You are on page 1of 15

The Anatomy of Games

First Draft. Feel free to help squish the bugs of nonsense in this first iteration.
Also, save me from disappearing up my own ass!
The Anatomy of Games
Introduction
This article is the result of aiming to think of a game and players in a formalized structure. It may
well be the first step toward a Game Structure Modeling Language. This could be useful in
grounding game design discussions in some common ground. Rather than talk in personalized
terms, a common nomenclature could be used assuming, that is, that this article is widely
accepted enough. It may not be! That is up to you to decide.
The article is open to change. If you feel something needs more clarification, or feel that
something simply isnt true, you can feel free to e-mail me, or post your comments on this page.
I will try to defend the article against sophistries, fallacies and straw man arguments. This is to
debate the article into a better form, rather than to overzealously defend my precious. If you
suggest a worthwhile change, I may not get around to it immediately, but I will get around to it.
Part 1: Basic Overview of Games Anatomy
This diagram shows the structure of a very simple game for one player. A good example would
be a one hundred meter sprint.

Fig1. A one player game.


Fig1. Describes the interactive loop between player and game system. This is seen quite often in
the Human Computer Interaction field under the guise The Model Human Processor.
Essentially the human being incorporated in the system is treated as a system in much the same
way as we would typically consider the game a group of algorithmic behaviors.
Thus, both of the elements in this interaction have fairly common anatomies in of themselves. In
the game model, we have a game state, which is affected by a set of rules, and outputs feedback.
In the player model, similarly, there is a perceived game state, which is changed by the players
interpretations. Output is produced from the player to affect the game as a result. These 3
elements to each model are fairly easily understood as twins. We can think of a player as a
special case of a game, or a game as a special case of a player, thus a conversation between two
people can be modeled so:
Fig 2. Player interacting directly with another player
Looking at Fig 2. it could perhaps be inferred that conversation is a game. It is not. The scope of
this article does not cover what constitutes a game. Instead, I will allow Jesper Juuls paper The
Game, the Player, the World [1] to cover that for us.
Game Definition vs. Game Instance
One distinction I will cover, however, is the tiny, but significant disparity between a game
definition, and a game instance.
One may stipulate, by this point, that the minimum required to define a game is this:

Fig 3. A game definition.


This diagram certainly incorporates elements of a game that we know to be common in all forms
of game (again, see [1]). The algorithms that deal with input, out put, data storage (whether
written on paper, implicit in playing tokens, or in abstract digital data types) maybe accurately
and explicitly defined. Yet it sits there, all dressed up, with no-where to go (or, more accurately,
no-one to go with)
Without any intelligent agent to imbue the games empty algorithms with play, there is no game
instance to witness. And without some physical or metaphysical representation of a game state,
there is no way for a player to ground his (her, or its) play within a structure, raising the activity
from playing to gaming. If a game definition defines the potential possibility space[2] of a game,
then an instance of a game (the playing out of game rules from initial state to victory, loss, or
simply end state) is one instance of play through those rules. Depending on the size of the
possibility space (depth, if you will), this instance of play may be unique. This uniqueness can
bring us to value a distinction between definitions of games and instances of games: A Game
(Chess), and The Game (Kasparov vs. Deep Blue, 1999) [3].
This brings us to an interesting point: Without structure, games would merely be play. Play must
be enjoyed by an intelligent entity. Without intelligent interpretation of structure and play, would
play have no value except to represent a permutation of game play? Is a game truly a game
without a player (human or otherwise intelligent)?
While a games algorithmic definition can exist in many forms (written, digital), does it still
require a truly intelligent entity to legitimize an instance of the game? Is Deep Blue vs. Deep
Blue merely a screen saver, interpretable, but not interactive; a resultant narrative from a
postulation of rules? A game is only truly an instance of that game to those who game, becomes
an unfolding narrative to those that spectate, but exists in isolation from intelligent interpretation
in the form of a definition of its algorithmic properties.
So there are three ways to interpret the game, now: a games definition explains the entire scope
of possible permutations of game play via algorithms; an instance of a game to its player is a
ludus veritas (true game) since the player is component and exponent, but an instance of a game
to a spectator is a ludus specto; narrative an unfolding, unaffected plot, merely grounded within
the rules of the miniature universe of the game.
Thus, while Fig 3. could be said to present a definition of a game, it can never be promoted to a
ludus veritas without a player. And if that player has no consciousness (a deterministic agent),
then at best, the game becomes ludus specto.
And with that unnecessary and anal distinction covered, we move on to inspect the finer points
of each element of the game; their domains, truths, and common fallacies.
In depth look at each feature
Game System
The metaphysical representation of a game that an intelligent player or spectator is able to
interpret has three common particulars: rules, the game state, and feedback.
Rules
The rules of a game iterate the game state dependant on the input received. Rules can be
considered a filter of player possibility. While a player may intend to kick a ball 5000 meters, the
rules of gravity, friction, and muscular ability are not going to allow this! Many games accept the
laws of physics as implicit on the games proceedings, since they are grounded in reality (Sports,
typically). Abstract games (on board, on cards, on paper, on computer, and always in the mind),
obviously only require the player to understand the rules of a game for play to ensue, while never
necessarily grounding themselves in the limitations of real physics (only the limits of human
thought). Aki Jrvinen, in his chapter Making and Braking Games: A Typology of Rules has a
far clearer dissection of rules than possible here.
Fig 4. Rules may take input into account when altering the game state. Or not.
In short, though, rules filter a players input into the game, allowing him or her to affect the true
game state in explicitly defined ways. Persistent rules may have to be kept up for every iteration
of play, possibly with respect to real time (objects accelerate under gravity, regardless of player
input on that point, can we model gravity as a special case of a player whose input is
constant, triggering the gravity rule, which in turn affects all objects? Just a thought that Ill
pick up on later).
Game State
The game state represents all the important temporary and ongoing information in a game. While
a games definition explains the data pertinent to the game explicitly, it should be noted that
factors outside of this domain (the mood and condition of players, the current state of turf,
weather conditions, theme, subject matter [4], etc.) are often considered to be implicitly defined.
Even a game isolated within a computer is not above the unlisted game considerations such as
ergonomics, or the ability of players to communicate with each other. Essentially, any pure data
that can be used for a player to strategize the next decision is part of the game state. A games
state, therefore, is not necessarily a game space[5], and may represent a domain of language,
knowledge, or anything else that is possibly recordable (positions, velocities and orientations
tend to be the considerations where space is an important descriptive factor in the game, but
should not rule out the possibility of other representational information packs of cards and their
orderings, words in a dictionary or encyclopedia, and other nuggets of data have no real concern
for spatial positioning).
It is hard to realistically limit the line we can draw around the useful information surrounding a
game. Even the time until ones supper could be a consideration in an online death match where
none of the players can dictate a time out! Or how about the boxer who has been asked to throw
a fight for pain and profit? While these factors do not change the rules of the game, they change
the way a player must consider his options, while not being explicit in the typical definition of
the game.
As a means of clarification, this article will resolve to ignore those data not explicitly required
for a game to ensue: that is, data that the rules can modify; data within the domain of the game.
Feedback
No matter where a game instances metaphysical state exists (in the computers circuitry, on
paper, on a board, on the track and field) its contents (all, or some) must be communicated to its
player in some form for a useful continuous feedback loop to be created. This can be through
visual, audible, haptic, smelly or tasty means: any or all!
The clarity of signal (noise to signal ratio) is the most important defining factor in how well a
game can present its game state to the player. Whilst many may quip that graphics are not
important, game play is all! we must give these people the benefit of the doubt and assume they
mean that graphical (or any type of) fidelity is unimportant as soon as it hinders informational
clarity. The quality of the imagery only serves to entice and sell the game, but typically neither
helps nor hinders the proliferation of useful information from the game state to the player(s).
This is not to say that either high or low fidelity feedback is better or worse in terms of clarity. In
WarCraft 3 and StarCraft, the high level of fidelity, en masse, can actually result in visually
noisy, ambiguous scenes, while low fidelity in sound can give equally ambiguous information.
And the ability to represent a mechanic by reference requires that the representative graphics be
recognizable at least some amount of fidelity is required to do this. Still, metaphor only helps
to explain rules to the new-comer. After the explanation has been digested, the gamer tends to
play the game qua game.
Perfect informational feedback is by no means a necessary tenet. It would make a game like
Poker irrelevant (how can one bluff if ones cards are exposed?). Instead, understand that it must
exist for a games quality of interactivity to be present a player must be able to perceive the
consequences of their actions. In Poker, feedback is taken from the reaction of language and
body language of other players from ones own actions. Feedback is intact, and thus, the game
can be played with a degree of skill, even without being able to act directly on the state of
everyone elses hand. Without feedback, how can one tell if one has succeeded, or indeed, failed?
Player as Model Human Processor
While a human being is easily analogous to the game system, he has the added ability to
strategize. This is the main thing separating the two systems. While one merely updates the
proceedings, the other also poses possibilities to its own internalized simulation of the perceived
true game state.
Interpretation
While rules in the game definition affect the true game state, each player reacts to the feedback
of the game by updating their own perception of the game state. Now, the feedback from the
game may or may not match up with their prediction of their previous action. If so, they must re-
assess their interpretation of the games rules and/or state. If not, the player is happy that his
perception of the games state and rules are somewhat consistent with the game itself. John
MacEnroes expletives demonstrate this perfectly as he perceives a ball to land out in tennis,
while a judge calls it in. The feedback from the true game state (which, in this case, is the
perception of the state held by the line judges, and fed through the Umpires good judgment)
does not match MacEnroes interpretation of the rules, and he is naturally irritated by this. In this
case, John expertise dictates that he cannot have been misinterpreting the rules all this time, and
thus, he must make a challenge to the umpire. His perception of the umpire as an entirely fallible
creature (much like himself) is the only justification for questioning the true game state that the
umpire holds in his mind.
Perception is all important here. If our senses deceive us, we will believe that something that did
not happen, did. If we then act on that interpretation, our interpretation of the game state will be
polluted.
Interpreted Game State & Strategizing
Once an interpretation of the rules has been built up in the mind of the player, he/she is able to
preemptively update his/her interpretation of the true game state. Additionally, this ability to
apply the rules and actions in the game to ones own internalized game state gives players the
ability to postulate an experimental strategy, and predict an outcome on the true game state.
[Added 21 July 2003] It should also be noted that such strategizing may be a result of knowledge
accrued outside the domain of the game: if a mechanical pattern in the game is easily identified
with a process that the player has learnt (either from experience or wrote knowledge) in some
other field, they may believe it prudent to apply this knowledge to the game. Whether the game
accepts their presumptuous implications is another question entirely, and would depend on how
closely the game's rules are married with the implied mechanic. Since some games (arguably)
live in total abstraction from real world models (and they must be at least a partial abstraction to
restrain them from being another instance of reality, and thus not a game), it is possible that they
may reject such strategies with extreme prejudice. The psychological effects of such a smack
down on a players presumptions may vary. Harvey Smith of Deus Ex and Ion Storm fame has
mentioned the ideas of good surprise and bad surprise in systemic level design an
unexpected outcome due to a consistent rule is accepted pleasurably from a player (I didnt
foresee that, but in hind sight, it makes sense), while consequences as a result of instantial
rules (that is, applicable in very special case scenarios) are met with indignant irritation (That
goes against everything Ive come to learn about this game or This game goes against the real
world metaphor it references in the case of a poor simulation). Thus, as a general rule of thumb,
one should be very careful how far one stretches their metaphor incase it uproots ones affinity
with the games representation abstract mechanics abhor abstract representation.[End of
Added. 21 July 2003. Thanks William]

Fig 5. Internalized predictions on the game state from experimental strategies.


As figure three shows, the player can iterate test strategies over and over, using their own
perception of the rules of the game to help generate an experimental game state. The player may
generate several of these and pick which is the best approach based on his short term and long
term goals.
If ones perception of the rules is not complete, or some of the game rules are non deterministic,
and if the rules are not well applied to the game state, then the player may not have a clear
impression of what the outcome will be. At this point a player is forced to either experiment (to
learn the outcome of a strategy that he/she does not know the result of) or build contingency
systems into his strategy (i.e. if I attack all out, but I get a 1 from the die roll, I will lose the
entire match, so I should be sure to be more cautious, and find a way to increase my chances).
Out Put
In the same way that the game must communicate with the player using feedback, the player
must communicate with the game using his/her possible range of actions on the game. There are
quite a few names for this set of actions, but I prefer the term vocabulary. If we think of every
singular action possible by the player in the game as a word, then the entire set of these actions
must be a vocabulary of words. Depending on the game, different amounts of actions are allowed
at any time. Chess only allows one of these explicitly defined actions at a time one word at a
time. It is arguable as to whether the implicit rules, such as bodily manipulation to move the
pieces around, or the reading of your opponents body language are part of the vocabulary in the
game. As explained before, they may not directly affect the game, but they can affect the
psychology of the players, and in turn, the true game state. It would be simple enough to block
off these forms of exo-ludic expression between players. Likewise, we could isolate our sports
competitors from audiences or distractions, who also alter a players psychology, and
occasionally interpretation of the game:
I can't do anything if the crowd fucking calls it Absolutely fucking ridiculous. At least replay
the point. Fucking ridiculous. Fucking ridiculous. It's fucking ridiculous. Frigging ridiculous.
Some wanker in the crowd changes the whole match. Well done. Well done. Absolutely shit.
Greg Rusedski, finding out that swearing is more fun than naughty audience-affected Tennis.
While ones method of expression in the real world is essentially infinite, only the explicit
vocabulary defined by a games rules can affect the true game state. Our ball calling hoodlums,
and psychologically imposing competitors can only hijack the feedback from the game state and
affect other players interpretations of it. Every action (as a result of deliberation of a useful
strategy, co-ordination, strength, dexterity, psychological state) is filtered through the possible
vocabulary into the game. The game state can only be affected by the vocabulary it defines.
The resultant acts in the game due to a word in the vocabulary are not always static (though they
may be consistent). Many words link into rules that act modally. There is the same sense of
polymorphism present as in actual language, where the semantics of words can be dependent on
context. Rules will interpret the vocabulary in different ways, dependant on the current game
state. Tomb Raider may be a good example of this. Moving between standing and running will
change the type of jump, and while airborne, pressing the jump button will do nothing at all!
Thus, a players ability to express themselves through the possible vocabulary is subject to the
games rules, which is why it is helpful to build a good understanding of those rules.
Conclusion of Part 1
When the rules are fully understood, the player is able to play the game in terms of the game, and
will have a closer connection with their abilities within the game. This is because the player has
the ability to accurately predict the outcome of an action (as explained earlier).
Early on in the learning process of a game, a weak metaphor to present a mechanic can harm
ones ability to play the game qua game:
In Return to Castle WolfenStein: Enemy Territory, the pseudo realistic world war 2 theme implies
that killing your own team mates is wrong, and bad for your team in general. But one rule trumps
this: promoted Field Medics are able to carry adrenaline syringes, which will restore a dead team
mate to full health. Savvy players, who played the game in abstraction, rather than playing along
with the realistic theme realized that it would be faster to shoot their own team members dead,
and then resurrect them to more health than they had in the first place, which happened to be a
lot faster than merely passing them health-kits. While the desirability of such a mechanic is not
in contention here, the rule does remain a non sequitor, which could understandably be jarring to
those trying to understand the game via metaphor, even if it is fun.
Metaphor is, therefore, merely a teaching and embellishing tool, as far as the structure of the
game is concerned certainly useful in explaining your metaphor, so long as the game
mechanics are able to stick to that metaphor. If not, the reasons for the mechanics can only be
explained through abstraction, ruining any immersion-through-metaphor desired. It is sad that
perfectly good rules become a slave to their subject matter, and are culled from a game, rather
than the subject matter expanding to explain the mechanic better.
Part 2: More Complex Models
Now that we have covered the basic structure of a one player game, we can begin to take these
base elements and mix them around to create structures representing other games.
Multiple Players:
Games of Direct Competition
The most obvious structure to derive from the single player game is the multiplayer game.
Fig 6. Multiple players affecting a common game state.
This diagram is a true description of a game where the players involved in the game are all
competing (or co-operating) in a common game state that is, the actions administrable by the
players on the game state may, through sufficient feedback, change all other players
interpretation of the game state. Weve already established that a player interacting with a player
is not actually an instance of a game, not least because both players have an imperfect
understanding of the rules, and a dirty model of the game state. Thus, their actions must be
filtered through a true game state, which provides structure to their play.
Depending on how important general communication is in the game (and its arguable that it
always is), we must also stream conversation and general body language through the games
body. This would be incredibly important in a formal debate, as the players ability to turn the
audiences opinion is entirely due to the vocabulary (pun) he/she can muster. Thus, this dinner
party conversation is not a game, since no node is one providing structure to the word play:
Fig 7. Not a game: The Dinner Party. Every arrow represents one persons feedback, and the
other persons interaction.
At the end of the meal, no-one is declared the winner (not officially, at least!)
Games of Parallel Competition
For sake of example, let us labor under the fallacy that there are very many examples of game
that remain competitive without having a common game state between players. We could say
that the structure of two players sprinting 100 meters as fast as possible, or two players aiming
for a high score in Pac Man is like this:

Fig 8. The parallel, competitive game.


But if a victor is to be decided scores and times must be compared, and in comparing, were
implicitly creating a game. Its an incredibly simple game, but a game none the less. Our
instances of play by each player (a ludus veritas each) become equivalent to a player providing
output to the game. As inputs, the comparison game takes the scores from each game of Pac
Man, or the times from each sprinter. It compares the two. If one score is greater than another, or
one time is smaller than the other, then that ludus veritas has won! Thus we fix the parallel
game into this form.

Fig 9. The directly competitive game, at one level of separation.


One can freely describe this game as competitive in parallel merely because it is a common
pattern in games.
Another Level (The Psychological Game)
The consideration of communication in the game is truly a point of contention where parallel
games are concerned. In directly competitive games, the answer is simple enough all
communication from other players is merely another kind of feedback from the game: a player is
able to lie about the game state (bluffing in Poker), or help explain the game state to other
players (warning team mates of enemies around the corner in any team based first person
shooting game you care to mention). It is useful information about the shared game state, and
thus it comes from other players, through the game itself, and not directly from player to player.
But in the parallel game model, communication of each players own game state can result in a
change in tactics in the other player, even if the game states are not common! Certainly we can
say that the stress of knowing that another player is winning, or the joy of knowing others are
losing, comes directly from the comparison game. But what happens when the information is not
about ones own game state at all? Can it affect a players choices?
Example: A player, Timmy, calls over to another player, Jimmy, while they both play through a
shoot em up: Hey, this red laser is really powerful. Jimmy responds to this as if it were simply
an inner voice, suggesting a new experimental tactic. So Jimmy, indeed, switches to the Red
Laser, and finds out that Timmy was lying, and merely attempting to slow Jimmys progress.
This deception could certainly be thought of as an entirely different game, going on at a level
above the comparison game.
But it is not. In the same way as communications are filtered through a game state in a directly
competitive game, (because communications are implicitly allowable, and not explicitly denied),
communication can invisibly, but implicitly, pass through the comparison game, then back
through the individual game states, and present themselves as feedback to the players.
In fact, if we were to represent a game as a single player interprets it, then all the other players
would be perceived as part of the game. Even their shouts of you suck! would merely be a
form of feedback from the game.

Fig 10. A multiplayer game as perceived by a single player (ludus veritas)


This brings up the interesting notion: how can a single player know that the other players
arent actually acting deterministically? Very sophistic!
Aggregated Players (Teams)
For many reasons (fair play[5], not the least) we may want to represent a number of players as a
group. Perhaps they all have the same objective and motivation within the game?
While each player in this game may be imbued with a different vocabulary as far as the game
itself is concerned, it may be useful to treat their interaction with the game as an aggregation
(and not in part to combat diagrammatic ambiguity and over complexity).
Fig 11. Team simplification
Since, in many games, the teams overall score is often more important than each individuals
contribution, it may be easier for the game to regard a team as an aggregated player a player
made up of lots of other players. This team takes in and interprets feedback, as a normal player
would, but interpretation is split between individual players some may have more, less, or
different information to work with than other players in their team. Then, all the players
resultant actions are passed to the game as one sentence (lots of words from the vocabulary). The
vocabulary in this case is representative of the team as a whole, and not of the individual players.
In the same way that individual players have different interpretations of the rules and the game
state, they may also have different vocabularies (A Goal Keeper in Soccer may use his hands to
pick up the ball while in his own penalty area, but other players may not). So a team is like a
multi cell organism, each cell acting differently to others, but with an aggregated outcome acted
upon the outside world as a result.
Why is this useful to know, at all? Partly, it clarifies our thinking when dealing with a large
number of players. Also, it allows us to compare the vocabularies of teams, which is important in
considering fair play[5] between teams.
Aggregated Games (Play Grounds)
Weve noticed that players can often be considered a special case of games, and vice versa. So
what is a games equivalent of a team? Well, if a team has many instances of players within it,
then an aggregated game must have several game states within it.
How does this manifest itself in terms of existing games? A perennial favorite, Grand Theft Auto
3 is a great example. At any given time during play, there may be quite a few games going on,
passively. One may optionally search for, and pick up secret packages for armaments and cash,
or one may want to build up cash by performing stunts, destroying vehicles, and killing passers
by for their wallets. These activities are all regarded separately, but increment a common score
(your cash), and most importantly, are played using a common vocabulary. Your ability to shoot,
run around, hi jack vehicles and drive around are all factors that can affect all of these games that
go on in parallel. Quite a nice result of these parallel rules is that one can occasionally un-
intentionally complete a goal for one game while intending to complete another. How many
times have you found a secret package as a result of an impressive jump, or midway through a
more formal, explicitly chosen mission?
Fig 12. Several smaller games, aggregated, can be affected by a common vocabulary, all at
once.
Mario 64, and Tony Hawks Pro Skater are also good examples of this approach. A common
vocabulary passed to the aggregated game can result in changes to the states of several games
(well, goals) that are going on.
Thus, a good analogy is a childrens playground, where boys are playing soccer on the same
gravel as girls playing hopscotch. The two games may interrupt one-another. A ball is kicked at a
girl, knocking her off balance. Or a girl kicks a stray ball back into play, and scores a goal from
her action. The common vocabulary of the players is the manipulation of their own bodies, to
affect their own position, and the position of the ball.
Agents as Players
It has already been suggested that if an instance of a game were played by an unintelligent agent
(say, a computer that would react to a game state as programmed, perhaps with some randomness
to their action, if desired, though randomness would not constitute free will), then the game
could never be a ludus veritas for any player wanting to exist as that instance of that player. It
could still be a ludus veritas for a player wanting to occupy another player slot in the definition.
For anyone else, the game is ludus specto.
Since these agents can fill the boots of players completely, were presented with a little bit of a
Turing test. Instead of our vocabulary of actions being text in a chat client, and the entire scope
of the world as a game state to refer to (as with a typical Turing test), the agent takes on the
vocabulary described by the game, and uses the game state of the game as reference.
In this way, we can start think of monsters and dumb enemies or entities in games as player
agents. Yes, they may have very different vocabularies to the typical player, and they may act
totally deterministically, and their actions may only be decided based on a tiny proportion of the
game state, but they CAN still be thought of as players. Namco are bringing back a five player
version of Pac Man, in which one player plays Pac Man on a GameBoy Advance screen, which
shows the entire game state, while four other players control the ghosts, but each ghost only has
a limited range of view on the game state. This is the perfect proof of the concept of agents as
players.
Agents as Rules, Players as Rules
The Ghosts in Pac Man always used to move in a deterministic way, based on the position of Pac
Man, or a basic roaming pattern. Their movements were rule based if I am close to Pac Man,
chase Pac Man, or else keep roaming. At this point, you could consider their movement as a rule
of a game. If you wanted, you could play the game on a board, and iterate the rules yourself.
Namco have similarly moved the enforcement of this rule to the players, by limiting the required
representation of the game state for each Ghost player to answer the question can I see
PacMan?.
Although human players have the ability to go against the rules (they dont have to chase Pac
Man if they see him, and in Tennis, they dont have to call a ball out when it lands outside the
line), this does not mean that a game definitions rules are undermined. They may, though, be
under enforced in the instance of the game, since the human being is not an accurate or
deterministic creature.
In exactly the same way that rules affect the game state, based on a players derived
interpretation of the game state and their subsequent action, a rule-as-agent can deliberate what
action to take based on the current game state, and then send a flag, or value (essentially, a
different vocabulary) to the handling rule. Thus:

Fig 13. A non player vocabulary related rule is enforced by an agent


So we finish with an interesting concept. A rule in a game (any rule) can be enforced by a player
if they are given ample information from the game state to be able to enforce the rule, and the
ability to alter the game state as required. An intelligent Basket Ball hoop could increment a
teams score every time a ball fell through it, if it could only know the distance the shot was
taken from (2 shots within the semi circle, 3 outside of it), the state of play (regular play/free
shot) or if the shot was from a foul (null point) etc.
This could be very useful to game developers wanting a flexible way to handle game rules.
Rather than having an invisible rule checking algorithm, each rule is embodied as a player that
responds deterministically to certain relevant elements of the game state.
Conclusion
A Game needs a player
A game can exist in abstraction as a definition of itself.
A game definition becomes an instance of a game if a player or agent imbues the definition with
play. The games structure turns their mere playing into gaming.
If someone who is not you plays the game, then the game is ludus specto an instance of a game
that you have no interaction with.
If you play an instance of a game, then it is a ludus veritas (truly a game), since you are both a
component and exponent of play.
By definition, an unintelligent agent (that is, a computer programmed to play a game, or a natural
consistent force like gravity) taking the place of players in a game can never create a ludus
veritas for itself, because it has no concept of self.
Interpret, Process, Communicate
Each element possible in a game instance (Game, Player, Agent, Rule Agent) shares a common
structure in that it receives an input which it must interpret, process and update its own
internalized interpretation of the game state, and then communicate out using its vocabulary set.
An audience isolated from play, however, would have the first two elements, but not the last, and
thus no means of interaction.
Player = Game = Agent = Rule Agent
Each possible element in a game, because of their incredibly similar structures, seem like special
cases of each other.
A player receives input from the feedback of the game, updates his or her interpretation of the
game, and in addition, is able to iterate experimental instances of this game state in his head, or if
this is not possible (due to a poor understanding of the rules) he or she is able to simply guess.
Their communication comes in the form of the vocabulary they can express on the game. This
expression is intended as an implementation of their decided strategy, but due to problems in
interface and imperfect hand eye co-ordination and limited strength or verbal skills, this resultant
action may not match their intended action. The result of their action in the game state may also
not match with their intended strategy simply due to a poor understanding of the rules.
A game receives words from the player/agent (the distinction is unimportant as far as it is
concerned) and uses its rules (which may query the game state) to interpret the inputs, and
updates the game state. It may feedback any or all information about the game state, but may
feed back different sets of information to different subscribers. If a game does not have any form
of feedback, then the interaction is broken, and there is no game.
An agent is most similar to a player, except that it is fairly deterministic. Its deliberations of
strategy are not affected by free will and experimentation in the way that a players is. Its
understanding of the rules is perfect. Its interpretation of the game state can be perfect.
A rule agent only exists to enforce a particular rule that is not implicitly linked to the actions of
the regular player. Thus, it only exists in an instance of a game and not in a games definition. It
acts only to instigate a rule if that rule needs to come into play based on the current game state.
Foot Notes
1. Jesper Juul, DIGRA 2003.
2. Possibility Space first termed by ?
3. It is not uncommon to think of instances of games as representations of emergent narratives.
Do you think in terms of the range of possibilities that game play can rise to form (bad and good
alike), or of specific, idealized (but ultimately shallow) ways of playing a game? Perhaps this is
the thinking that separates narratologists and ludologists?
4. For more on theme, see Akis chapter again.
Posted by Aubrey in Articles on 2003.07.09 at 02:52 PM

You might also like