adf logo

Australian Defence Force

Background for the Basic and Advanced Analytic Wargaming Courses from the Naval Postgraduate School in Canberra, 17 February to 6 March 2020


Adjudication is the process of determining outcomes of game events. The most obvious of these are when the moves of players result in contact between opposing forces. This may be only the detection of one element by another (perhaps without a subsequent engagement), or a skirmish between small forces, or even a major clash on the battlefield.

In wargaming, adjudication is often provided by a facilitator. This is particularly true if there is the activity includes a seminar war game. In a seminar war game, a facilitator is typically used to take participants through events in a scenario or vignette. If no pre-arranged rules have been established for adjudication, it may solely up the facilitator to make a determination on results of player decisions.

An aspect of adjudication that may be overlooked relates to non-combat activity, e.g., re-supply, recovery of battle casualties, arrival of replacements, re-fuelling and re-arming of aircraft for subsequent sorties. If the purpose of the war game is to analyze combat results in the short term, such oversight may be relatively unimportant. But if the timeframe is longer, e.g., multiple days, then elements like logistics and reconstitution become more important and there should be appropriate adjudication procedures for them.

Purposes of Adjudication

The adjudication process in a game serves two purposes. The first, and obvious, one is to provide players with an outcome based on their decisions and actions. This is a critical contribution during the play of the game. To be effective in this respect, adjudication must be plausible and timely. The level of plausibility can vary. At one extreme, there may be sufficient plausibility if players are willing to accept outcomes as reasonable. At the other extreme, almost any discrepancy from expectations may be too much.

The second purpose of adjudication is to support analysis. For this, a simple replay of a game, as is usually available following a computer-supported game, is rarely adequate. This will certainly show the unfolding of a specific game but it will be missing many parts that are critical to good analysis, particularly the player reasoning that when into moves, and diagnostic records so the analysis team can backtrack apparent anomalies to determine the source.

Four Types of Adjudication


As Kriegsspiel developed in the nineteenth century, it was accompanied by a rule book of increasing complexity. At a much simplier level, rules might resemble the rubrick "Attacker with 3:1 advantage wins." Rules in Rigid Kriegsspiel were extended to cover combat under a wide variety of extenuating circumstances, but also logistics, and many other elements of military operations.

An extreme case of rule-based adjudication is when a game is played with computer support. When computer support was added to manual wargames it was often used simply to assess the probabilities of engegement outcomes and draw a random number to determine which outcome the players would have to contend with. As hardware and software became more powerful and elements like digitized terrain became available, combat simulations (like the Janus war game) were used to track movement, line of sight, detections, engagements, and damage ("catastrophic", "mobility kill", "firepower kill"). The software behind elaborate combat simulations like this resembles the rules of Rigid Kriegsspiel, although the complexity (captured in algorithms and data) is generally hidden from players who see only a user interface.


As a Free Kriegsspiel, many modern war games are played with adjudication using relatively simple rules. Indeed at the extreme the resolution of events may be left completely to the discretion of the umpire or adjudicator or facilitator. Of course, in such a case considerable onus would be placed on the adjudicator. The adjudicator would have to be highly knowledgeable about all potential outcomes and have appropriate judgment to choose an outcome that players are satisfied is fair.


Matrix games. Chris Engle developed matrix games in the early 1990s as an umpired alternative to more rigid, rules-based games. When a branch point arrives in a matrix game, players make arguments about potential outcomes, the acknowledges the strength of these arguments by assigning probabilities to the various outcomes. Then rolling a dice (or some other randomization process) determines which branch will be taken.

When using the matrix game approach, it may be useful to impose certain protocols for efficient use of time. There are such protocols to keep players from waffling. One of these is called "pros and cons": a proponent outlines the pros for a preferred outcome and other players summerize cons. Another protocol is "three reasons"; players are must limit themselves three best reasons for the outcome they support. Once arguments are offered following the previously agreed protocol the umpire assigns probabilities each possible outcome and "rolls the dice". Players then implement the outcome determined by the random process.

Matrix games can be seen as a consensus method in the sense that players agree that the umpire has assigned acceptable probabilities to the various outcomes. Then it is the roll of the dice that determine which of the possible outcomes will obtain.


War games originated within the military community over two centuries ago. The adjudication techniques have since been adapted to a wide range of endeavors. In practice for any game, adjudication processes will typically be a blend of the three above: rigid, free, and consensus. For example there may be rigid procedures for determining how far/fast a unit might move in specified terrain. Ten some unantictipated circumstance might be adjudicated soley using the umpire's judgement: there are no rules for the unexpected situation and players cannot reach a consensus. But there may be other situations, say the result of a political action, where an appropriate might be easily agreed by all concerned (consensus).

Subterfuge, False Alarms, Misunderstandings, and Fratricide

An aspect of military operations that is particularly difficult to reproduce in war games is the impact of false information. The source of such false information be malevolent, e.g., an information operations campaign by an opponent. But the source may also be relatively innocuous, but with consequences that can be as catastrophic as if it had been designed by a highly motivated enemy.

There are complicated psychological aspects to how false information affects decision making. This is a broad topic that goes well beyond how to replicate the factors in a war game. Interested parties should consult the many references on psychological effects on the battlefield. Needless to say, these effects are still incompletely understood in real life and, consequently, difficult to reproduce faithfully in models and simulations.

Engineering Level Data on Weapons and Lethality

Within the US defense community, the Joint Technical Coordinating Group for Munitions Effectiveness (JTCG/ME) provides detailed algoritms and data for determining the effectiveness of weapons. The group's products are used both operationally in mission planning ("weaponeering" to determine which weapons should be used for engaging a specific target) and also for models and simulations to determine the results of a hypothetical engagement.

Algorithms and data specifially for land engagements are provided by the Army Materiel Systems Analysis Activity (AMSAA), which is the US Army's primary participant in JTCG/ME.

A review of weaponeering provides an unclassified description of JTCG/ME and its methods.

Deterministic and Stochastic Outcomes

In a deterministic model, the result from the model is fully determined by the parameter values and the initial conditions (or "inputs determine outputs"). So, if an adjudication process had the same inputs for two cases, the output would also be the same for both. However, in a stochastic model there is some form of inherent randomness. Rerunning a stochastic model with the same inputs (parameter values and initial conditions) for two cases would typically result in a different outcome each time.

Applying stochastic models to war games is usually implemented by assigning probabilities to potential outcomes then using a random process to choose one of those potential outcomes. For example, we might say the chance of killing a target in an engagement is 1 in 6 (16.7% or 0.167). Then we could roll a six-sided die and claim a kill if the side with one pip turns up. The result for the other five sides would be "no kill".

Kill Lookup In an example at a more sophisticated level we could say there is a 12% chance of a catastrophic kill (sometimes called a "K" kill), a 9% chance of a mobility kill (an "M" kill), and a 14% chance of a firepower kill (an "F" kill). There remains a 65% chance that the damage to the target is not sufficient to take it out of operation (or no kill). In this example, the random number that was drawn was 0.18321; so the result was an "M" or mobility kill.

A stochastic model of weapon engagement can get even more sophisticated, say by determining the trajectory of fragments through the target and the likelihood those fragments will intersect with vital components to the degree the target will be damaged. Another layer of sophistication would allow for cumulative damage to a target: one engagement might be insufficient to affect performance, but an accumulation of "mosquito bites" could take it out of action.

Aggregated Units

Many wargames use aggregations like batallions, brigades, divisions, and other formations rather than individual vehicles or platforms. In such cases individual weapons or sensors are not represented and instead some parameters for strength and weakness must be used.

War Game Counter (example) As an example this counter from a board wargame shows the second brigade ("2" on the right side) of the 10th Mountain Division ("10" on the left side). Across the bottom of the counter are the numbers "5-4-3": by a common convention this means a strength of 5 in the offense, 4 in the defense, and a movement rate of 3 (rules for a specific game should be consulted as not all games use the same conventions). The use of such parameters to determine results is explained in the rules of the specific game.

Determining appropriate parameters can get very complicated. Apart from using the number and effectiveness of weapons and sensors, factors like training standards, target prioritization during engagements, morale, and quality of leadership can come into play. While using parameters for strength indicators in aggregations (along with rules for their application) removes some requirement for good judgement from the adjudicator, the requirement for good judgement is merely moved to the game designer.

Rewind the Clock

When some particular outcome can be traced back to a specific contentious adjudication, it may be appropriate to rewind the clock to the event in question and to restart the game with a different adjudication outcome.

There are several impediments to this approach:

  1. The game apparatus may not allow it; this is particularly true of games that rely directly on computer support or indirectly on computer-supported command and control systems. Time for a program on a computer-based system is generally based on the system time (which is based in turn on the real time). Unless the program was designed with some facility to "play with time", it is notoriously difficult to reset the game time to some specific time in the past.
  2. The players may not be able to "clear their minds" of what happened on the first branch. This can influence their play of an alternate branch.
  3. The analysis team may not be able to keep up with events. It may, for example, be difficult to keep the data for one branch separate from that of another parallel branch.
  4. The time available for playing the game may not allow it. Often the available time of players (and other participants) is fully booked, so there is no spare time to replay any part of the game.
  5. Biases may creep in to the play of an alternate branch. For an example, a player may get into trouble by being too agressive the first time through a game; if part of that game is replayed that player may now act more cautiously.
  6. Keeping track of various replays may challenge the menthal agility of humans - players and analysts. Players (and analysts too) often have trouble keeping track of details for just one run of a game. If there multiple replays are conducted, the participants have to keep several "pictures" straight. Analysts will have to plan in advance to code data with the number of the replay in addition to the time of the event. Activity that humans may associate with some specific event, might actually be from a parallel time line where the various activities seem to be the same.