THE OPTIMAL SIZE OF A TERRORIST NETWORK

Distributed, dynamic terrorist networks cannot scale like hierarchical networks. The same network design that makes them resiliant against attack puts absolute limits on their size. If so, what are those limits?

A good starting point is to look at limits to group size within peaceful online communities on which we have extensive data -- terrorist networks are essentially geographically dispersed online communities. Chris Allen does a good job analyzing optimal group size with his critique of the Dunbar number.

His analysis (replete with examples) shows that there is a gradual fall-off in effectiveness at 80 members, with an absolute fall-off at 150 members. The initial fall-off occurs, according to Chris, due to an increasing amount of effort spent on "grooming" the group to maintain cohesion. The absolute fall-off at 150 members occurs when grooming fails to stem dissatisfaction and dissension, which causes the group to cleave apart into smaller subgroups (that may remain affiliated).

Al Qaeda may have been able to grow much larger than this when it ran physical training camps in Afghanistan. Physical proximity allowed al Qaeda to operate as a hierarchy along military lines, complete with middle management (or at least a mix of a hierarchy in Afghanistan and a distributed network outside of Afghanistan). Once those camps were broken apart, the factors listed above were likely to have caused the fragmentation we see today (lots of references to this in the news).

This leads us to optimal group size, which according to Chris Allen's online group analysis, can be seen at two levels: both small and medium sized. Small, viable (in that they can be effective at tasks) groups (or cells) are optimized at 7-8 members. A lower boundary can be seen at 5 (with groups less than 5 not having sufficient resources to be effective) and an upper boundary at 9. Medium sized groups are optimal at 45-50 members, with a lower limit of 25 and an upper limit of 80. Between these levels is a chasm that must be surmounted with significant peril to the group. This is due to the need for groups above 9-10 members to have some level of specialization by function. This specialization requires too much management oversight to be effective given the limited number of participants in each function. At 25 members, the group gains positive returns on specialization given the management effort applied (a break-even point).

This chasm (between 9-25 members) nicely matches the problem period in the development of terrorist and guerrilla networks that studies of guerrilla groups refer to. The amount of damage a small (7-8 member) group can do is limited to narrow geographies and therefore does not represent a major threat. Once a network grows to 45-50 members, they can mount large attacks across multiple geographies. They are also very difficult to eliminate due to geographically dispersion of cells. However, during the transition to a larger group they are vulnerable to disruption. This vulnerability necessitates fast counter-terrorist action (this gives credibility to the military strategists who claim we didn't have enough troops in Iraq immediately after the war, nor were we quick enough to establish martial law) during that short period of time a network is transitioning in size.

This size dynamic can also be seen in criminal organizations. The mafia (BBC), despite their widespread influence, has closely mirrored the limits on group size:

• The Genoveses are the largest of the five families in New York and they recruited nine new foot soldiers, bring their total to 152.
• The Gambinos, had a terrible year from 2000-2001, losing 33 members, but they still managed to retain 130, making them the second largest in terms of manpower.
• Meanwhile the Luccheses have initiated three more gangsters, lifting them to third place with a total of 113 hoods on the streets, according to FBI reports.

A recent Washington Post article on Islamic terrorist cells in Iraq says:

Dempsey said he estimated there were only about 100 "foreign terrorists" in Baghdad, organized into about six cells. In Anbar province, which stretches across western Iraq and includes the strife-torn cities of Ramadi and Fallujah, Maj. Gen. Charles H. Swannack Jr. of the 82nd Airborne Division said he believed there were a total of 50 to 80 foreign fighters in eight to 10 cells.

This indicates a cell size (the optimal size of the smallest viable network) of between 5-12 members.

Note: The limits on organizational size does not mean that terrorist or crime organizations can't expand their ranks on a temporary basis. There are plenty of "contract" employees available. Also, there is also the potential for intergroup cooperation (we see this in both crime and terrorism).

Comments

[Gil Friend]

How is this analysis applicable to other (non-terrorist, non-criminal) organizations? Like companies, NGOs, government agencies?

Have you seen Thomas Malone's new book "The Future of Work?" He posits (this from dust cover) that "technological and economic factors - particularly the rapidly falling cost of communication - is enabling a change in business organizations as profound as the shift to democracy in governments."

[Christopher Allen]

Very interesting analysis -- I appreciate the source of the info on Mafia sizes. Can you give me a source for "matches the problem period...lots of the guerrilla warfare literature refers to"? I'd like to use it in a followup to my Dunbar article.

[jeff]

It was for me great pleasure to visit and enjoy this site. Thanks

[Valdis K]

What happens in truly "social" networks and "covert" networks is NOT the same! Covert networks do not demand/need the constant grooming of social networks.

Also networks can 'fold' for task related communications and then go back to a more covert mode to avoid detection. Compare Figures 2 and 3 in this article:
http://www.firstmonday.org//issues/issue7_4/krebs/

Small clusters can connect to form a larger small-world network... again the activity of the connections may be determined by task requrements. Several 8 person networks connect to form a 24 person network for a particular task.

The key for the gloabl guerillas is adaptation via recombination.

[John Robb]

Amazing insight! Thanks Valdis.

[John Hawks]

I am very sympathetic to this point of view about network size, but relying on Dunbar's ecological model imposes some limits on its applicability. Remember that Dunbar's case is that humans have language in order to allow social groups of arbitrarily large size, because the groups can be structured in ways that do not require all interpersonal relationships to be binary.

For a limit to apply to a human group, there needs to be some reason to reintroduce a time constraint on maintaining relationships. For terrorist networks, this is not hard to defend, because network members must be trusted, which means that their activity must be well-known to other groups members. Thus, monitoring network members must take a significant portion of somebody's time, and if there is no centralized control, this duty must be shared by many network members. But the level of trust necessary for successful (i.e. non-infiltrated) operation of the network is not a constant. Instead, it depends on (a) the extent of shared commitment to the terrorist enterprise among network members, and (b) the frequency of attempts by outsiders to infiltrate the network.

Thus, there may be no single optimal size for a terrorist network, since the time spent verifying the bona fides of network members depends on the force and effectiveness of people attempting to infiltrate those networks. Any increase in enforcement capabilities might be expected to decrease terrorist effectiveness, particularly if network size or connectivity could be dropped below some critical threshold.

[John Robb]

Thanks John. I think that it is important to point out that this is a geographically dispersed virtual community (that uses the Internet to communicate -- cell phones have been abandoned). As such, the lessons about online community we've learned over the last 9 years apply. We've found that the Dunbar number applies to online community. Additionally, certain forms of organization, that allow groups to scale beyond small groups, aren't possible given the pressures they are under (particularly since we have shut down the camps).

Unfortunately, there is a work around that answers the question: how do relatively small autonomous networks operate in concert to conduct war? It's called the bazaar (I have a brief on this on this site, and a longer explanation in my book). It is formed via a combination of mechanisms we see at work in the open source world (to build extremely complex software) and stigmergic learning. The bazaar explains how the guerrilla movement in Iraq works. This model will probably serve as the basis of the next generation terrorism.

[John Hawks]

OK, I get what you are saying concerning the bazaar model. Essentially, the cost of information transfer among groups is assumed to be very low (through mass media or internet-anonymous communication), so that a very large number of groups may maximize the effectiveness of attacks without direct coordination (or even knowledge of each other). One may add that this is an especially potent form of assault upon Western societies, because an obvious strategy is for potential targets of terrorist action to spread disinformation to disrupt this communication, but a free society will not tolerate much disinformation.

My own interest in this subject is the application of such priniciples to group dynamics during the evolution of humans, which presents a slightly different spin. In these contexts, all group sizes were small and exogamous, meaning that people had relatives in neighboring groups, to a greater or lesser extent. In these contexts, a bazaar model also has its application, since new foods, medicines, hunting techniques, and technologies could spread fairly quickly once they were developed by somebody.

But in this context, it becomes a challenge to explain why large-scale differences among groups emerged. Were these the result of limits on information transfer, antagonism between groups, or ecological barriers to various kinds of knowledge acquisition? We know that these large-scale cultural differences did arise (they still persist) and that they did not begin to have great importance until the past 50,000 years or so. The recency of their evolution implies that some constraint appeared that had previously been absent.

So for me, finding the constraints on a bazaar model is a very interesting subject.

[Watching Them, Watching Us]

"One may add that this is an especially potent form of assault upon Western societies, because an obvious strategy is for potential targets of terrorist action to spread disinformation to disrupt this communication, but a free society will not tolerate much disinformation."

Are you really sure that a Western "free societeies" "will not tolerate much disinformation" ?

It seems to me that a vast amount of propaganda ranging from tabloid celebrity sleaze stories through to actual political spin, media manipulation and classic "cold war era" style "disinformation" and "black propaganda" are tolerated every day. The rise of the Internet has vastly affected the signal to noise ratio for the worse, despite spreading the nuggets of "good" information farther than was possible before.

[ep]

Interesting analysis but strikes me as sorely lacking in context.

Seen The Power Of Nightmares?

Seen this?
http://tinyurl.com/dco2z
"apparently Aswat was working for British intelligence"

Or the list of links at the bottom of this post?
http://tinyurl.com/ar7qj

Are you sure you're not just chasing a phantom?

[enigma_foundry]

"This chasm (between 9-25 members) nicely matches the problem period in the development of terrorist and guerrilla networks that studies of guerrilla groups refer to. The amount of damage a small (7-8 member) group can do is limited to narrow geographies and therefore does not represent a major threat."

No. To create a group of 6 that could kill millions with a properly executed biological attack is not out of the question. The agent that would be created would have to be relatively lethal, and preferably slow to act. There is no reason to think such an agent would be impossible to create, or that this technology is out of reach of terrorists. An educated micro-bioologist could find all the necessary equipment on ebay.

The infrastructure needed to defend against such an attack is not in place, and the correct execuation, against those who are largely not insured in American urban centers could play on both class and racial divisions quite well.

The fact that the present regime in America will not consider universal health care as necessary line of defense just gives those who would attack the USA a very clear road map to victory.

There is every reason to believe that someone with malevolent intentions has built a bio-warefare production facility on US soil (remeber the anthrax attacks, co-ordinated in time with 9/11?) and will us it when the time is optimal. Probably near an election.

[enigma_foundry]

" 'One may add that this is an especially potent form of assault upon Western societies, because an obvious strategy is for potential targets of terrorist action to spread disinformation to disrupt this communication, but a free society will not tolerate much disinformation.'

Are you really sure that a Western "free societeies" "will not tolerate much disinformation" ? ..."

I am certain they will begin to create more disinformation, when faced with a real attack those in power will initially deny the scale of the attack, and thereby lose all credibility.

Those who will attack us have considered our leaders responses, which will fully expose our present leaders moral duplicity.

Our leaders will be just as much our enemies as al-Qaida is, and we will, I predict, see our leadeers killed by angry mobs.

It will certainly be a sad day, but one that our leaders in their utter selfishness will have brought on themselves.

[freeDissent]

This interesting post on the optimal size of a guerrilla cell reminds me of the fictional story, "The Moon is a Harsh Mistress" by Robert A. Heinlein and his portrayal of a revolution comprised of a unique hybridization of people and technology. Heinlein described a system to organize cells of three where any person in the network only knows three other people, but is still able to coordinate the entire network through the use of a sophisticated technology system. Heinlein wrote fiction, but we are seeing something very similar to this system in Iraq.
http://en.wikipedia.org/wiki/The_Moon_Is_a_Harsh_Mistress