Why character creation is always going to suck
So a couple of years ago I bought Dragon Age Inquisition, even though there was another game, doing pretty much the same open-world RPG schtick only better in pretty much every way (I’m talking about the Witcher 3, no I haven’t played that game yet). Yes, I knew what I was doing, it’s because I was invested in the lore.
So far though, I’ve spent about nine hours playing the game, and I’m pretty certain that half of that has been in the character creator and you know what? I still look ugly as fuck. Being the narcissist that I am, I tried to model the character after myself, and after two abortive tries, I ended up with this.
Yeah. I mean, not necessarily super ugly, just kind of… off. So clearly the Witcher was always going to win this round because there you had professional face-makers spending two years lovingly face-making every handsome scar on that handsome Witcher-face, whereas Dragon Age is going to have normal people who really don’t know what they’re doing, just bumbling around with features like earlobe size and jowls (seriously, what the hell are jowls?).
This is a problem for Dragon Age, but actually, Dragon Age characters are always going to be the post-accident Gerard Depardieu to the Witcher’s George Clooney and the reasons are always going to be more complicated than me being a bumbling amateur and the Witcher people being proffesional Wicher-makers. Like, I totally swear you guys, this is not just me deciding that because I couldn’t make a good character, all character creation is forever the pits.
So yes,, this is not necessarily a knock at the Dragon Age Inquisition character creator. As far as these, things go, it’s a lovely effort, you get to multiple sliders for every feature you can imagine, and then some (fucking Jowls). The problem is that the process of character creation as it stands is always going to be deeply fundamentally flawed. This is true for pretty much every game out there and it has to do with the way that the brain process faces.
To explain why I’m going to have to drop some psychology on you. See, usually, when we look at an object, say a tree, the brain takes the different parts of the image, branches, trunk, leaves, and all that, and then puts them together again, one by one, like a puzzle. When we look at faces, though, it’s different. The brain doesn’t take in eyes, mouth, nose separately, instead, it processes the entire thing in one go. In fact, there’s even a specific brain area tasked with doing only this kind of processing.
This is why police sketches end up looking so ridiculous, because witnesses are asked to describe the face feature by feature, nose, ears, one by one. And they’re astoundingly bad. When people are asked to describe celebrities to police sketch artists, the recognition rate for people who later look at the picture is in the low single digits. As in the low single percentage points. And these are faces like Brad Pitt and Tom Cruise, that both the witness and the judge should be intimately familiar with.
If the above process sounds a bit familiar to you, it should, because this is also how we make characters in video games. The problem is, as we are sliding the size of the nose bridge up and down, our brain has no idea what’s happening, but it knows it doesn’t like it. We may have a rough idea of how we want the face to look in the end, but getting there is just really, really hard. In this case, getting more options may actually be a hindrance rather than a help.
Seriously, what the hell is going in in this video?
So, is there nothing we can do? Are we just doomed to using ugly characters for the rest of our lives? Well, not necessarily. Psychologists have developed tools that allow eyewitnesses to create sketches that are actually pretty accurate. At identifiying celebrities. Which is, you know, a part of the process. Basically, what the most successful approaches do is take a bunch of random features, smash them together to create seven or eight faces and then ask the witnesses to say which faces are the most similar to the person (the celebrity) they saw. The face that is chosen is then used as a baseline to create new faces and the processs is repeated until you get something that’s reasonably accurate.
This is all well and good, but I think the question we’re all really asking is can this be applicable in gaming? And really, I don’t see why it shouldn’t. So hey, Bioware, if you’re read this, and I’m pretty sure you are, you know, get on it.