DARPA is looking for advanced war gaming
- By Michael Peck
- Nov 09, 2017
War gaming has become more popular as the Department of Defense struggles to anticipate what future conflicts might look like. A little judicious war gaming now just might stop bad strategies or bad procurements before they cause disastrous consequences down the road.
War gaming comes in all flavors, from simulating the actions of individual soldiers to designing battle plans for entire armies and fleets. But the hardest war gaming of all is strategic gaming. To model the behavior of an entire nation, with its complex interplay of political, economic and social factors, is a daunting task.
That's why DARPA's Foundations for Strategic Mechanism Design wants to see whether it’s possible to devise a better high-level wargame that will prevent the U.S. from being surprised by the actions of an adversary, or enable the U.S. to surprise an opponent with its own actions. However, the game that DARPA envisions is the opposite of the usual Pentagon simulation: while most military war gaming aims to determine how a given plan might work out if implemented, DARPA wants a game with a predetermined outcome. The game is there to tell the military how to achieve it.
"We would want to shift from a 'simulation' mindset to thinking about the creation of the rules of the game itself," DARPA spokesman Jared Adams told Defense Systems. "For someone that has done a lot of war gaming, this is the hardest part: designing the scenario, objectives and rules of the players to explore certain decisions in an intelligent way. We want to do the inverse problem: given a desired set of strategic outcomes, could you define the rules of the game in such a way that the decisions will lead to that?"
DARPA recently put out a Request for Information to see whether industry and academia can provide solutions that can help the Pentagon solve the problem of how to create a useful high-level game. "War gaming at the strategic level is decision-centric and heavily dependent on both priming of the players and the question construction to elicit meaningful responses," notes the DARPA RFI. "Even when successful, defining strategies that can achieve objectives requires repeated assessment of scenarios that must be carefully constructed. This war gaming 'art' can be complemented by modeling methods to capture details that may influence decision makers (e.g., relative combat power of military assets), but principled inclusion of relevant factors such as adversarial reasoning, information warfare, and economic incentives is lacking."
DARPA envisions a game that encourages players to meet various objectives through mechanisms such as dynamic economic and trade structures, diplomatic alliances, military posture and action, and infrastructure. Yet a big problem with strategic gaming becomes obvious by a glance at the newspaper: while games reward rational behavior such as maximizing a player's score, the behavior of nations -- including the United States -- is often anything but rational.
But DARPA researchers wonder whether advances in fields such as artificial intelligence and the social sciences, will allow a realistic simulation of irrationality. For example, research in behavioral economics has demonstrated that social factors explain consumer behavior as much as "rational actor" models that simply assume that consumers want to pay the lowest price for a product. "If you understand the social contexts, environments, and influences that shape how people perceive their problems they are seeking to solve, and the strategies they use to try to solve them, then their behavior becomes much more understandable," Adams said. "And, if you think of it this way, potentially much more amenable to simulation."
While strategic war games are complex, some of DARPA's solution may not be. The RFI points to the classic board game of Diplomacy, a simple, yet abstract game of negotiation -- and backstabbing -- in pre-World War I Europe. "Simple exploration of mechanisms employing multiple modalities of deterrence could be explored using modified rules within the game Diplomacy."
Michael Peck is a freelance contributor to Defense Systems