Gaming the Past

This is a reprint of my post at PlayThePast from March 2012. This essay generated some insightful comments, for which see http://www.playthepast.org/?p=2594

What simulation games do best as interpretations is present the past in terms of problem spaces. This is a concept I have co-opted from games and learning theorists (most notably Henry Jenkins and Kurt Squire) for use in thinking about how we teach and learn about the past and use simulation games. I have an article in the works that addresses  some aspects of this, but I want to test these ideas out in the meantime. I am also inspired by the recent posts from Trevor and Rebecca and the enthusiastic feedback they generated and wanted to make some sense of that in relation to my work.  Any feedback would be most welcome.

I.

The concept of problem space is a highly useful tool for studying historical simulations, teaching history, and using the former to help in the latter.  Simulation games are interpretations of the past designed as problem spaces. A problem space, at least as I currently define it, has the following features:

• Players, or in the physical world, agents,  have roles and goals generally contextualized in space
• They make choices in an effort to achieve these goals
• The choices available and their success or failure are shaped by:
  • The affordances of the space (which can include quantifiable resources, cultural frameworks, psychological tendencies, etc.)
  • The constraints of the space (which can include finite quantifiable resources and scarcity, cultural frameworks, psychological tendencies, etc.)

The problem space design of a simulation game is, in some respects just a more sophisticated articulation of the basic core of game-ness. By most definitions games require players, conflict, and a quantifiable outcome. Players have affordances and constraints embodied in rules. What a historical simulation game does, however, is craft a virtual problem space that represents a real-world one.

The problem spaces in simulation games are subject to some particular constraints. One of the most important is the constraint of quantification. Simulation game programs, as computer games, must be reducible to 0s and 1s – this is the only language through which a CPU can receive orders. Consequently, all elements of a historical simulation game, including agents and their motives, must be reducible to numbers. My favorite example of this is the happiness metric found in many city-builders. This takes a very imprecise real-world concept and transforms it into a precise number that can be increased or decreased in precise ways (often through food, housing, jobs, taxes, and amenities) that have a precise effect on the city population (generally determining whether immigration or abandonment occur). Happiness, productivity, popular unrest, attack and defense strength, espionage effectiveness, cultural influence—no matter how qualitative the concept in the real-world, if it’s an actual mechanic in a game, it is strictly quantified.

Additionally problem spaces in simulation games, however open-ended they might appear, are closed. To function, these games must be working, closed systems, completely operational once the player joins the mix. As expansive as a game might be in its treatments, it will impose arbitrary limits on its subject. These limits begin with the roles and goals of the player, decisions that shape the entire design. Perhaps more problematically for critical scholars of video games, simulation games are, as games, teleological in their focus. The quantifiable gameplay elements and mechanics all, in a tightly designed game any way, factor directly into whether the player achieves their goals.

II.

There has been excellent discussion on PtP about the appropriateness of, and methods for critiquing simulations historically. Teasing out the ramifications that they are interpretations in the form of quantifiable problem spaces can provide some important insights on this issue. It suggests considerations for rigorous and meaningful criticism that is holistic and sensitive to the medium.

First of all, just like any other interpretation of the past, simulation games will select certain aspects of the past as their theme and not others. This is true of all historical interpretations—after all an interpretation that includes everything is not an interpretation at all. To be playable and appealing, a game needs to have a set of core mechanics that are tight and cohesive, modeling one overarching system well. Consider the standard genres of simulation games that have developed over time: city builders, nation management, trade, war, diplomacy, politics, etc. Though there is always room for crossovers and new genres, the existing genres of successful simulation games point to a constraint that a compelling game—just like a focused narrative or analysis—must focus on some things and neglect other things.

Because simulation games must function as working systems, however, the choice of problem space, or more specifically the choice of whose problem spaces to represent necessarily locks the game into certain portrayals of the past. Other media are not subject to the same constraints. One could easily conceive of a textual narrative/analytical work, for example,  that devotes time and space to a variety of viewpoints and agencies.  Still, no narrative or analysis covers all or even most viewpoints and all are subject to authorial predilections. Most importantly when considering the difference between simulations and these media, written texts are not quantitative rule sets executed by a computer to form a problem space. Even if writers wish to analyze the past in terms of problem spaces, they are free to select a variety of roles and goals that may have at times only tangential relationships. Further, they can select affordances and constraints that do not always form a complete system, and digress on important philosophical and practical comments in ways game designers simply cannot. The game designer must think in terms of a complete working simplified system in ways the writer does not. This is certainly not a bad thing when one considers the goal of designers is to entertain and interest players, After all focused games fare better than those with tacked on elements that do not contribute to the whole.

III.

All this may seem fairly obvious, but there is an important point of criticism here that is not always fully appreciated. When trying to understand why an element of a simulation exists in the way it does and what it suggests about attitudes towards the past—whether why Colonization codes native peoples the way it does, why Civilization does not deal with social issues in cities, or why East India Company does not treat the tensions between English and Indian customs—one needs to consider holistically the problem space selected by the designers. First off, one must consider the goals and roles of the human players set out by the design plan. Certainly in the real world there can be many agents in a problem space with different roles and goals that can complement, conflict, or have little bearing on one another. Simulation games, too, can represent multiple agents with varying roles and goals. In single-player games these additional roles are handled by the program’s artificial intelligence routines. In multi-player games human players can take on additional roles. Generally speaking however—and I welcome examples where this is not the case—simulation games, especially pleasurable and/or commercially successful ones  must commit to a very small set of roles and goals, often one role and one goal. Even where roles and goals differ and conflict they tend to be set up as binary opposites or at least draw from the same well of resources and affordances. So both sides may want to hold a territory or win an election, one group may want independence while the other wants centralization, the city ruler wants profits while the citizens want material niceties, etc. This is in large part, again, because games must be closed functioning systems: each part must connect to every other part. So a game cannot represent roles and goals well that do not fit into the core choices, affordances, and constraints of the chosen problem space. Therefore the commitment to a particular articulation of a problem space will shape every other aspect of the game and any analysis of an element of the game, not least of all an agent, must consider the framework of the problem space.

Let’s apply bits of this theory to a concrete example. Hegemony: Philip of Macedon by Longbow Games, a real time kingdom management / strategic and tactical war game hybrid. The player assumes the role of Philip, the fourth century king who unified Macedonia, subjugated Greece, and, in the process, built the formidable military system that his son, Alexander, would use to conquer the Persian empire. Consider the portrayal of slavery in the game. When one defeats an enemy unit, the survivors can be captured and enslaved. If the survivors are not captured within a short time, they will escape. Slaves can be used to work mines, perform general construction tasks, and transport food supplies. Left untended it is possible for slaves to escape. Now, this is a reasonable sketch of aspects of slavery in the ancient world, not least of all the matter-of-fact nature of a system we find repugnant today.

Suppose, however, one wanted to criticize formally this historical representation of slaves. One might start by noting that these slaves have very little agency. Granted, they have the goal of escaping in this problem space and can do so if left unattended for too long. This goal does little more than add a constraint to the player’s problem space, a reason to take care attending to slaves and spending resources on watchtowers. This is not much of a depiction of agency at all and so slaves become nothing more than affordances, resources for the player to exploit in the game. This sort of portrayal might inspire comment and critique that even enslaved people had agency beyond escape, the ability to make choices and have a degree of ownership over their lives despite the horrible constraints of their status. (I recollect from my limited work with U.S. history that this sort of development took place in the historiography of American slavery, from first not studying slaves to studying slaves as victims, to studying slaves as agents while still recognizing the oppression of enslavement).

Why does the game not portray the agency of slaves? How Longbow defined the primary problem space, the human player’s problem space, is a critical answer. For the player Philip king of Macedon is the role with a goal of uniting Macedonia and building a Balkan empire. With this role and goal driving the articulation of the problem space, depicting slaves in the game as affordances is fully understandable. One could attempt to flesh out the slaves feelings on their situation, but it is difficult to see how that would fit into the mechanics of this particular problem space, the one the designers chose.

It is important to note, however, that saying a portrayal of ancient slaves, native Americans, Hessian mercenaries, railroad barons or any other agent or aspect of the past, takes the form it does because of the problem space is not meant to be a tactic for ending discussion or defending an implementation (One could imagine such a chilling effect: “why are they portrayed this way? Because the problem space demanded it. Oh … okay, so what’s for lunch?”). It is meant to focus criticism on a game holistically and consider how the affordances and constraints of the simulation game medium and the interests and goals of a game’s creators (their concerns, assumptions, hopes, attitudes, what have you) shape a game’s interpretation of the past. At the risk of being too meta, but in all seriousness, one really needs to consider the problem space of the game designers when considering the elements of the simulation they designed. Once the commitment has been made to make a commercial simulation game, as opposed to using any other medium of interpreting the past, the affordances and constraints of conceptualizing history as problem spaces place great pressure on the final product.

So, what kinds of questions might one ask of a simulation game as a problem space and what kinds of meaningful criticisms/evaluations can be make? A few, necessarily incomplete suggestions:

• One might meaningfully question why the particular main roles and goals for the game were selected in the same way one can meaningfully question why certain generations of historians privileged one set of topics and questions over another. Indeed meaningful answers to such questions can be given based on careful research of prevailing ideas at the time. Simulation games, for example, tend to be inclined to issues of domination whether in political, military, or economic forms – discussing why this is continues to be a lively debate.
• One absolutely should question whether the roles and goals selected for the players are historically legitimate. In other words, do they reflect what our evidence suggests were some important roles and goals in the past? There is little question that Philip wanted to dominate the Balkans. In other cases, such as Colonization where the goal, as stated on the Firaxis page is for colonists to “negotiate, trade and fight as they acquire great power” http://www.2kgames.com/civ4/colonization/  one might very well explore the cases in the past where this articulation of goals was and was not valid. That’s a great conversation to have, and it has great bearing on the validity of each element in the game’s interpretation of the past.
• One can rightfully question why each and every element of the game is portrayed as it is. But these questions should not be divorced from the consideration of the problem space as a whole, especially the historical roles and goals conceptualized by the designers. A thorough consideration of why slaves are mere tools in Hegemony, happiness is the defining metric for success in CivCity: Rome, Indian culture is not represented in East India Company, or any other element in any game, should  consider the goals set out for the game and the supporting game mechanics to be thoroughly compelling.

So, suppose that one accepts the roles and goals of a game as historically valid goals, i.e. goals that reasonably represent what good evidence suggests motivated some peoples of the past. That might well mean that a thorough challenge to the portrayal of some historical agents in the game could only be made by suggesting that:

a)      That the agents could not reasonably be conceived to play that role  in the problem space from the point of view of the player, the primary agent

b)      By suggesting what more legitimate roles the agent could have played in the game that would mesh into a system, or at the very least not conflict with, with the player’s roles and goals in the problem space. Considerations of this sort need to be very aware to the developer’s presumed goal to create a playable, enjoyable, and commercially viable game.

c)      (A variant of b) By suggesting what roles and goals non-player agents in the game could have played that would have worked in a system centered on the player’s role and goals.

Note: The day before posting I was browsing through old PtP posts I had missed, and noted happily that Trevor made comment in this spirit on Mark’s essay on Detroit. Trevor essentially agreed Mark’s critique of the absence of race in SimCity/Micropolis is valid, but asked him to suggest how he would have race represented in the game.

So challenging the portrayal of slaves in Hegemony, if I accept the historical validity of the role and goals (which I do), would require suggesting how slaves could have been portrayed more complexly and validly within the defined problem space, how they could have had a greater portrayal of agency through expanded roles and goals. This certainly can be done. Consider another nation management / strategy game, Paradox Interactive’s  Europa Universalis III. One of the problems the player-ruler faces is nationalist revolts. When a nationalist revolt erupts in a territory, the rebel units besiege the territory and, if unchecked, ultimately seize control of it. When the nationalists seize the province they preventing the player from taxing it, drawing troops from it, or administering it.

(Note I was mistaken in the characterization of the next few sentences. Check out Liam’s Burke’s comment on the original PtP post. Fortunately, my point was to illustrate the structure of a critique, not levy a challenge against EU III)

What the nationalists do not do, however, is actually create an independent state, i.e. a functioning state along the lines of the hundreds of others in the game. This limits the nationalists’ agency severely, making them serve as nothing more than a constraint, rather than a fleshed out agent. Indeed once the player can commit the necessary troops to reconquering the province, it will be reintegrated into the kingdom. The designers presumably could have easily made successful nationalist revolts result in the creation of new small states; there are many states in the game, after all who only control one province. Doing so would have represented the agency of nationalists and the goals of nationalist revolts more realistically, rather than the current state of affairs where they simply lock down a player province. Such a change could mesh well with the game’s overall problem space—though there may still have been compelling design reasons not to have done so.

On a final note, it is not my intention to criticize existing game analyses, or suggest that I have fully practiced these ideals — I’m quite sure my critique of CivCity: Rome from last year did not offer specific suggestions for remodeling the game. Nor am I suggesting that thinking in terms of problem spaces is the only meaningful way to conceptualize simulation games. I’m simply reaffirming that simulation games are human interpretations of the past subject to certain constraints, as sources and media they should be considered holistically, and this can be done by thinking in terms of problems spaces. Next week, I’ll add a shorter post considering how the concept of problem spaces and the use of simulations to illustrate problem spaces can be very effective in the classroom.  In the meantime, I hope this will generate some comments; there is a lot more to be said on these matters.