The Perception Layer

Download Audio

The perception layer (in my terminology (which is what I use to communicate)) is an intermediary between reality and memory. It models the fact that, when you perceive something, your perception doesn’t always match what’s really there. In fact, it never matches it perfectly. Occasionally, perception and reality are far divergent.

I think of the perception layer as a piece of glass. If it is smooth and clean, you can look through it easily. If the glass is clouded, painted, or covered in dirt, your perception is going to be obstructed. Even worse, if the glass is uneven, warped, or otherwise malformed it can give a clear but distorted view of the other side.

Fog of War is a common example of a perception layer in RTS games. In the “real world” of the game, there are your units and the enemy’s units, all existing clearly and definitely. But in your perception, much of the map is hidden. The perception layer clouds your ability to view the enemy’s units until you’re close enough, or have radar, or clairvoyance, or other methods to gain a broader perception.

The perception layer has so many benefits! In the abstract, it:

  • Allows slop for narrative inconsistencies (unreliable narrator)
  • Eliminates perfect extrapolation
  • Introduces realistic misunderstandings
  • Limits the player’s grounds for nitpicking
  • Provides limits to required model fidelity
  • Allows for perceptual conflict
  • Allows for “imaginative” acceptance of images and models as reality.

Briefly, if what is presented to the player (and the other controlling entities) always represents the exact underlying game state, then there can be no subjective truth in-game. Adding perception allows the narrative to bend more easily, since any inconsistencies could be chocked up to perception instead of writer error. Perfect extrapolation is a problem in many games because it allows the reduction of interesting choices to simple calculations. But without perfect information, there can be no perfect extrapolation. Misunderstandings, too, arise from differences in perception, which can not occur if every entity percieves an identical state. The player is freed from the burden of nitpicking as well, since what they see is not precise their criticism must necessarily be more broad. The model itself of the underlying game, and of all the actors, can be reduced in fidelity. If perception only works to a certain depth, there is no need to model the finer details, which will rarely be percieved anyhow. Characters will be able to support truly different views, instead of simply different philosophies, which will lead to conflicts of perception and arguments over truth (a common pastime in real life). Once the player is freed from perfect perception, they can begin to imagine details, inflections, and implications which, while not present in the game, will nevertheless impart a sense of reality. No doubt I have overlooked some aspects, but these examples should serve to point in the direction I am headed with all this.

Unfortunately, the perception layer is often ignored as a game mechanic and a source for interesting game interactions. This is a real shame because there is so much potential for so little cost.

The cost does exist of course. A perception layer imposes a large burden on the referee to disseminate and differentiate different perceptions. Thus it is best suited to digital games where this burden can be born by the computer, which generally doesn’t mind doing menial tasks for our enjoyment.

One of the main technical hurdles is that a perception layer requires essentially an entirely separate set of “game state” for each player character (or, more generally, for each character in the game). While this may have been an excuse in the past, growing hardware capabilities and multi-player architectures have made this the ad-hoc state of game engines, rather than some exceptional technical challenge. Today’s digital game systems are more than capable enough to handle this duplication of detail.

Another challenge is that of dealing with how imperfect perceptions effects the game-play. Most of the time, perception (and the faults therein) is never modeled because it is something frustrating to deal with in real life. We hate it when we mis-hear, or remember incorrectly. Often these are terrifying symptoms of internal systemic failure, or the slippage of sanity. Yet these minor errors of observation occur all the time, and are a constant presence in our every-day lives. Really, the question is “How can we possibly implement believable character systems without perception?” instead of the inverse. We are naturally attuned to account for perceptions. It seems that including this ability in our games is a reasonable expression of a human truth, and should allow other systems to present much more simple models without sacrificing believably.

The largest problem I see with the perception layer as a game design feature is that it draws attention to the artifice itself. It is so similar in nature to the purely perceptive interface of games as a medium that it has the potential to completely shatter suspension of disbelief. Now, I don’t have a problem with this, as suspension of disbelief seems like something that should be won instead of assumed. However, if your game is unconvincing, adding a perception layer may only draw attention to the fact.

It seems the uses of the perception layer fall into two broad (and overlapping) categories: Imagination, and Lies.

  • Imagination, creativity, and communication. Characters in-game could perceive their plans as having already occurred, or having the same “realness” as “actual” events. Then actions would be merely guiding the “actual” events to match some percieved “ideal” course. In this way, characters can be satisfied when something goes “according to plan” because they posses a subjective reality in which the plan is already accomplished. Likewise, characters could be dissapointed when things don’t turn out how they thought. These subjective realities could be traded by communication, with the danger that always comes from mis-communication.
  • Lies, deception, and confusion. The perception layer begs for these aspects, and in fact has often been implemented (in a limited way) to aid games of stealth and strategy. The ability to present an apparent reality (especially one that a foe is fixated on, either with desire or with fear) which differs from actual reality is only possible when a robust perception layer is in place. Likewise, warring perceptions can cause confusion, as is so often the case in battles both small and large.

There are so many specific applications where I think the perception layer can (and eventually will) be put to good use. Because perception is itself a characteristic (that is, a property of characters) It will mostly be character focused. But of course all things are made up of characters so, the closer our games draw to reality, the broader the applications will be. Here’s one specific example:

Deceptive “Fog of War”

As mentioned above, the Fog of War is a common concept in strategy games. This ignorance of conditions, especially about enemy movements, mimics the real world battlefield uncertainty, but it is rarely taken very far. The simple (and common) implementation is to show the terrain, and then reveal enemy units as they come within a “sight range” of your units. When the foes leave your sight range, they disappear from your view. Even this could accept a great deal of improvement. Stealthy or camouflaged movement (at a cost, perhaps to movement speed) to allow units to close with the enemy without being spotted. “realistic” sight range, with obstacles. Unit memory (ghost records of the last sighted locations of enemy units) would also aid in keeping track of where and when the enemy was spotted.

But this is just the beginning. Once there are ghosts and memory we can begin extrapolating and guessing. Buildings too (which are really just large units, but that’s a different matter) could have ghosts (which some games already support, since buildings move more rarely than units), and even extrapolated constructions based on the plan you expect your foe to pursue. Once we take the step to extrapolations, ghosted units could display a possible location zone showing where you expect them to be able to move since you saw them last. You could employ these tools for your own forces as well, to make plans for the future based on your own units extrapolated possible positions. All of these aspects would have no necessary effect on the underlying mechanics.

And once we start tying the perception layer firmly into the mechanics things can get very interesting indeed.  The enemy could spend resources to plant false information, make units look like different units, and change the apparent quantity or strength. You could send both your allies and enemies misleading or partial reports on your deployment.

And, once we get down to the level of each unit having its own perceptions, then we have nested perception layers. For the player is but one unit, and can not actually see anything other than what his forces report.  You could disguise your units as the enemy (though they will probably get shot by your own units if you’re not careful), or trick your foe into shelling his own forces.  You could spend resources to improve your units ability to penetrate disguises. Units and buildings of all sorts could have differing perceptions, with conflicting reports on the enemy’s strength and movements. Morale could affect perception, with high-spirited forces underestimating the foe, and cowed units blowing their enemies out of proportion.

All this (and more!) awaits the bold designers willing to implement the perception layer, and care for it until it yields fruit.

Until next time, my friends, see clearly!

Leave a Reply

Your email address will not be published. Required fields are marked *