"Loopy" Wargame Design

I found David's 27 June piece on Principles, Models, and Algorithms interesting and figured the post deserved more than I could produce in a short comment to his post directly.  After some thought, I tackle David's essay.

To briefly summarize David's essay, he argues that rules can be understood at three levels (or layers): principles, models, and algorithms.  Principles are broad goals that require judgment and may be ill-defined or vague.  Models are the underlying ways in which we try to understand reality and quantify those perceptions into workable solutions.  Algorithms are the mechanisms used to resolve game actions without any bias or judgment.  David suggests that this process carries a definite hierarchy with Principles at the top followed by Models and finally Algorithms.

David's perspective on algorithms (rule mechansims) is that they are useful since discrete rules can be applied consistently by anyone.  That is, if the rules are clear and without ambiguity!  A potential downside pops up when these algorithms mask the principles and models underlying them.

David's hierarchical way of considering rules and scenario writing is a helpful approach.  The distinction between principle, model, and algorithms is useful.  My experience in evaluating ruleset design illustrates that many of the arguments about wargame rules come from confusing these levels or from expecting one layer to do the work of another.

To respond to the post, I first relabel David's levels to terms that carry more meaning to me.  "Objective" replaces "Principle" and "Mechanism" replaces "Algorithm."  "Model" remains the same.  I tend to think of these levels not as hierarchical, one-way progressions but as layers that interact with one another.  When a rules mechanism exposes a flaw in a model, the objective may need to be re-examined.  In practice, designers move back and forth between each of these levels in a loop as shown in the diagram below.

In scenario design, the idea that objectives sit at a higher level of abstraction than mechanisms feels exactly right.  The aim of a game is not always simply to defeat the enemy.  Once objectives shift to delay, escort, screen, or hold key terrain pieces, the whole character of the game changes.  Objectives should not be vague or ambiguous.  In a sense, objectives help to shape and classify the historical problem we are trying to represent on the table.  Victory conditions, laid out as an objective, tend to shape player behavior more than mechanisms do.

The idea of a model is also useful, but some clarification is warranted.  A model in wargaming is not only conceptual but takes the role of creating a selective abstraction of causation.  As a designer, we make choices.  In other words, a model tells us what we think matters in a given historical situation.  That seems especially important in wargaming, where two rulesets may look similar on the surface but actually rely on different assumptions about morale, firepower, cohesion, or command.  There are different types of models in wargame design as well.  Models can include physical models (ranges, formations, movement rates, weapons effects), behavioral models (morale, command friction), and perceptual models (fog of war, limited intelligence).  Different combination of models or different implementations explain why two rulesets with the same, core mechanisms can feel very different.  

On mechanisms removing judgment altogether, I find this not the case.  Mechanisms certainly channel responses, which is valuable, but they do not eliminate judgment from the game as a whole.  Rather than eliminating judgment, mechanisms relocate it.  Players still exercise judgment in movement, target selection, risk-taking, timing, and so forth. Perhaps a better point is that mechanisms confine judgment by removing arbitrary rulings.  Mechanisms allow for a more explicit and repeatable structure.

In the end, the most valuable part of David's essay is that it reminds us to ask not only whether a rule works, but why it works.  A game may have clean mechanisms but still fail to express the historical problem it is trying to represent. Conversely, a simple mechanism can work quite well if it is supported by a sound model and a clear objective. A looping design and development structure helps to focus the interaction between the three levels of objective, model, and mechanism.