Tuesday, November 5, 2013

Game Design, and Beating Yourself Conscious

There's a lot about what we're doing with Sinister Dexter that makes me really
proud of our team, and I'd like to discuss one challenging angle. We're trying to deal with social issues of race and normativity in a game that technically needs neither graphics nor a story. We're trying to do that in a game with literally zero budget, a tight timeline, and - as chance would have it - a team made up mostly of white men*. Yeah, there's a lot of ways to mess this up.

*Note: Fixer Studios has no formal recruiting policies around race, age, gender, nor sexuality. We're small, and recruiting the most talented friends willing to donate time as independent contractors who may never get more than $10 each. We'd take any talented person of any demographics and haven't intentionally restricted our hiring in any way. I'd love to bring in more diversity and talent period, so please let me know if you can recommend someone that brings either or both!

In 2002, when I first started started getting paid to work with games, I was a fledgeling academic. I worked at Simon Fraser University's School of Communication and tried to put my few years of studying the social sciences to work in thinking about games. From 2004-2006 I did my MA in Communication at UW while I was researching and writing about issues of semiotics, political economy, gender, and violence in games. These years were spent somewhat frustrated with the industry for 'their understandable but tragic failure to comprehensively think about the messages they were embedding'.

Since I started working directly in the industry in 2007, I've try to keep my idealistic ivory tower thinking despite the realities and pressure of making games work. I've had mixed results, because these jobs can leave little room for the required navel gazing and even less room to grow authority and take appropriate action. Here, at Fixer Studios, I have that level of responsibility and a team that supports doing the right thing above anything else.

Our first playable build had three characters sketched and coloured well enough to use. One happened to be dark skinned, one light skinned, and one green skinned. The hand icons for the game had just been finished as black-and-white line drawings and we were at our first crossroads.

I could not morally show a game at PAX in which characters have whitewashed hands; even if that's simply due to them being outline drawings instead of actually shaded as caucasian. I requested that our artists produce three skin tones of hands for every hand icon in the game, at every resolution.

Mobile developers, that's your cue to say, "But... the file weight of your sprite sheets is going to get huge! Make enough decisions like that and how will you keep your game below the magical 50MB size limit for over-the-air downloads from the app store?" Yeah, that problem was on my mind before I even opened my mouth. Building a time-consuming lazy load system was out of the question for a team with one full-time engineer...

My initial hunch was that we could create one of each gesture icon at each different asset size in a sort of grayscale, then we could apply a specially masked-off, semi-transparent skin tone layer above those gestures in-engine. We'd probably need two of these layers, I was thinking, because there's some shading involved as well. After all, if we used grayscale shaded hands as the base and then we applied a skin tone over top - well, pink plus gray equals zombie... not European.

An artist and engineer spent a lot of their time working this out in Photoshop and in engine, both trying several revisions before landing on a solution of "sepiascale" hands instead of grayscale. It works pretty well for human-coloured hands and allows us a variety of skin tones while keeping file size down. It was a bit of investment to get right, but it had to be done.

Beside all that, we've also committed to attempt having a broad range of races represented in our game's characters plus try to break up race norms in storytelling. At every step of the way we're pushing ourselves to ask, "Is that character bringing diversity to the table? Are they breaking storytelling norms? Are we doing this in a way that also enriches the story and adds depth? Why is that character of that ethnicity and not another? How is this going to work geopolitically for localization and globalization?"

Meanwhile, our Associate Designer raised some good points and challenged the team to push ourselves even further down the path of doing what's right. Three human skin tones may push people into subcategories that don't work. Also, what about faces? Just because our team has default conceptions for the skin tones of different characters' faces doesn't mean we should push those conceptions on our players. He challenged us to have five different, player-selectable human skin tones applicable to all character faces and hands. We're also evaluating having up to five non-human skin tones in game, and those pose even more challenges. For instance, putting blue over sepia doesn't look any better than putting pink over grey, so our sepiascale solution doesn't work there.

For a tiny game in which having any characters at all was an open question for a long time, well, that's ambitious. It may also be the right thing to do, and damn all the extra time the team spends working on this at the cost of something else. Character portraits are a lot harder to handle masking layers for, and this likely involves applying separate hair colours as well... all in a game with no character customization engine.

There's additional tough questions too. How do we deal sensitively with the idea of taking a particular face shape and making it work with skin tones that weren't part of the original art design of that character? Besides, there's a lot more to race than skin colour; does offering a palette-swap actually let players choose our characters' racial identities for themselves, or could it be construed as an offensive oversimplification of race? Uh-oh. I get nervous just thinking about it, and I guess that's the point.

Heck, here's a less real-world example I'm actually struggling with. We have a cat-person character concept that I was going to cast as Tarot's "The Fool", our protagonist's loyal, naive and awkward best friend who is clearly marked out as different than the rest of the cast, (like C3P0, Samwise Gamgee, Hugo "Hurley" Reyes, or Ron Weasley). However, I personally see different things in cats and it's crucial for the gameplay experience that our players are able to get a sense for this character at a mere glance while ignoring text.

Given that most people are predisposed to think cats exemplify unreliability, fickleness and unpredictability... I'm not sure this cat character can work as The Fool and I've recast them as Tarot's "The Wheel of Fortune" (like The Millenium Falcon, Gollum, Charlie Pace, Hedwig). Here I'm changing a writing decision to fit people's preconceived notions about the values of a character based on how they look. Is that decision discriminatory against cats?

You can see how this sort of thing can really slow down development, and there's a constant trade-off between trying to tell a story visually, get a game done on time, and simultaneously be an exemplar for socially conscious game design. I haven't even gotten into our goals for how we treat gender and sexuality in our game... and we're just a tiny mobile game in which story is almost peripheral.

I guess what I'm saying is that I can't promise that Sinister Dexter will make all the right decisions around how to treat race in our game, but I do promise that we're going to keep thinking about it and we're going to try.


  1. Knowing nothing about the game, it's hard to leave any 'Practical' feedback. Obviously presenting a socially conscious face is something to be commended... but putting this work into a feature that is not 'actually' part of the game experience is questionable, especially when its impact is potentially nonexistent. I don't know that it 'is', I'm merely stating that if say the character portraits / hand icons in the game could be replaced by inanimate objects without affecting the game experience, then the differentiation may not be practical to implement. Perhaps a better way to send a message in 'that' case would be to have a 'different' (Not a 'White Male') race / gender for the main character, as that's bucking the trend more than allowing for a choose your color system. Again, this is a blind comment, so I may be approaching it from the wrong perspective all together.

    1. Thanks so much for sharing your feedback, and I personally tend to agree with all that you're saying. We sort of stumbled across this issue like you might in a first-person shooter. All you might see of yourself is your hands, and whatever "skin" your opponent has doesn't change their hitboxes or headsize for headshots... so it's all superficial. That said, if we're going to have faces in the game we wanted to try to buck the trend on some stuff.

      I probably shouldn't share this yet, as it may change... but the current plan is to have a young black woman as the main protagonist. We chose the story and character personalities before we selected art to match them, and she simply had the "impetuous, innocent, self-doubting but relentless in the search for knowledge" look about her face. We're making similar choices for other characters right now, and I'm proud of that process. I'm looking forward to sharing more, and apologize if this shifts around for story reasons I hope you and other readers will understand.