Curtain

What role do you take on when you go to see a movie? What is your relationship to the characters onscreen? You are always part of the viewing audience, watching the events unfold. Likewise, when you read a book, you are always the reader-- the interpreter, perhaps, but you maintain a consistent, detached role. You are outside of the story, observing the events unfold.

When playing a game, you are always the player, but your relationship to the characters onscreen can vary greatly from one title to another. What role do you play? Do you inhabit the character itself? Are you an abstraction of the character's will? Are you just the guy outside the screen, playing a video game? Or do you inhabit some vague omniscient third precence in between? Probably the most concrete metric of the player/character relationship is the UI, and the raw quantity being measured is the flow and availability of information. What does the character know, and what does the player know, and when, and how? I'm only talking about human-level, single-protagonist games here; obviously guiding the rise of the Roman empire or fostering a species as it evolves is on a completely different level. And I'm not talking about the simple difference between first and third person. The player/character relationship is only partly determined by the position of the camera.

A couple of the games I've been playing have informed my recent thinking on the player/character thing: Condemned: Criminal Origins and Hitman: Blood Money. Condemned is first-person, while Blood Money is primarily third person with an optional first-person toggle. What got me started thinking down this track was the Blood Money HUD. Blood Money is a game about a super-assassin who works by infiltrating heavily populated sites to eliminate his target by any means necessary. As such, the protagonist, Agent 47, often disguises himself to pass undetected through restricted areas. The HUD breaks down into four persistent readouts:(click to enlarge)

1) Contextual action menu

2) Compass
3) Health
4) "Tension meter"

The first three are pretty standard. The tension meter is the element that stands out. Being discovered by hostile forces isn't a binary event in Blood Money; you can tell when guards or passersby are starting to freak out by observing the tension meter, and quickly backing off or silencing any nearby witnesses to avoid your cover being blown. The question is, where does the raw info for this meter come from? Obviously on a code level it's drawn from the AI states of the NPCs in 47's vacinity, conveyed graphically. But in the context of the game, is this a representation of 47's perception of the people around him? If so, it seems too accurate, and also takes into account the states of people outside 47's range of vision. If the info comes from an omniscient third entity-- 'the computer'-- is that info also available to 47, the character? Or is it only available to the player? If the info on the tension meter is only available to the player--if the player knows something that the character onscreen doesn't-- it serves to separate the character, the computer, and the player into three distinct entities. "I am the player; I am observing the information provided by the computer; this informs how I direct the character in the video game." Any HUD element that draws the player's locus of attention out of the character's gameworld, and onto a graphical element that exists only in the player's world, breaks the player's experience into something more artificial than it might be.

I believe it's possible for the player to 'inhabit' a character in the third person. For one, the player identifies with the character's actions through the familiar act of visualizing their own physicality in everyday life. Right now, you probably can't see anything of yourself but your hands on a mouse or keyboard, but you know how you're sitting, what your posture looks like. When you walk, or run, or duck down, you can feel how that must look and can easily picture your own performance of these actions. I believe that the player inhabits a character in the third person by projecting their own sense-memory onto the character's actions, and treating the wider camera as a surrogate for the peripheral vision missing from the first-person experience. As such, for the player to maintain inhabitance of the character, everything the player knows about the gameworld must be directly observable through the physical gameworld itself; everything the player knows, the character must also plausibly be able to know. So, in Blood Money, the player's knowledge of surrounding guards' tension levels would ideally be dictated by directly observing their reactions to 47's actions, their body language and vocalizations, the sounds of movement in the next room. The player would need to swivel the camera around and carefully observe the results of their actions within the gameworld to gauge how much leeway they had in their actions at any given moment; likewise, the game's designers would need to ensure that NPCs' AI states were realiably readable through consistent body language signals and vocalizations depending upon their current state. In this way, the act of determining nearby NPCs' tension levels would no longer entail the player removing his attention from the gameworld to check a meter that only exists as an artificial construct on the screen, but instead to more closely examine the gameworld itself, drawing him further into the inhabitance of the character onscreen.

Another common HUD element in action games is the ammo readout. At a glance, the player can see how many shots they have left in their gun, and how much ammo they have left total in store. But who's keeping track of these things? Is the character counting how many dozens of shots have been fired from their gun, and how many remain? It seems unlikely this is a function of graphically depicting the character's perception, and instead is being conveyed the the computer, showing the player information the character doesn't actually know. Condemned addresses the old ammo readout issue, but only goes halfway:


There is no persistent ammo readout onscreen. The player must input a specific keypress to physically check their ammo count if they lose track (or when they first pick up the gun.) However, upon doing so, a graphical ammo readout appears briefly, and the count is not physically observable by the player. In the screenshot above, would you be able to tell how many bullets are in the magazine the character is holding? No, you must again break out of observing the gameworld, and shift your locus of attention to an onscreen readout. This seems easily avoidable. I mean, it's a problem that's been solved by firearm manufacturers in the real world:
In Condemned, the problem is almost solved in game design terms, but then forces the player to rely on a HUD element instead of his own observation, breaking him out of the gameworld.

I believe that the ideal solution is a game that exists primarily in the third person, but allows first person when useful (such as, for instance, when checking the ammunition in a gun's magazine.) The most important thing is that all information available to the player be physically observable, and equally available to the player character. Being in third person only facilitates this, by allowing simultaneous observation of the player character's state and the gameworld surrounding it. This also gives the designer a means of eliminating the mainstay of video game HUDs: the health meter. If the player is constantly observing the state of his character, and is capable of swiveling the camera around him at any time to examine him from every angle, the character's physical state can be reliably observed without a graphical readout. The character must react to damage incrementally, gradually slumping more, slowing down, visibly bleeding from wounds, shaking, dragging injured limbs, etc. This is less easily quantifiable than a numerical value displayed onscreen or a graphical health bar, less exact, but more meaningful to the player. I find it more useful to see that my character is physically limping
and bleeding than to read that my character has "26 Health" while still acting as they would with 82 Health or 100 Health. I believe that we can do without acknowledging the computer as an entity of interpretation, that we can completely omit the man behind the curtain from the experience. What should be important is making the player's experience as congruous with his character's as possible-- creating a seamless player/character relationship. This is HUD-elimination not in the pursuit of the cinematic. When the designer remains aware of the gap between character perception and player perception, and concentrates on eliminating it, the game experience itself becomes that much more pure. It just makes sense.