Peter Perla, author of The Art of Wargaming, works for the Center for Naval Analyses, which has a number of his papers available online, along with others related to wargaming. A few of particular interest are:
An Introduction to Wargaming and Its Uses and Wargames, Exercises, and Analysis
Two short papers from the mid 1980's, these provide a good basic introduction to the terminology and the concepts that the military (particularly the Naval War College) applies to wargaming. Most notably, in both cases, wargames are presented in the context of other types of analysis. These papers also demonstrate the way that strategic-level political-military games (of the type this blog spends the most time considering) are viewed as one segment of the broader wargame genre. I found that seeing this sketched out helped me better understand the approach that some military and defense sources take when discussing these pol-mil games.
Wargame-Creation Skills and the Wargame Construction Kit
This paper and accompanying kit are not a complete course in wargame design, but they do lay out some of how the Naval War College approached creating such a course. Some of the material was deleted when the report was declassified, so the kit is no longer a complete wargame in and of itself, but I found a lot of useful ideas in both the report and the kit.
Game-Based Experimentation for Research in Command and Control and Shared Situational Awareness and Gaming and Shared Situation Awareness
These papers describe two series of experiments conducted using a simple computer game (called SCUDHUNT) to measure the degree to which players developed a common mental model of their operational environment. The focus of the game is tactical/operational rather than strategic, but the approach represents a new and fascinating way to study fundamental questions of perception, decision-making, leadership, and other issues, all of which have great relevance when considering human action at any level. In addition to the practical military consequences of understanding shared situational awareness, it seems to me that a better understanding of how these shared mental models are formed (or not formed) would help game designers address the fact that different participants in a game (particularly a large one) will have very different experiences, which could lead to vastly different conclusions being drawn. I don't know if anything has come of the authors' proposal for a game-based laboratory to apply this sort of technique, but I hope something will.