Posted by: Christian | February 3, 2009

Afghanistan is a Computer Simulation

Oh dear God… please, no. Noah Shachtman has found some sort of bizarre government pre-tender for a computer model simulation of future conflict in Afghanistan. Noah with the details:

NATO commanders in Afghanistan want a virtual version of the country, to test out battle plans and forecast future unrest. […]

Last last week, NATO began its search for for the newest “simulation capability.” This one should “be able to model the Afghanistan engagement space in the Political, Military, Economic, Social, Infrastructure and Information (PMESII) domains,” a call for white papers notes. With all that information in hand, war planners can then “assess and validate how specific future events or actions could impact on the current situation through the creation and simulation of a hypothetical/simulated environment.”

And if you are really itching to get the details, you can read the government pre-tender here. You have until February 25 to get your proposal in. But don’t bother. I’ve already got you beat. I unveil to you my foolproof conflict forecasting software for Afghanistan:

please purchase me!

Dear government procurement deciders: please purchase me!

My asking price is $US 1.4 billion payable in small unmarked bills physically deposited in my account in the Caymans or at my lobbyist’s office in DC. Idiots.

So yeah, I don’t have Photoshop and the government is still stupid and wasteful. Noah made the obvious point:

Of course, this assumes the program is loaded with next-to-fool-proof data on Afghan economics, politics, and culture — and understands how all those various elements interact. Not bad for a “COTS [commercial off-the-shelf] solution which may require limited development work.”

Agreed. The whole idea is summarized by the government typo:

Description: RFI for possible sources of supply for Provider for ISAF Simulation Capability (Course of Action Analysis) in Afghasnistan

They can’t even spell the name of the country. That’s about as good as it gets. It just goes downhill from there, as you will find in this pdf description of the proposed “simulation tool.” Even the most hardcore empirical/quantitative “big N” sociologist or political scientist who works with computer models would find this to be near-fraudulent. An example: America has been combed over and picked apart by sociologists, anthropologists, political scientists, economists, historians, journalists, etc… There are millions of articles, books, studies, etc… published on America. There is a massive amount of accurate data. There is no language or culture “barrier.” And yet we can not come up with an accurate predictive model for elections, riots, social mobilization, social change, the economy, etc…until it is so close as to not require a complex tool since any person can see it coming. And despite all of this somebody thinks that a model can be built for Afghanistan that will be useful?

The military should invest in its people, not this pathetic excuse for “science.”


  1. That… that was beautiful. I’d fund it.

  2. too many players, both non state and state , USA lacking any cogent strategic or operational vision,NATO other than USA and UK marking time , US diolomats punching holes , passing time,USAID , USACE corruption ridden, how can anything be forecasted.In any case its dofficult to subject wars to models.


  3. Chris, I thought you might react to this. So did Rex Brynen, who knows a thing or two about conflict simulation. Here: Simghanistan?

  4. lol. How on earth are they going to compute Nuristan into a relevant binary pattern? Chaos-mathemathics is the closest I get to describing it.

  5. Actually, societies of Afghanistan’s type are far more predictable than an Open Society with active Civil Groups. CIA models worked fine in Latin America of the 70ies and 80ies. And the CIA has far more advanced tools today, than it did twenty years ago. The reliability of the software, is pretty awesome, in short-term horizons.

  6. nice blog

  7. Actually, I’ve yet to meet anyone in the intelligence community that thinks that software can replace a good analyst on issues of this sort, and I doubt there are many who think that you can effectively model the complexities of a COIN and SSTR environment.

    Software can be used in conjunction with subject matter experts to facilitate identifying issues that might arise in implementing an operations plan, but it would be silly in the extreme to put much weight on its predictive capacities.

  8. O Ye of little faith!

    There is nothing a model can’t do. We can now model the global climate for the next 50-100 years, with 95% confidence. Our models are so good, we don’t even have to worry about the fact that the yearly and 10 year forecasts (so far) have been wrong, by more than the margin of error.

    And indeed, to determine the required carbon reductions, we model not just climate but the next century’s global demographics, economics, energy discoveries, and technology advances. And it’s all scientific!

    Afghanistan would be trivial compared to that.

    Call it the federal government corollary to the precautionary principle: all funds must be spent, in case they take the money away.

    OUTSTANDING blog, by the way.

  9. If their proposal is for COA analysis alone, then it makes sense. COA analysis is another name for wargaming. The object of wargaming is to provide planners with additional “what ifs” that enable comparison and decision. It does not take the place of visualization or design, it merely provides a more detailed understanding to flesh out multiple possible outcomes rather than most dangerous or most likely ECOAs. Simulations cannot replace that aspect.

    Love the blog.

  10. Most poeple in Afghanistan think the the western world are operating in the country with a long term calculated strategy, and that they know every hole and rock of the country. now I realize they don’t know a thing. I really feel scared knwoing that the future of this country is going to be in the hands of a computer program. Oh my God! is this really what modern civilization can do to poeple?

  11. […] What weight should one put into the above “code”? Certainly not this much. […]


%d bloggers like this: