A few years ago, Patrick Redding, then Far Cry 2’s narrative designer, and I discussed how to evaluate and interpret the meaning of a player’s input into the system we were designing. We had engineered a system wherein character relationships would evolve over time based on player actions, but the player’s input into the system amounted to either doing missions for people or killing them. Pat came up with a name that described this scenario: the fat pipe-thin pipe (FPTP) problem.
Stated simply, the FPTP problem is an issue of discrepancy between the bandwidth a game uses to communicate to the player and the bandwidth the player has to communicate back. A game’s capacity to output rich, nuanced information exceeds that of film or television, yet a player’s capacity to reply with equivalently rich and nuanced statements is massively constrained by our input devices and our game designs. In a sense, from the perspective of a game, players would appear to suffer from some extreme form of autism; our inputs suggest that we take the game’s output at such a literal surface level that we appear to either not understand or not receive all the cues the game gives us.
Games are fundamentally about dynamics, and dynamics are about dialogue. In singleplayer games, the dialogue is between you and the game system. So in these games, game and player haggle over the value of positions in the state space. If the capacity for one side to make sophisticated arguments about what the game means is massively constrained, then almost by definition the resultant meaning is of low fidelity.
A modern game, in theory at least, could dynamically simulate the subtlety of body language, the nuance of pupil dilation and blush response, and the subtle variation of tone and metre of voice exhibited by a couple having a quiet romantic dinner. A player in this theoretical game might deduce from this rich information that now might be a good time to finally propose marriage. That deduction would be meaningless, however, if all the player could do was press X to activate the ‘get married’ mission, or shoot their partner in the face. The coarseness of the possible inputs delimits the fidelity of the meaning and makes all the subtlety and nuance in the output irrelevant. So what can we do to enable more sophisticated negotiation of meaning in our singleplayer games? There are two approaches, but both are extraordinarily difficult.
The first approach is to widen the thin pipe. This means increasing the bandwidth of our possible inputs. To a certain extent, this is the promise of new hardware, such as motion, voice, or potentially facial recognition controllers. These new input devices theoretically give players more ways to communicate with the game. Unfortunately, for the most part, the current technology in these devices seems capable of detecting only the crudest of player expressions. They give us the ability to generate more low-fidelity input, not the ability to generate higher fidelity input. Enabling a player to communicate with their digital dinner date by beating their chest, screaming and making faces into a camera does not add subtlety and nuance. As a consequence, games where you play an angry gorilla work well using these new input devices. Games about dinner dates? Not so much.
The other approach lies in designing different sorts of games. A standard dual-analogue controller is capable of enabling half a dozen degrees of full analogue input (so analogue input into at least a six-dimensional space). That theoretically enables us to design a game where we control our body language, pupil dilation and blush response while constructing dialogue and modulating the tone and metre of our character’s voice. The difficulty here is that we don’t have a good sense of how to imagine such an interaction model. Even if we could control those things, what is the goal in such a game? What’s classed as a success? What’s a near miss? What’s a failure? How do we design the subtle responses from the AI so they are clear and readable when their ambiguity in reality is so important?
One thing is certain; until we start tackling these sorts of problems – and failing an awful lot – we will always be constrained by the FPTP predicament, and the dynamic meanings of our games will always seem to be of lower fidelity than meanings that can be achieved through more traditional authorship.
All that said, there remains the interesting idea that this inequality isn’t a plumbing problem that needs to be fixed. Perhaps the very thing that makes modern singleplayer games entertaining at all is the absurdity of having to perform every surgery with a chainsaw, of having to whisper every sweet nothing through a bullhorn, of having to vault every hurdle while wearing a giant foam sumo-suit. The idea that it is not only OK for games to be insane, absurd and playful, but that the irreconcilable mismatch between boorish action and suave reaction is the very point of the singleplayer game is worth considering – if not openly embracing.