1 st -Person Tactical Shooters: COTS Games with Military Training Potential?

Transcription

1 1 st -Person Tactical Shooters: COTS Games with Military Training Potential? Michael Barlow; Peter Morrison and Adam Easton School of Computer Science University of ADFA Northcott Drive ACT 2600 Abstract: One of the more popular genres of modern computer games is that known as the 1 st -Person Shooter. In a 1 st -Person perspective view players control heavily armed characters fighting in a fully-realised 3D world inhabited by multiple participants. A relatively recent innovation to the genre is that known as tactical (or squadlevel) shooters; these have more detailed combat physics models and an inherent recognition of hierarchies of command and the section as a unit. The paper discusses a preliminary evaluation of one such game Operation Flashpoint. Scenarios were designed for the game based on small unit tactics training- with the game being enhanced to support Australian Army units, operations up to company level, and other dynamic behaviours not in the original game. The system was demonstrated to a number of Lieutenants of the Australian Army who had recently completed tactical training and their responses elicited. The article concludes with a number of recommendations. 1. INTRODUCTION This paper is a preliminary study of the potential military training aid offered by a COTS (Commercial Off The Shelf) game. The potential, and associated risks, offered by employing computer games for military training or analysis are not yet well understood. However a number of such applications of computer games are beginning to emerge [1,2,3,4]. The article has the following structure. The following section outlines the features of the First Person Shooter genre, and more particularly that of the tactical 1 st person shooter. The next section details the investigation of the game Operation Flashpoint - giving details of its features, the scenarios that were developed as a test of its versatility, and the subjective evaluations performed. The article concludes with a discussion of the potential this genre of game may offer and the challenges that must be met for that potential to be realised. 2. THE 1 ST PERSON SHOOTER GENRE One of the more popular genres of modern computer games, accounting for 3-4% of sales [5], is that known as the 1 st person shooter. The genre was started in 1992 with the famous title Wolfenstein 3D from ID Software [6]. That release created a new game genre, known as the 1 st person shooter, that has lasted a decade, and for which many titles are available and in development at any one time. First-Person Shooters (FPS) share a number of common features a 3D environment is shown to the user from a first person s perspective; that is through the eyes of the character they are controlling. FPS games are action (fast paced) titles in which players invariably control a heavily armed character in a hostile environment both computer controlled forces and, potentially, characters controlled by other human players seek the destruction of the player s character. Other common features in modern FPS games include a range of weaponry and equipment that the character may be armed with, increasingly sophisticated physics simulation engines, multiplayer support, and AI (Artificial Intelligence) opponents - computer controlled characters. FPS games differ in the setting - the backdrop against which the action takes place; these range from fantasy worlds through modern settings to futuristic and science fiction. A currently popular setting is modern military or paramilitary. Examples of this include Operation Flashpoint from Bohemia Interactive Studios and Ghost Recon from UbiSoft. Further, popular titles such as Quake III and Unreal Tournament attract an active and dedicated community of users interested in extending the game. These extensions are known by two names. Skinning is the activity of adding new objects or altering the surface look and feel of the game. Modding is more sophisticated and involves employing the game engine in new ways - new types of settings and ways to play the game. Currently Urban Terror mods - those focusing on modern settings and equipment, with lethal combat engines, are very popular. 2.1 Tactical 1 st person shooters While a number of the features of FPS games make them a possible training or evaluation tool for the military, a number of other features of the genre have meant their full potential in this regard has not

2 been realised. Increasingly sophisticated combat physics engines have emerged ones that semiaccurately model modern ordinance and its effects such as the ability of a single round to kill or of structures to be demolished. However play has typically revolved around a style known as deathmatch players roam a set of rooms and corridors killing all other characters they encounter. A sub-genre of the FPS that has recently emerged is that known as 1 st Person Tactical Shooters (FPTS), or alternatively 1 st Person Squad Shooters. Examples of games in this area include Tribes and Tribes 2, Operation Flashpoint, Rainbow Six, and Ghost Recon. These games have built-in support for team play at the section or squad level. They also tend to support realistic and open environments, have a richer set of interactions between the characters and the environment (e.g., vehicles that can be entered and controlled), and more realistic combat systems that standard FPS (single shots can kill, damage to different parts of the body incapacitates in different ways). FPTS appear the most likely candidates of the FPS genre as a training tool for the infantry. The game engine more closely supports the level of realism demanded for military training and the settings more closely match those of real life and the potential scenarios used in training infantry. 3. OPERATION FLASHPOINT In order to determine whether a FPTS could be employed as part of infantry training a pilot study was conducted. The most likely candidate from the current pool of FPTS was selected. That game was then analysed in two regards. First, how easily, and to what extent could the game be tailored for a military training environment and scenarios? Secondly, just how effective was the modified version? The game that was selected for evaluation was Operation Flashpoint (OFP) [7]. Released in mid 2001, the game takes place against a mid-80s coldwar in which America and Russia have begun a limited conventional war. The game has won numerous awards for its playability and veracity. Due to the modern equipment available to characters, the open outdoor environments, the realistic game engine, and powerful editing capabilities the game was selected as the best possible candidate currently. Indeed this closely parallels a recent decision by the USMC to employ a version of the game for training [2]. To that end a number of copies of OFP were installed on dedicated desktop machines. The machines were networked and equipped with headset microphones for inter-player communication. 3.1 Evaluation Scenarios Operation Flashpoint comes provided with a number of scenarios for single and multi-player play. These scenarios are intended for entertainment, and while well designed are not necessarily true to military doctrine little alone specifically targeted to test the military training potential of the game. In order to examine this aspect a suite of scenarios was designed. OFP possesses a number of powerful scenario creation features that enabled specific doctrinebased scenarios to be created. The game is equipped with a scenario editor that allows the placement of units on the map and basic scripting of behaviour. Movement can be programmed as a series of waypoints, markers can be placed, and triggers can be used to initiate new actions when game events occur. A powerful scripting language is also part of OFP. The language has many features of a 3 rd generation programming language with many game specific operators. With the scripting language it is possible to program complex behaviours and actions not incorporated in the original game. Finally, it is possible to modify the multimedia aspects of the game adding news skins textures for the 3D models, or sounds. All these features make it possible to extend OFP in new directions, and thus potentially explore its utility for military training. Three different scenarios were designed. Each scenario represented a different tactical operation assault by a mechanized platoon, a defence or fighting withdrawal, and a special operation in the form of a search and destroy. Each scenario also served a different secondary role in terms of exploring different game, or added features. In one case heightening the sense of immersion and relevance to Australian soldiers, in others allowing what-if scenario analysis or high-level control Mechanised Platoon Assault Scenario The first scenario constructed and also used in later evaluation was a single-player mechanized platoon assault. The scenario was entitled Operation Axe Point. The player controlled one of the sections involved in the assault and careful attention was paid to ensuring that the briefing and mission itself were conducted according to Australian Army doctrine. To this end considerable effort was expended in not only structuring the attacking and defending forces appropriately but also in heightening the sense of reality or immersion of the mission. Military map symbols were created and utilised for the 2D map based briefing, which was also performed according to the military appreciation process. Further, models of Australian Army units, not in the game, including small arms, cammo, and vehicles were employed. Radio traffic in support of the

3 assault was recorded with Australian voices and the squelch and noise transmission qualities of the service radio equipment was added to those recordings. It was found that OFP supported a mechanized platoon assault scenario with a high degree of doctrinal veracity and one that was made directly pertinent to Australian soldiers Fighting Withdrawal Scenario The second scenario constructed was a multi-player fighting withdrawal. The scenario was entitled Forward Echelon and involved a platoon facing a company level assault ahead of which it had to withdraw in good order. Players could play one of the three section leaders, or indeed one of the individual section members. Emphasis was placed in the scenario construction on checking the game s ability to support the coordination necessary between the different sections in order that the withdrawal was made in good order. To that end voice communication via headset microphone was found to be particularly useful for player coordination. Further, a sophisticated script was incorporated in the scenario that allowed a human player to control a section of which they were not a member by employing the in-game radio. This allowed even a single player to control the entire platoon as well as supporting multi-player play in which one or more sections did not have a human player as member. The script was found to be very powerful but limited by the simple model of radio communication that the game supported Special Operations Scenario The third scenario constructed was a multi-player special operations mission. Players could play either side in the mission on one side a section that had penetrated deep into enemy territory and needed to escape. On the other side platoon members of a larger search to find and destroy the insurgents before they escaped. The mission was entitled HQ Discovered. Emphasis was placed in the scenario on checking the game s ability to support both cooperative and opposed team play. The scenario was designed to allow each side to attempt a number of approaches in achieving their goal supporting a form of whatif analysis. Special scripts were written that extended the game in a number of ways. The insurgents could call for a dynamic helicopter extraction, or summon a boat to one of several pre-defined pickup points. The searching side had a number of computer controlled units involved and scripting was employed in order that these units would report any contact they made with the enemy. In order to make pursuit of the computer controlled insurgents (if no human was controlling them) more interesting a script was employed so that different escape choices and routes were selected each time the scenario was replayed, making the computer oppnent unpredictable. 3.2 Subjective Evaluation In order to complement the constructive evaluation discussed in the previous subsection, a subjective evaluation was also performed. A group of twelve (12) Australian Army lieutenants was enlisted All officers were graduates of RMC, and familiar with infantry minor tactics and low level operations. Further, all had experience with computer-based training such as the Weapons Training Simulation System (WTSS), and all were familiar with the operation of computers and computer games. The evaluators were systematically guided through an introduction to and evaluation of the game. On average this took 1-hour per responder, at the end of which period the evaluator completed a series of 13 questions pertaining to the game. The evaluation process comprised an introductory briefing followed by hands-on play in a simple scenario that taught movement and firing skills. That was then followed by play in 2 of the 3 scenarios designed as part of the constructive evaluation - the multiplayer search and destroy mission known as "HQ Discovered", and the single player mechanized platoon assault known as "Axe Point". The culmination of the evaluation for each participant was the completion of a questionnaire. The questionnaire comprised 13-questions pertaining to the training potential of Operation Flashpoint. For each question responders selected between one of 6 responses ranging from very poor (-5), through poor (-3), below average (-1), above average (+1), good (+3) and very good (+5). Evaluators were also provided space to write comments on each of the questions. The questions fell into two categories. The first group pertained to the veracity of the game - its ability to realistically model low level operations. The responses to these questions are found in Figure 1. The second category pertained to the training potential of OFP. The responses to these questions are summarized in Figure 2. Table 1 is a complete summary of the 13 questions showing the mean and standard deviation of responses to each.

4 Question Mean σ Simulation of Physical aspects Sense of Scale of Combat Env Utility of 2D Map 4 Tac. Appr Realism of carrying capacity Veracity of command structure AI decisions & actions Section Level potential Platoon Level potential Company Level potential Tactical Training potential Teamwork & subunit training Battlecraft training potential Training Aid potential for Australian Army Table 1: Mean and standard deviation of responses to the 13 questions asked of the 12 army lieutenants in the subjective evaluation. Responders chose from one of 6 possible responses ranging from very poor (-5) through to very good (+5). The fourth question of the veracity set was How realistic is the carrying capacity of individuals in the game? Some participants believed that the carrying capacity was unrealistic, whilst others thought that it was accurate. The primary concern of a number of participants was that the speed of movement appeared unaffected when heavily loaded. According to doctrine, units in the game carry more than in reality. The first group of questions numbered six and pertained to the veracity of the game - its ability to model the physical aspects of a low-level infantry engagement, and provide a sense of immersion within that environment. Those results are summarized in Figure 1. The first question of the veracity set was How effective is the game in simulating the physical aspects of combat? Generally responders thought that the physical representation of combat was good. Weapon effects were considered accurate, especially for those of an explosive nature. Damage from projective weapons was seen as too weak participants thought that a bullet would affect aiming and movement to a greater extent than in the game. Some also thought that the movement speed of units in the game was unrealistically fast. The second question of the veracity set was How well does the game provide a suitable sense of the scale of the combat environment? The officers were sometimes put off or disoriented by the computer screen which affected their sense of scale, in that they were unable to gain an accurate sense of scale. Generally people thought the size of objects were realistic compared to landscape features of the terrain. The third question of the veracity set was How effective is the 2D map provided in facilitating the tactical appreciation process? Participants commented favourably on the 2D map within the game, primarily because it was of so much higher detail when compared to similar first-person games. The topographic nature of the map suited RMC graduates who were used to that form of terrain representation. However, the inclusion of every tree and bush on the map was seen as unrealistic. The map was viewed as an appropriate planning tool for scenarios within the game. Figure 1: Subjective evaluation by 12 Australian Army officers of the veracity of Operation Flashpoint. The 6 criteria were physical representation of combat, sense of scale of the combat environment, utility of the 2D map, realism of carrying capacity, match of command structure to field commands, and reality of actions of the AI. Responders rated each criterion on a 6-point scale from Very Poor to Very Good. The fifth question of the veracity set was How closely do the commands available to you as a commander in OFP match the commands you would employ in the field? [Generally, participants thought that all of the required commands were present, given that OFP doesn t support intersection groups. The method of target indication was considered accurate, although it differs slightly from the methods used in reality to identify the position of features on the ground. Some participants suggested that the commands must match those used in reality to a greater extent for the game to be really useful. The sixth and final question of the veracity set was How well do the tactical movements and actions of the computer controlled section members resemble those you would expect from trained soldiers in the field? The animations of characters firing and moving were considered realistic. However, tactical

5 movement in reality is done in fire teams, pairs or groups, and as such this was not simulated. The characters used cover and concealment where available which was viewed favourably by the participants. Actions such as reloading and changing weapons were completed slightly faster than in reality. Six questions were also asked in order to gauge the training potential of OFP - its applicability at various tactical levels as well as to particular tactical skills. Those results are summarized in Figure 2. Figure 2: Subjective evaluation by 12 Australian Army officers of the veracity of Operation Flashpoint. The 6 criteria were effectiveness for section, platoon, and company level training, effectiveness as a tactical training aid, effectiveness for treamwork training, and effectiveness for training battlecraft skills. Responders rated each criterion on a 6-point scale from Very Poor to Very Good. The first question of the training potential batch of questions was How effective do you believe the game could be in representing section scenarios? OFP was seen as a good medium to represent section scenarios. Command techniques within the section were easy for most participants to pick up, and accurate movement of section members was possible. However, the lack of correct section subgroups (gun, assault and scout groups) was commented on. In reality groups within the section support one another, for example, during an assault the gun group may cover the scout and assault groups moving towards the objective. This is impossible to achieve in OFP. The second question of the training potential batch of questions was How effective do you believe the game could be in representing platoon scenarios? Participants thought that Operation Axe-point, one of the designed scenarios, was a very realistic representation of a platoon assault, and that OFP can represent platoon scenarios well. However, controlling sections via the radio the work-around employed to lift OFP above section level by simulating a platoon commander s role - was considered very difficult by the participants. This was due to the limitation of the radio being the only form of input available in the game. The third of the training potential questions was How effective do you believe the game could be in representing company scenarios? Participants thought that OFP would not be good at representing company scenarios. The primary reason is that on a 19-inch monitor situational awareness is not very high. Although it was considered feasible to participate in a company assault, platoon level scenarios were considered the highest level that OFP could effectively simulate. The fourth of the training potential questions was How effective would OFP be as a training aid for tactics? Participants commented that OFP would be a good training aid for tactics. The high-detail map combined with realistic contacts and weapon effects was seen as suitable for representing infantry, armoured and combined tactics. The fifth of the training potential questions was How effective would OFP be as a training aid for teamwork and sub-unit cohesion? Generally, participants suggested that OFP could provide an above average training aid for teamwork and subunit cohesion. Situational awareness was raised as a limiting factor, as situational awareness on the battlefield is seen as very important. Voice communications over the game was viewed as very beneficial when working in a multiplayer environment. The last question of the training potential set was How effective would OFP be as a training aid for individual battlecraft skills? Some participants thought that individual battlecraft skills such as fire and movement, choosing appropriate cover, battle drills (ambushes, moving in harbours etc.) and action on contact could be improved by OFP. Weapons drills were noted impossible to simulate in the game. Aiming of the weapon, however, was considered fairly realistic (for example aiming off when firing at a moving target and long-range targeting). Finally, a general summative question was asked concerning the suitability of the game as a training aid for the Australian Army: How effective do you feel OFP would be as a training aid for the Australian Army? The results are summarized in Figure 3. With a mean response value of 3.8 most participants felt it was a good to very good option. Participants enjoyed their experience with OFP and

6 felt that it could be used in a number of pertinent training areas, both at the soldier and officer levels. Much work remains to be done in this area far larger studies in which such games are actually employed as part of training are necessary. New games with more sophisticated features are emerging all the time. Finally, the risks posed by employing these games for training also need to be studied. 5. ACKNOWLEDGEMENTS The authors would like to acknowledge the assistance of Bohemia Interactive Studios in providing the models of the Australian units. Further, the authors wish to thank the participants in the subjective study. Figure 3: Subjective evaluation by 12 Australian Army officers of the training aid benefit potential to the Australian Army of utilizing a game such as Operation Flashpoint. Responders rated the potential on a 6-point scale from Very Poor to Very Good. 4. DISCUSSION The paper has presented a preliminary evaluation and study of the potential infantry training benefits offered by a COTS (Commercial Off the Shelf) FPTS (First Person Tactical Shooter) known as Operation Flashpoint. A number of scenarios were created to investigate the ease and extent to which the game could be altered to suit infantry training. That was followed by a subjective evaluation of the game and its potential by a dozen Lieutenants from the Australian Army. Using a combination of the game s own setting and objects, the in-built scenario editing capabilities, and non-trivial programming in the game s scripting language it was found possible to create attack, defence, and special operations scenarios comparable to those studied as part of an infantry minor tactics course. Indeed, the fact that the game offered the powerful mission creation and scripting capabilities was vital in the success of this aspect of the study. The subjective evaluation highlighted a number of areas in which participants felt the game had significant potential as a training aid. In particular training for section-level tactics and operations was seen as one area where the game could be of aid. All participants believed the game had potential as a training aid for the Australian Army. 6. REFERENCES 1. Robinson E. (1998) The Pentagon Finally Learns How to Shop, Business 2.0 Magazine, tml 2. Virtual Battlefield Systems Homepage, 3. Appleget J.A., and Illingworth J.E. (2001) Land Warrior Training Initiative, SimTecT 2001 Conference Proceedings, p Smith J.D. (2002), Private Communication 5. Interactive Digital Software Association, 6. History of ID Software, 7. Operation Flashpoint Homepage,