the director of Last Chance Media (Dr. Dobb's Challenge 2).]