Can a Round of Poker Solve Afghanistan's Problems?
by Major Richard J.H. Gash, Small Wars Journal
Can a Round of Poker Solve Afghanistan's Problems? (Full PDF Article)
Analyzing the ill-structured problem that is southern Afghanistan noticeably tests the bounds of traditional military planning doctrine. Identifying and framing the problem, isolating centers of gravity, and even articulating an attainable end-state given the tangle of tribal violence, narco-trafficing, and religious fanaticism can drive otherwise mild mannered planners to the verge of physical violence. Training and Doctrine Command's Pamphlet 525-5-500 goes as far as predicting an expected "lack of professional consensus" when tackling such a "wicked problem"1. Fortunately (or unfortunately), ill-structured problems do not exist solely in the realm of military conflict. Social scientists, political theorists, and economists routinely grapple with their likes. Over the past century they have devoted much scholarly effort toward their mitigation, if not solution. One theory that may particularly apply to southern Afghanistan is that of games. Although one can quickly become bogged down with the mathematics of game theory, a rudimentary understanding of its basic principles can prove quite beneficial to military planners. What follows is a brief primer and simple demonstration of how game theory can be applied to help military planners frame the problem of developing a viable counterinsurgency strategy in southern Afghanistan.
Can a Round of Poker Solve Afghanistan's Problems? (Full PDF Article)
About the Author(s)
Comments
Major, this is a great article, and I truly enjoyed reading it (perhaps showing my geek side a little more than I should).
This is somewhat akin to Monte Carlo analysis, which is used to model particle transport systems, the stock market, games in Las Vegas and games elsewhere. After all, John von Neumann also designed the Monte Carlo process which was used in part for criticality calculations for the Manhattan project.
In Monte Carlo - let's take a particle transport system - one may not intuitively know where the high scores will be, where the low scores will be, and where the poor sampling will be. Adjustments can be made to the system and the game to handle all of those exigencies. Or, one may intuitively know where there might be a high score, but not with certainty. Running the game tells us this information.
Your description of the application of the game to the towns and the Taliban is insightful and applicable, and of course as you say this can become much more complex. But the problems with this approach emerge quickly.
First of all, it presupposes that one has access to game theoreticians who can perform this modeling. Second, it assumes that one has access to good, if not comprehensive, information as input to the game (e.g., the nature of people's concerns, the economy, the nature of the insurgency, etc.). In other words, one needs a full time anthropologist or team of them, as well as very good intelligence. Third, it assumes that the game can be adjusted on the fly, since this is a dynamic rather than a static system. Most models that could be set up at the Pentagon would be too time-intensive, too impervious to change, and untimely in their conclusions. We need this sort of thing on the ground, run in real time. Chances of this happening are statistically insignificant (and we know this without running the game, pardon the pun).
With that said, I still enjoyed the article, and it is this kind of out-of-the-box thinking that we need.
V/r,
Herschel Smith, P.E.