What Makes Heroic Strife?

The Economist offers a piece about breakthroughs in modeling civil wars.  I, for one, am highly skeptical that our wunderkinder can do anything of the sort that is truly faithful to the massive complexity of human interactions.


For in the war-games rooms and think-tanks of the rich world’s military powers, bright minds are working on the problem of how to model insurrection and irregular warfare. Slowly but surely they are succeeding, and in the process they are helping politicians and armies to a better understanding of the nature of rebellion.

Your rating: None
Tags : civil wars, modeling, simulation


War is forcing some other govrnment to submit to the will of a foreign power. COIN is forcing some populace to submit to a situation it finds intollerable, and that can be by either a foreign (resistance insurgency) or domestic power (revolution or separatist insurgency). The difference may seem subtle to many, but it is significant.

Too often where foreign powers struggle most is that they assume that the ability to succeed at the former automatically converts to success at the latter, but the two are completely unconnected. If what one forced through war onto some foreign government is deemed acceptable to the affected populace, than one's COIN efforts are likely to be relatively easy (consider US post WWII occupations of Germany and Japan); if however, what one has forced through war (or coercion or undue influence) onto some government is deemed unacceptable, then no amount of COIN is likely to result in peaceful, natural stability (such as in Vietnam and in Afghanistan).

Understanding this fundamental aspect of the nature of such conflict is far more important than possessing a sophisticated process for modeling the character of the resistnace to one's actions. "Might" may well "make right" in State on State conflicts, but will only buy one a ticket of admission to an insurgency. Our frustrations in recent conflicts are not due to our inablity to assess WHAT was going on around us, but rather due primarily to our unwillingness to accept WHY it was going on around us. This not because insurgency is more complex or complicated or dangerous than war. It is just because governments refuse to think of insurgency differently than war.

When we seek to control the outcome, and pick an outcome that is unacceptable to a significant portion of the populace, there arises an unsuppressable energy of resistance. One can suppress the fighters through the applicaiton of sufficient effort, but such effects will only be temporary until such time as that resistant energy subsides. Understanding what it is the people are resisting and why they resist gives one the best appreciation for how strong and durable that energy is apt to be.

In Afghanistan my assessment is that resistant energy is growing stronger. The more we push against it, the stronger it grows. Assess this energy, not one's objective metrics of progress in programs designed to suppress it. The programs are growing. So is the resistant energy.

Ha! Brilliant. I read this today in the magazine. It reminds me of the mass number crunching of data that was done in the Cu Chi area of Vietnam. Back then, the DARPA boffins were selling the idea that they could predict which way insurgencies would go just like the weather...decades later, we see the same crap being peddled in Afghanistan looking at metrics like the price of fruit at the market, which is a valid metric of stability, but we don't need some Harvard "genius" to tell us this. And now we get this. It's dangerous: the opposite of the NeoCon arrogance that took us to Iraq- this is techno hubris: intelligence as a platform for arrogance.

Peter, well done for bringing this future porn nonsense to our attention. We are right to be wary.

I will push back a little.

I do not perceive the work surveyed in The Economist as having any particular linkage to a specific political ideology, be it on the right (e.g., neocons) or left (?).

It is also unclear to me whether the work surveyed has much to do with the data crunching performed by MACV or McNamara's Defense Department. (Although, upon proofreading this comment, I somewhat anticipate a rejoinder to this paragraph/sentence.)

I think the best epistemological objection to some of the ideas expressed is that people are going to engage in modeling, broadly construed, in some way or another; sometimes intuition is faulty. Thus it makes sense to test what seems logical but may be inaccurate, and to press one's logic, with as much rigor as possible.

I think, furthermore, the real questions (not well addressed in the article) are: What insights (if any) does this work provide? What explanatory or predictive power does the work presented have? Is that explanatory or predictive power better than what currently exists? Does the work presented yield any counterintuitive, surprising and/or novel results?

Some similar questions (think of this as akin to a Quantitative Manifesto of sorts, to use the same grandiose nomenclature employed by Abu Muqawama some time ago): What data does the work presented require, and how robust (e.g., valid and reliable) is it? What special software packages or programming languages must one know or utilize in order to replicate the models? In simple terms, what is the basic mathematical logic at play (e.g., regression to the mean)? How do inputs (independent variables) relate to outputs (dependent variables)? What is the basic causal logic being put forth? What is the likelihood that any associations (correlations) are spurious, both in mathematical terms (e.g., p-values) and plain English (why would one expect to find the associations/correlations observed)?

I think The Economist's journalism here is not great. But I think the potential worth of the work surveyed can - potentially - still be defended. In sum, the jury is still out.




Adding to your post if I may, it is the idea that if we only wargamed the thing better we would have had a better understanding and would have won. Shoot i bet you and your boys understood your area quite well, yes? But of course in war armies can learn and adapt all they want and understand perfectly the war they are fighting and still the war might be lost.

This is what we get from the learning-and-adapting fetish (aka soup with a knife) and the idea that if the big stupid conventional army only LEARNS coin then we shall win. This Coin learning and adapting grammar still dominates much thinking in the Army, and heck it is all over the pages of Paula Broadwell's book on General Petraeus.

I will go out on a limb and bet they didn't understand their area well. They probably didn't understand it anymore than we understood our areas in Iraq or Afghanistan. No matter how many times we walked the ground and spoke to people living their through an interpreter, our understanding of what motivated them was shallow and often wrong. At best, over time, in some areas we could discern enemy tactics such as likely ambush sites, where they would launch their motars from, etc., but we didn't understand the people, and why they fought.

Instead of gaining understanding we applied Western ideas and thought they were fighting because there was no electricity or schools, and if we just fixed that the people would love us. 10 plus years later, too many in our ranks still embrace these ideas despite overwhelming evidence this approach has failed. With or without technology we will continue to fail to understand our operational areas from a human perspective if we can't limit our preconceived biases.

On the other hand technology like SCARE (in the article) has helped analysts determine the location of targets that were not previously identified using non-technical analysis. I agree with Zen, don't throw the baby out with the bathwater, but also maintain a realistic view on the limits of this technology. I have seen a few successes over the past years, I have also seen quite a few "so what" insights, but the technology, and more importantly the techniques for using it will improve over time. It will help enable targeting, but I have seen no evidence it will help us understand the conflict and the people, and as the article states, recommend courses of action. That is a bit worrisome and is the ultimate form of micro-management. I'll produce a program that view the world the way I do and program recommended generic solutions. It will be simplistic at best, software crunches the data and notes several attacks around City X, determines people in City X must support the insurgency, computer model states you must win hearts and minds, recommends you build a school, take a picture of smiling kids, down play continued fighting, focus on smiling kids, show the media the picture, tell them we won this city. Continued attacks to not compute :-).

It is pretty clear from the sources (primary and secondary) that the US airborne drops in Normandy were quite effective in disrupting German operational response to the Utah beach landings, which is what they were designed to do all along. The irony though is that was caused the severe disruption to the Germans was the chaotic and unplanned dispersion of the drops. The German high command thought the Americans were doing some kind of new high-fangled swarm tactics and could not figure out what they were up to with paratroopers falling all over the Cotentin peninsula. So even when things dont go according to plan things sometimes work out anyway. Such is war. Drop me a line at gian.gentile@usma.edu and i can offer up some specific sources if you wish. With regard to the German airborne drop at Crete, well what the Germans learned from that operation was the futility of putting elite infantry up in a symmetrical way against defending infantry without any advantages in firepower and protection where they were hammered and hammered hard. This should be important for the US Army to consider as it moves toward a light infantry dominated force and gives up the traditional advantages of American firepower and protection.

agree, in no way am i suggesting that modeling--aka wargaming--is not important and needs to be done well. As a graduate of Sams and a former planner in the 4ID I appreciate the value of modeling of the enemy and conditions when done correctly. I just grow weary of this never-ending and absurd reduction of higher end combat to the simple, like any dolt can do it.

'All models are wrong, but some are useful'

Mathematical models are all around us and grow in strength every day. Piezoelectric fuel injectors rely upon such models to run the diesel trucks that delivery daily food and fuel to our stores of choice. FEMA uses one and two dimensional hydraulic and hydrology models to predict flooding which provide information for those who are not interested in living in flood zones. Finite element modeling helps us build structures (skyscrapers, etc) that protect us from the elements. Computational fluid dynamic modeling helps us to find where chlorine disinfection is not optimal in our water treatment plants and ensures that our drinking water does not sicken us. Everybody knows what Google is capable of and how our lives have changed because of it...

For those still unsure about the power of models, and with practical experience in an arena such as war, some common sense questions about how a 'war' model's predictions were calibrated with real world events will provide some idea about the actual efficacy of the model.

Models do not absolve leaders of responsibility, but when used wisely they can increase the chances of making an informed decision. For those who read history, perhaps they remember the stories of the many european troops in WWI who thought they would be back home in just a few months? For those of us who experienced OIF1, remember how many of our young troops thought we were going back home 'early'? Hopefully the models described in the article, and other models, will help leaders make better decisions about the necessity of war.

While I generally agree with the criticisms offered here by Gian - while adding that computer modeling tends to be oversold whenever/wherever it is applied - let's not throw the baby out with the bathwater:

"RiftLand is being developed on the navy’s behalf by Claudio Cioffi-Revilla, a professor of computational social science at George Mason University in Virginia. It is specific to the part of East Africa around the Great Rift Valley (hence the name). That this area includes Congo, Ethiopia, Rwanda, Somalia and Uganda, each of which has been the scene of present or recent civil strife, is no coincidence. But the ideas involved could be generalised to other parts of the world, with due alteration for local conditions."

Local conditions are usually the rub with universal rules :)

This is a tool that seems useful in the sense of adding another kind of "rule of thumb" for a commander to consider, and I suspect it would be a far more useful enterprise if the computer scientist was joined by associates with area/local knowledge and military command/planning expertise. You'd get better answers wth the right questions using the right variables. It's not unhelpful to know where subgroups like to congregate (young males, sectarians, wheeler-dealers, criminals) where traffic ebbs and flows, patterns of recurrent social conflict.

The danger is some CYA-minded fool ordering the program determine or drive military planning vice just being an interesting data point to think about.

Moving the past the debate on whether regular or irregular warfare is more complex and bringing the debate back to computer aided wargaming. I don't think the article addressed wargaming, and I'm thankful for it, because I agree with the danger that Zen pointed to in his last sentence. On the other hand, using computational modeling to assist predictive modeling has already proven to be value added in irregular warfare. The key is to success is having a lot of historical data to derive trends, and we don't have that type of data in most places. Some police organizations are using computer modeling to help identify problem areas/times and adjust their patrols to prevent instead of respond to crimes. Most of the efforts I have seen, and I have seen quite a few are largely tactical in character.

What may be promising as an aid to predicting to potential flare ups is the ability of some systems to pull multiple forms of open source data (social media, news media, etc. to discern potential risks.

Using computers to crunch available data to gain insights that may not otherwise may be discernable is a worthwhile effort. To use computational models to predict the outcome of battles (reglar or irregular) appears to me to be a methodolgy that will almost always be dangerously misleading.

ADTS: I know Dave and I keep pounding this drum, but it is important to keep pointing out that war in general is difficult and the graduate level; however the purveyors of Coin have concluded that it is the opposite and quite easy. It is not. If we are not careful with this kind of thinking combined arms skill in conventional warfare will be kicked under the rug because folks assume that it is easy and therefore it doesnt need work and training.

Carl and ADTS: The best book on Omaha Beach planning is Adrian Lewis's "Omaha Beach: A Flawed Victory." It is true that Bradley and his planners were aware of Pacific amphibious operations just as they were quite well aware of army amphibious operations in the Med. What Bradley and his planners did, after considering these different experiences, was to construct an "out of the box" landing and assault doctrine for Omaha. They reasoned that what they were dealing with was so different and knew it required jettisoning other experiences and methods. In the end this "disruptive" and different doctrine and assualt plan assumed too much about the efficacy of air delivered fire power, of tactical suprise, and of an overly tight timeline that offered no flexibility at all to move quickly across the beach. Instead of coming up with something different, they should have used their experiences in the Med and in the Pacific. To me it teaches that "disruptive thinking" or out of the box thinking is not by rule always the answer.

Peter: Agree, I dont trust the whiz kids either.

Gian Gentile:

Thanks. Along a somewhat similar line, as long as I am asking, and you are willing to provide book recommendations: Are there any books that evaluate the effects of the air drops into Normandy, or (relatedly) evaluate *critically* similar operations in World War II (e.g., Crete, Netherlands/Market Garden, etc.), or even in conflicts other than, or including, World War II? I suppose what I am really asking is whether there is literature asking the question: Do mass air drops "make sense"? My initial thinking is that the comparative ease of securing Utah Beach and the surrounding area may well have been due to the actions of the 82nd and 101st Airborne Divisions, but I would be curious to know how carefully this has been established (or disproved, for that matter). For example, it may be that the real salutary effects of the 82nd and 101st Airborne Divisions transpired by accident rather than design (e.g., by dispersing forces across such a wide area as to confuse the German command and control structure as to Allied dispositions). Once more, I would be very curious as to whether someone has examined this with some rigor. I think I've seen some argumentation to this effect, but more in passing rather than at length.

Thanks once more.


First, I think the Go-Gos said roughly the same thing some time ago as Gian Gentile and Dave Maxwell: "It look[s] so easy, but you know, looks sometimes deceive" ("Head Over Heels").

Second, I, like Carl, would like to know more about interservice (e.g., Army and Marine Corps) and intraservice (e.g., Army in the Western Pacific and Mediterranean theaters) and perhaps even transnational (e.g., Krulak's observations of Japanese operations at Shanghai) development, dissemination and diffusion of WW II amphibious warfare doctrine and TTPs.

Third, the author of the article, per the usual for The Economist, does not have a byline, but I suspect her/his/its beat is probably not even military affairs. (Does The Economist even have such a person?) I would imagine the author's beat is probably social science, academic research, science writ large, or even computer science, etc.

Fourth, what strikes me, first, is the extremely wide range covered by the article. I think a more accurate description of the latter half of the article's content is the use of social media to monitor and predict. Moreover, I would contend most of the phenomena predicted is not necessarily *war* per se, but rather "political instability" which may or may not be a precursor to war. Really, arguably the only reference to modeling "small wars" is the reference to SCARE. (I am a bit surprised the Human Terrain System did not make the article.)

Fifth, Jay Ulfelder at "Dart-Throwing Chimp" worked on the Political Instability Task Force (another unfortunate omission, I would perhaps think), and while I don't read it regularly, I think his blog is fairly pertinent to much of what is covered in the article. Similarly, I would perhaps be at fault for not mentioning "The Monkey Cage": while it is a general political science blog, it is a very good one, and at least one if not more of its contributors could probably provide intelligent commentary.

Sixth, in terms of net assessment with respect to convenional warfare, I have passing familiarity with the work of Mearsheimer and Posen in International Security in the early-to-mid 1980s, and consider it pretty solid.


The article may have value, but it is premised on the seemingly never-ending notion of irregular warfare as the graduate level of war, and that conventional warfare is somehow easier. In this article the author says that conventional war is easy to figure out...

"Guerrilla warfare, however, is harder to model than open battle of this sort, and the civil insurrection that often precedes it is harder still."

I dont buy it. Shoot, Bradley and his subordinates thought that they had Omaha Beach pretty neatly modeled yet they got it wrong, very wrong, and the troops that hit the beach paid a heavy price in their mistakes. Was Bradley et al just stupid because they could not figure out a simple problem of open battle on Omaha Beach. No, maybe in fact, open battle in whatever form it takes in the future is in fact quite complex, perhaps even more so than the experts of coin and irregular warfare have come to believe.

I propose that the author of this piece, just as many so-called experts on coin and irregular war may indeed know a good deal about those things, but they know squat about what terms like "open battle" actually mean, nor do they know squat about military history and this history of warfare in general.

I have to agree with Gian. The graduate "level" of war is ..well,... war . - the entire spectrum of war takes the highest level of intellectual as well as physical abilities. I wonder why Napolean said the moral is to the physical as three is one? If one thinks that large scale maneuver warfare is simple then I would suggest studying a little more history. We really do need to get past these "rice bowls of war" and think about the strategy needed to support our national security objectives regardless of where we might find ourselves on the spectrum of conflict. But as the old saying (adapted) goes: "war is hard, but it is even harder when you are stupid."


This is just a question that may be beside the point, but I read once the the European planners didn't listen as closely to the guys from the Pacific as they might have and shorted Normandy on naval gunfire support, especially destroyers close in. Is that true?