Less Is Sometimes More: How Distributed Games Can Be Used for Training & Analysis
In today’s fast-paced world, it seems like there is a new gadget born every minute! Just when I was finishing loading my address book into my now 4-year old Palm Pilot with its black and white display, my husband tells me he has got to have the latest and greatest PDA – a Compaq iPAQ H3600 pocket PC with a color display, wireless modem, and the capability to download real-time NEXRAD weather while he flies (he’s a pilot).
Sure, the latest and greatest technology is attractive, and it’s fun to be the first person on the block with the latest PDA, pager, and GPS system, all nestled in an attractive black leather holster that reminds you of Star Trek, but is this really useful?
In today’s simulation world, where millions are spent on the modeling and simulation equivalent of the Ronco Vegematic (it slices! it dices! it juliennes fries!), there are other alternatives to consider.
Consider the thorny problem of determining how a team develops shared situational awareness (SSA). This is quite complicated and depends on many factors, including the team members’ backgrounds and personalities, the team’s purpose, how they work together, etc.
In 2000, DARPA’s Wargaming the Asymmetric Environment program asked ThoughtLink®, together with the Center for Naval Analyses (CNA), to explore this problem. The specific focus of the experiment was how the use of shared visualizations and different modes of communication (e.g., telephone vs. text chat) affected a team’s SSA. One approach we could have taken was to use a complex simulation environment, complete with a real-world scenario and authentic command and control devices, such as GCCS. However, we concluded that this environment introduced so many variables, identifying the key factors contributing to a team’s SSA would be hard to uncover. What we needed was a simple, abstract game designed to focus specifically on this particular problem. Hence, we designed and developed an on-line game we called SCUDHunt.
SCUDHunt
SCUDHunt is a simple game of command, control, communication, intelligence, surveillance, and reconnaissance (C3ISR) played by distributed players over the Internet. The game is similar to the popular game Battleship - but played in teams, with each player playing from their own computers. In SCUDHunt, each player commands one or more information assets and the objective is to work with other team members to find the 3 hidden SCUD launchers on a 5 X 5 grid map representing the fictional country of Korona. For more information about SCUDHunt, see http://www.scudhunt.com
SCUDHunt Gameboard
We designed the game so that the distributed team members would need to share information in order to do well. We also
designed it so that the players' interpretations of the information would be easy to record and measure. In the game board shown above, the left-hand display indicates that the communication intelligence device (COMINT) was placed in Column 1, Row D. The right-hand display shows a shared visualization of the results returned for all of the information assets.
The initial experiment we conducted included six four-person teams. Each team played six different versions of the game, each version using a different combination of communication capabilities and visualizations.
The experiment’s results were very successful – we were able to indicate that both communications and shared visualization greatly contribute to a team's ability to develop shared awareness – and we learned that there was no significant difference in the way the teams communicated (text chat vs. voice). For more information on the results, see http://www.thoughtlink.com/wae.htm
Another important factor was that by using an easy-to-use programming language (Visual Basic), shareware tools, and other low-cost web technologies, this experiment was relatively low in cost – a few hundred thousand dollars vs. millions of dollars. The broader results indicate that simple games like SCUDHunt, designed to target specific experimental goals, are a promising technique for conducting research in the fields of C3ISR, team behaviors and other complex areas such as network-centric warfare and Operations Other Than War.
ThoughtLink® is currently under contract with OSD’s Joint C4ISR Decision Support Center to mine additional data from the SCUDHunt experiment, looking for the key factors contributing to the quality and timeliness of decision making. SCUDHunt is also being used by the University of Arizona in experiments on leadership at a distance.
SCUDHunt shows promise as a research tool - as well as an educational tool. One potential educational application might be to teach the importance of sharing information and developing team processes when working distributedly. In addition to rapidly developing specific distributed games for analysis or training, a great deal can be learned from commercial games. Commercial games tend to push the current limits of technology while still considering cost and marketability – hence, a great deal can be learned from their implementation. The following is a review of a relatively new on-line, multi-player game, called Majestic.
Majestic
I recently played the pilot episode of a new on-line game called Majestic, developed by Electronic Arts/Anim-X. Majestic is an on-line adventure game with a twist. As opposed to persistent state games like multi-user domains (MUDs) where you exit and enter the virtual environment when you want to play – Majestic interweaves itself into your real life, using a wide range of communication modes to contact you and give you game information.
Players (real or virtual) contact you via AOL’s Instant Messenger (IM), they send you e-mails with links to web sites, they send you faxes, and finally (and probably the most eerie for me) they telephone you.
The game has no 3D environment – which some might argue reduces the game’s fidelity. But, Majestic’s creator Mike Griffen summarizes it nicely when he argues “No graphics engine can possibly rival the human imagination when it comes to creating realism.”
Another intriguing twist: the developers have identified how to use intelligent agents and event triggers to move the storyline along automatically. By tracking what web sites you have visited and/or what information you have accessed, the game knows what should unfold next. Although I think they use some real role players, my guess is that most of this game is automated.
The interface for the game is very simple. There is a Majestic player web site where you go to see news, your progress timeline, links to applicable sites, and contacts (an archive of the phone and fax messages you have received). More important is the Alliance Application. This window (shown below) lists all of the other players in your alliance (working on the same conspiracy – since conspiracies are the major theme in Majestic).
Majestic’s Alliance Window
The Alliance Application also includes documents, videos, calls, and other items you acquire as the game develops (on the right-hand side of the Alliance window). It is very helpful to have these accessible during game play to refresh your memory and refer to them for additional clues.
Almost instantly when you start playing the game, other players (whose IM names appear in your Alliance window) are contacting you. Many of the players you interacted with were actually “chat bots” – intelligent agents, programmed to have an intelligent conversation with you. The developers did a very good job at designing the responses so it would appear like these were real players. Majestic’s use of “chat bots” was very sophisticated. This technology might easily be used to mimic synthetic participants in on-line military training and analysis environments.
Majestic’s automated environment that takes advantage of the player’s real-world communication devices seems to me to be a realistic environment both for training and analysis – particularly for the operational and strategic levels where the 3D environment is less important and communication with the other players is more important. I feel this type of environment holds great potential in the operational world as well – to help track complex crises. For instance, during the coordination of a unfolding disaster, event triggers could prompt different user’s on key issues, like who to contact or what resources might be used to address a specific problem.
In Conclusion
I received my early indoctrination to the world of Distributed Interactive Simulation (DIS) while I was working as a Research Staff Member at the Institute for Defense Analysis (IDA). At that time, I could see the many benefits of using man-in-the loop, entity-based simulations. While working on a project for DARPA back in 1995, I spent a great deal of time with Ft. Knox’s Virtual Training Program and saw how effective this form of simulation could be for training tank platoons. I later discovered the benefits of using this form of simulation for conducting analysis – the DIS tapes captured all of the events and made it relatively easy to playback the simulation and evaluate what had transpired.
However, I do not believe that the use of virtual or constructive simulations is a cure all. In many instances, this approach is too complex, too expensive, and clutters what it is you are trying to train and/or analyze.
I have become a big fan of using web-based technologies as a way of doing low-cost training and analysis. Games provide simple and elegant abstractions of complex environments, offering an excellent means for analyzing a particularly hard problem or for providing training focused at a particular training goal or objective.
Oftimes, people opt for complex simulations because they want higher fidelity. But if the problem is a nail, then the solution is not a battery-operated multi-purpose hammer/screwdriver/table saw.
The problems associated with large-scale simulations (like ModSAF, JSIMS, JWARS, JMASS, etc.) are that they often demand sophisticated and expensive hardware, the software itself is complex and expensive, and they generally demand contractor support to set them up and use them.
I argue that web-based technologies, specifically the use of distributed games, can offer many benefits to analysis and training. They can even on occasion offer a more realistic environment and they can save a lot of money. Games are particularly useful for training or analyzing groups in the following areas: cooperation, coordination, communication, risk taking, problem solving, leadership, group dynamics, and team building.
Specifically, my experience has shown that current simulations are not very advanced when it comes to representing Operations Other Than War (OOTW) environments. The areas that are often overlooked in these environments are the soft factors including building trust in virtual teams, learning how to communicate with individuals and organizations with different cultures, knowing what their capabilities and resources are, and building a shared picture (SSA). These things can be easily trained and analyzed using an on-line, distributed environment.
There are many examples of war games being used for analysis and training. To learn more, see http://www.thoughtlink.com/recommend.htm. Earlier this year, MindsimTM Corporation recently published a CD-rom with a complete encyclopedia of war games called WarGames 2000.
The CD also includes contains books and articles, databases, games and rule sets, World Wide Web resources and, an impressive bibliography. The CD is available for $29.95 from MindSim’s web site (http://www.mindsim.com).
Mindsim’s WarGames CD