there are usually models that are designed to adjudicate player interactions. Finally, wargames must provide players with information that they will then use to make their next decision. A wargame that only provides the players the information that would actually be available to them is called a closed wargame. In closed wargames, each side (typically red or blue) is sequestered in its own cell, and the wargame’s white cell makes the determination of what information each side’s collection assets would produce, and then communicates only that information to the appropriate cell. Wargames that allow all the players to have all available information are called open wargames.
And, as for the model or models that are used in simulations and wargames, users must heed the statistician George Box’s warning: “All models are wrong, some are useful.”4
An Abbreviated History of Wargames and Simulations
Before the rise of the computer, the primary form of combat simulation was wargaming. Wargaming has a rich history and has been used by many cultures and in many different forms. The ancient games of chess and Go are but a few of the games that were believed to have usefulness in training and testing military commanders’ decision‐making capabilities, and this belief led to the designing of games focused on modeling combat for the development of military leaders. In the nineteenth century, the Prussians developed Free and Rigid Kriegspiel as methods to educate their officers. Rigid Kriegspiel focused on the calculations of combat, with the hypothesis that good combat leaders had to be able to employ a type of “combat calculus” to mathematically understand what decisions should be made on the battlefield.5 Free Kriegspiel used battle‐tested Prussian officers to assess junior officers’ responses as they were presented with possible combat situations to which they had to react.6
In the early part of the twentieth century, F.W. Lanchester proposed two sets of differential equations that could be used to simulate combat, referred to as Lanchester’s linear and square laws. Although he proposed these laws to simulate aerial combat, they gained traction for use in modeling ground combat. The square law rewards a combatant’s ability to concentrate forces and was seen as relevant to modern warfare, and the linear law has been accepted as a model of ancient warfare where combatants were unable to mass fires.7 Lanchester came to realize, through studying the Battle of Trafalgar, that if a battle can be decomposed into a series of concurrent and consecutive sub‐battles, separated by space and time, then it is more appropriate to apply the square law to each sub‐battle, and sum all sub‐battle losses to gain a more accurate accounting of the entire battle than it is to apply the square law a single time to the entire battle.8 As computer‐based combat simulations were developed in the latter half of the century, many ground aggregate simulations used some adaptation of Lanchester’s laws to model attrition.9
In the first half of the twentieth century, the US Navy made great use of wargaming to examine a potential war with Japan, beginning over two decades of focused wargaming in 1919 at the US Naval War College.10 This detailed examination of war in the Pacific proved to be so successful that, after the conclusion of World War II, Admiral Chester Nimitz said “…nothing that happened during the war was a surprise – absolutely nothing except the Kamikaze…”11
Wargames and Computer‐Based Combat Simulations: From the Cold War to Today
As the North Atlantic Treaty Organization (NATO)–Warsaw Pact “Cold War” consumed the US defense establishment for the latter half of the twentieth century, two events greatly impacted the use and characterization of combat simulations in the US DoD. One was the development of the closed‐loop combat simulation. The earliest computer simulation of ground combat is believed to be CARMONETTE, developed by the US Army’s Operations Research Office in 1953, and used to inform defense decisions from 1956 to 1970.12 The other was the US DoD’s embracing of Defense Secretary Robert McNamara’s “Systems Analysis” philosophy (McNamara served as SECDEF from 1961 to 1968), which led to the thinking that every important defense concept or procurement program required quantifiable justification.13
In the late 1950s, the use of CARMONETTE, a stochastic tactical‐level closed‐loop combat simulation was still referred to as a “wargaming” technique. It was used for research purposes such as “testing the value of new weapons, fighting techniques, or war plans.”14 The developers of CARMONETTE realized that the concept of repeatability combined with modeling some of the processes of combat probabilistically could be used to remedy one of the identified shortcomings of wargaming: the results of a wargame are only a single realization of a range of potential outcomes. They determined it would be necessary to “repeat the battle calculations allowing nothing but the play of chance to vary and so identify the spectrum of the possible outcomes and the associated frequency distribution.” They also realized that the scientific method would be useful if they were to attempt to compare battle outcomes of forces equipped with two different weapons systems; in particular, “we must be able to repeat the battle simulation many times while holding fixed all parameters except that one under investigation.”15
However, these same developers also came to the realization that simulations such as CARMONETTE could not reproduce the complex decision process a military commander uses in combat:
The design of the simulation is such that it can create a realistic representation of close combat during the brief intense engagement phase lasting for approximately 1 hour or so. Continuation of the simulated combat beyond 1 hour becomes unrealistic because a decision fundamental to the execution of the maneuver would undoubtedly be made at that point. CARMONETTE has no capability to reproduce a military commander's mind, and thus the simulation must be terminated, and the results reviewed. The simulation may then continue with an appropriate order from a commander if desired.16
The implication here is that CARMONETTE could be integrated into a larger iterative process where a commander’s human decision‐making process would shape the approximately hour‐long CARMONETTE engagement, and then, after the commander reviewed the engagement results, another human decision would be made that would shape the next CARMONETTE engagement.
In the 1970s, independent of the budding combat simulation community, military wargamers realized that some of the book‐keeping functions that wargames required could be better performed by computers. In 1976, the Training and Doctrine Command Systems Analysis Activity integrated a Wang 2200 minicomputer into the Dunn–Kempf manual wargame to create BATTLE, the Battalion Analyzer and Tactical Trainer for Local Engagements. They used the computer to calculate the results of simultaneous direct and indirect fire engagements, and to perform “book‐keeping functions” such as tracking movements and ammunition expenditures. They realized that the computer could be leveraged to lessen the burden on wargaming staff to free them up to focus on more vital aspects of wargaming such as the tactical decision‐making process.17
In the 1980s, the use of computerized combat simulations became ubiquitous in