The Pentagon is pouring tens of millions of dollars into mathematical models that might one day help America's armed forces win a counterinsurgency. Too bad the U.S. military is almost totally unprepared to model irregular warfare.
The Pentagon is interested in modeling because it's a cheap, fast way to calculate whether your equipment and tactics will be effective against whatever the enemy is throwing against you. The problem is that, for years, modeling and simulation focused on conventional war with the Soviets. And it hasn't quite adapted to today's guerrilla conflicts, as I discovered when I wrote this article for Training & Simulation Journal. Which means a "Sim Afghanistan" won't be ready for a long, long time -- if it's ever ready at all.
The Army's traditional simulations focused on the sort of kinetic physics stuff that's fairly easy to model: X number of M1 tanks firing at Y number of T-72s at Z meters will kill N number of targets. They were attrition-based systems that tended to favor whoever had the most firepower, but they didn't address the intangibles of war, such as morale, training and leadership.
Now, the military is paying for that firepower focus as it grapples with simulating counterinsurgency. How do you model non-combat factors such as the reaction of an Afghan village to U.S. troops patrolling their streets? How do you mathematically calculate the effects of propaganda, psychological warfare, religion and economic hardship on the attitudes of a civilian population?
What it boils down to is whether it's possible to create a computer model of human behavior. Humans are so complex that it seems virtually impossible to construct a valid model (Wall Street has tried, and look how well it worked for them). The same applies for nations; it would be easier to simulate Iraq as a single monolithic entity, but this would be useless for grasping the essentials of counterinsurgency. The alternative is to treat Afghanistan as a bubbling cauldron of competing interest groups, but simulating each one requires the ability to accurately model a multiplicity of political, economic and cultural variables.
"I don't see any way that we're going to have some kind of 'plug in the data, let it run' simulation and have it spit out a nine month counterinsurgency campaign," said one Army expert.
That feeling was echoed by many of the modeling and simulation (M & S) experts I talked to, who seemed neither confident nor enthusiastic about building human behavior models. Yet daunting as the task may be, they know they will have to try. Because warfare is too expensive to learn the hard way, and we need to give planners and commanders tools so they can predict the consequences of various strategies. But garbage in, garbage out, and a model is only as good as its algorithms.
Creating valid algorithms requires a solid theoretical framework. You can't model human behavior unless you have a viable theory that accurately predicts human behavior. Darpa wants to build a lab to test social science simulations, and I wish them luck. As one Army researcher pointed out, social scientists never agree on the validity of each other's theories, which makes there won't be much consensus on whatever model the military builds.
An overarching one-size-fits-all model of counterinsurgency and nation-building is unlikely. We will probably see a variety of models, plus board games and BOGSAT (Bunch of Guys Sitting Around a Table) seminars. The model-makers know they had better get it right or else. "In our business, your work had better stand up to scrutiny. Because the scab-pickers will show up, and they'll start picking apart your models and sims," said one.
Still, there is some good news, at least at the level of individual soldiers. Serious Games are giving them the opportunity to practice interacting with civilians. In another piece for TSJ, I looked at BiLAT, from the ever-clever developers at the Institute for Creative Technology at the Univ. of Southern California. It's a single-player game, designed to give junior officers and NCOs a chance to negotiate with virtual Iraqis -- without violating too many of their cultural norms. I had a chance to test BiLAT, and after having my Iraqi counterparts storm out of meetings numerous time, I now know why it's good to make small talk before getting down to business, and why it's bad to wear body armor to a conference.
[Illo: Game Production Services]
ALSO:
