Can data-rich war games improve decisions?
- By Patrick Marshall
- Jun 03, 2019
Military strategists have long relied on war games to gain insight into how human decision-making might impact a tense confrontation or actual conflict. The current generation of war games -- in which a small number of military and political officials conduct table-top exercises -- are so limited in scope that analysts don’t get much more information than they might from watching a game of Risk.
"Because you have a limited player set and only play through a few scenarios, you don't get enough data from these scenario-based discussions to draw statistical inference. You may only get an idea of how these specific people would react," Bethany Goldblum, a researcher in the University of California at Berkeley's nuclear engineering department, said. "This is why traditional wargaming is often described as an art rather than science."
Signal, however, is a new cloud-based, multiplayer game designed to collect decision-making data from a much broader set of players. It runs on Amazon’s cloud platform and employs state-of-the-art game engines to deliver a more realistic experience for participants. It was funded by the Carnegie Corporation and developed by researchers from the Lawrence Livermore and Sandia National Laboratories along with a team from Berkeley that included members from the departments of electrical engineering and computer sciences, nuclear engineering, political science and the Goldman School of Public Policy.
Besides amassing data from choices made in thousands of games and a variety of players, Signal also collects basic demographic data about participants to aid in analyzing decision-making behaviors.
"Traditional discussion-based exercises have led to a theory-rich, but data-poor environment, said Andrew Reddie, a team member and Ph.D. candidate in Berkeley's political science department. "Using new tools, we're arguing that board games and electronic games can complement traditional wargaming by providing a science-based, experimental framework for gathering data."
The premise of Signal is simple: Each player controls a hypothetical country and by making decisions about cyber, conventional and nuclear capabilities -- including investments in infrastructure and military resources and forging alliances -- tries to increase that country’s influence. The research question being asked is: How does the introduction of different weapon capabilities affect escalation or de-escalation during conflict?
In the game, players “signal” their intent to build up civilian and military infrastructure or to launch a cyber or nuclear attack and then negotiate trades and agreements with other players to deter attacks in an escalating conflict. The game tracks every move players make and their communications with each other through online chat logs.
During the game, players’ scores are updated so they can see the impact of their decisions. And the game’s pre- and post-game surveys collect basic demographic data about the players that includes their age, education, political affiliation, work experience and knowledge about national security issues. “We do not collect information beyond basic demographic data used in most social science experiments,” Reddie told GCN. “There are no psychometric or alternative testing procedures in Signal.”
The data collected is then anonymized and can, according to the team, be used as input for machine learning algorithms to create models of optimal behaviors given certain experimental conditions. “By tracking demographic data and automatically collecting player and game data in real time, the platform allows for quantitative analysis of the game outcomes," Goldblum said.
In the mood for a little war gaming? Signal is available to the public here from 1 p.m. to 5 p.m. PDT every Wednesday and Thursday.
“Signal offers a first attempt to use experimental gaming to ask and answer social science questions related to international security issues,” Reddie said. He said the team is also looking to test the differences in game play when the numbers of players are increased or decreased and to generate additional scenarios, such as the introduction of drones and other emerging technologies.
Patrick Marshall is a freelance technology writer for GCN.