The Dunbar Number
Last week, I posted a link to an essay titled Monkey Law, by David Wong. The essay makes reference to ‘The Dunbar Number’, established by the research of R.I.M. Dunbar on the relation of neocortical volume to social group size.
The Monkey Sphere has a tongue-in-cheek description of Dunbar's discovery:
Also known as ‘the Law of 150’, this is of interest to virtual world developers in establishing the ‘proper’ size for social groups in online games. Typically it is applied ham-fisted as a sign-post of optimum guild sizes. The logic is something like, “Someone smart has established that the optimum social group size is 150, therefore we should make our online game support guilds of about 150 people, expect guilds to have 150 members on average, balance content and so forth with the number 150 in mind. 150, 150, 150!”
Or developers may fall back to the Laws of Online Game Design and read:
…and apply the same logic process (“Well guilds are communities, and the ideal size is… uh… from 0 to 250… so… uh…”) and conclude that they should implement guild systems post-launch. They are, after all, pressed for time.
So I started thinking: This is antithetical to everything I have learned, seen, heard and smelled in my many, many years of playing online games, and in my fist-full of years making them. Guilds, being the most common manifestation of online communities (in MMORPGs, I mean), just don’t typically get that big. When they do, they are frequently, in practice, multiple subcommunities working together to overcome a common enemy (such as, say, obstacles designed to require massive armies, because the developers expected most guilds to be that big).
I would expect the most common guild size to be the ‘optimum’, which would imply that the optimum guild size is a lot lower than 150.
My blog, my observations:
I know, this approach is contrary to the “Anthropologists say optimum community size is 150, and therefore we should assume guilds to be ~150 members strong even if they aren’t”, but I’m crazy that way.
It doesn’t fit my own personal observations, but I could be wrong. So I did some googling, and ran across another skeptic, the esteemed Christopher Allen.
In his blog entry titled The Dunbar Number as a Limit to Group Sizes, he has some nice charts of guild sizes in Ultima Online, Allegiance Size numbers from Asheron’s Call, and personal observations of other member-size averages (and apparent limits) in assorted other online groups. The charts are from Raph Koster's Small Worlds presentation.
He speculates:
Sounds good, except language was supposed to be what provided humans with the more efficient grooming (so to speak) that a social group of 150 requires, not the other way around (the limit isn't 150 because that's all we have time for, but rather the limit is 150 because of our brains, and that requires a more efficient mechanism than grooming to maintain).
The reason optimal size for 'creative and technical groups' is smaller than the size for, say, a village, is not due to their function/purpose. It is because the village is an all-encompassing group. It is everyone you know.
Meaning:
There’s nothing in Dunbar’s research to suggest that the limit to human group size is 150 (or 250, or any other number) due to time-limitations. Or to put it another way, even if we had a mechanism for maintaining social cohesion that was even more efficient than language – telepathy, let’s say – then we would still be limited to social groups of about 150 people, because that is all our brains can support.
Or for an even third way of putting it, this time in the form of a question: What is the maximum group size humans could maintain if the limit were based on the efficiency of language as a tool for maintaining social cohesion? Answer: We don’t know!
So, I got lucky. Chistopher Allen’s going somewhere else with all this, and didn’t hit on the point I want to make. So now I can still make my point, and I’ll just refer to Raph's charts and Chris' observations to validate my suspicion that ‘typical’ guild sizes are much, much smaller than either our language or our brains can support.
You can take a closer look at just what exactly R.I.M. Dunbar actually wrote, but I’ll give you the stupid-person summary:
This point is mistaken for being useful to the virtual world developer. We want to know the answer to the question, “Well what size is that?” The answer depends on which formula you prefer. Dunbar’s equation isn’t simply a measure of neocortex size, volume or surface area, but rather… well, knock yourself out:
That gives us ‘The Law of 150’, Dunbar’s number, really 147.8.
A different equation (based on absolute neocortex volume) gives us 248.6, which might be the origin of the 250 in the Laws of Online Game design.
In order to understand why online communities are smaller than the Dunbar Number, we must keep in mind that the Dunbar Number is not a property of the group, but rather it is an attribute of the individuals in the group. Groups don’t have neocortical volumes, people do.
We, human beings, cannot maintain more than 150 social relationships.
Guilds in online games are smaller than this because their members have other important social relationships that they must maintain. Don’t you have any friends or family offline? No individual member of the group can maintain more than 150 social relationships, total.
So let’s say, offline, you have a 100 social relationships. You could go online, join a guild of 50 people, and maintain all of your social relationships, both online and off. The optimal guild size for you is 50. If your guild is larger than that, then it ceases to be ‘your online community’.
More likely, you will have 150 offline social relationships, and as you invest yourself in the online community you begin to sever ties with your old offline friends. Ever heard of that happening? ‘Course.
Different people have needs (or perhaps I should say, “available capacities”) to form different numbers of relationships online, on a case-by-case, person-by-person basis. This is determined by how many relationships they are maintaining in other groups, including (especially) offline ones.
You, being so popular with your rugged good-looks and smooth-talking ways, might be maintaining 145 social relations, straining your little brain to the very limit just to maintain those. And then you join an online game. You could form some new relationships with a tight knit little crew of six people. If all six of your posse are in the same situation, then the ‘optimal group size’ for this hypothetical happy band is isn’t 250 or even 150. It’s six! Even if they aren’t all in the same situation, since you are, then the optimal group size is still six. If those other members want more relationships, then they need to join other groups (of various sizes, up to their limit).
This may even dictate what sorts of games a person wants to join. If the game is especially geared toward very large groups then it might even be prohibitive for the average person to form the sorts of meaningful relationships which lead to cohesive online communities (which lead to greater retention).
Again, not due to time-limitations, but because our brains cannot handle maintaining more than 150 total relationships simultaneously.
My conclusion to all this leads to requirements for MMORPGs to support multi-guild, micro-guild and persistent-partner relationships. To stop thinking of guilds as 'online communities' and to instead recognize that online communities exist within subsets of even relatively small guilds, communities which should be supported and encouraged to form via system mechanics (similar to the system mechanics which support guilds, only different).
But I’m going to leave that essay for another day.
Update: Previously I had incorrectly attributed a quote to the PWoT article. I dont know where it came from, but it didn't come from there, so. Fixed now.
The Monkey Sphere has a tongue-in-cheek description of Dunbar's discovery:
You see, monkey experts performed a monkey study a while back and discovered that the size of the monkey's monkey brain determined the size of the monkey groups the monkeys formed. The bigger the brain, the bigger the little societies they built.
They cut up so many monkey brains, in fact, that they found they could actually take a brain they had never seen before and with a simple dissection, analysis and a quick taste, they could accurately predict what size tribes that species of creature formed.
Most monkeys operate in troupes of 50 or so. But somebody slipped them a slightly larger monkey brain -- but a monkey brain nonetheless -- and they estimated the ideal group or society for this particular animal was about 150.
That brain, of course, was human. Probably from a homeless man they snatched off the streets.
Also known as ‘the Law of 150’, this is of interest to virtual world developers in establishing the ‘proper’ size for social groups in online games. Typically it is applied ham-fisted as a sign-post of optimum guild sizes. The logic is something like, “Someone smart has established that the optimum social group size is 150, therefore we should make our online game support guilds of about 150 people, expect guilds to have 150 members on average, balance content and so forth with the number 150 in mind. 150, 150, 150!”
Or developers may fall back to the Laws of Online Game Design and read:
Community size
Ideal community size is no larger than 250. Past that, you really get subcommunities.
…and apply the same logic process (“Well guilds are communities, and the ideal size is… uh… from 0 to 250… so… uh…”) and conclude that they should implement guild systems post-launch. They are, after all, pressed for time.
So I started thinking: This is antithetical to everything I have learned, seen, heard and smelled in my many, many years of playing online games, and in my fist-full of years making them. Guilds, being the most common manifestation of online communities (in MMORPGs, I mean), just don’t typically get that big. When they do, they are frequently, in practice, multiple subcommunities working together to overcome a common enemy (such as, say, obstacles designed to require massive armies, because the developers expected most guilds to be that big).
I would expect the most common guild size to be the ‘optimum’, which would imply that the optimum guild size is a lot lower than 150.
My blog, my observations:
- The typical guild size is 30 to 50 people.
- Many guilds are even smaller than that.
- Guilds which are larger than that are frequently fragmented, internally, into smaller groups.
I know, this approach is contrary to the “Anthropologists say optimum community size is 150, and therefore we should assume guilds to be ~150 members strong even if they aren’t”, but I’m crazy that way.
It doesn’t fit my own personal observations, but I could be wrong. So I did some googling, and ran across another skeptic, the esteemed Christopher Allen.
In his blog entry titled The Dunbar Number as a Limit to Group Sizes, he has some nice charts of guild sizes in Ultima Online, Allegiance Size numbers from Asheron’s Call, and personal observations of other member-size averages (and apparent limits) in assorted other online groups. The charts are from Raph Koster's Small Worlds presentation.
He speculates:
This all leads me to hypothesize that the optimal size for active group members for creative and technical groups -- as opposed to exclusively survival-oriented groups, such as villages -- hovers somewhere between 25-80, but is best around 45-50. Anything more than this and the group has to spend too much time "grooming" to keep group cohesion…
Sounds good, except language was supposed to be what provided humans with the more efficient grooming (so to speak) that a social group of 150 requires, not the other way around (the limit isn't 150 because that's all we have time for, but rather the limit is 150 because of our brains, and that requires a more efficient mechanism than grooming to maintain).
The reason optimal size for 'creative and technical groups' is smaller than the size for, say, a village, is not due to their function/purpose. It is because the village is an all-encompassing group. It is everyone you know.
Meaning:
- Non-human primates use grooming to maintain social cohesion
- Grooming is too inefficient with group sizes of 150
- So humans use language, instead of grooming, to maintain social cohesion
There’s nothing in Dunbar’s research to suggest that the limit to human group size is 150 (or 250, or any other number) due to time-limitations. Or to put it another way, even if we had a mechanism for maintaining social cohesion that was even more efficient than language – telepathy, let’s say – then we would still be limited to social groups of about 150 people, because that is all our brains can support.
Or for an even third way of putting it, this time in the form of a question: What is the maximum group size humans could maintain if the limit were based on the efficiency of language as a tool for maintaining social cohesion? Answer: We don’t know!
So, I got lucky. Chistopher Allen’s going somewhere else with all this, and didn’t hit on the point I want to make. So now I can still make my point, and I’ll just refer to Raph's charts and Chris' observations to validate my suspicion that ‘typical’ guild sizes are much, much smaller than either our language or our brains can support.
You can take a closer look at just what exactly R.I.M. Dunbar actually wrote, but I’ll give you the stupid-person summary:
- Group size is a function of relative neocortical volume.
This point is mistaken for being useful to the virtual world developer. We want to know the answer to the question, “Well what size is that?” The answer depends on which formula you prefer. Dunbar’s equation isn’t simply a measure of neocortex size, volume or surface area, but rather… well, knock yourself out:
log(N) = 0.093 + 3.389 log(CR) (1) (r2=0.764, t34=10.35, p<0.001), where N is the mean group size and CR is the ratio of neocortex volume to the volume of the rest of the brain (i.e. total brain volume minus neocortex) (Dunbar 1992a). Use of both major axis and least-squares regression, as well as alternative indices of relative neocortex size, all yield equations that are of about this same magnitude.
That gives us ‘The Law of 150’, Dunbar’s number, really 147.8.
A different equation (based on absolute neocortex volume) gives us 248.6, which might be the origin of the 250 in the Laws of Online Game design.
In order to understand why online communities are smaller than the Dunbar Number, we must keep in mind that the Dunbar Number is not a property of the group, but rather it is an attribute of the individuals in the group. Groups don’t have neocortical volumes, people do.
The group size identified by this relationship appears to refer to the maximum number of individuals with whom an animal can maintain social relationships by personal contact. It is not necessary that all these individuals live in the same physical group: chimpanzees (among a number of other species) have a fission/fusion form of social system in which at any one time the community (the group in the sense defined above) is divided into a number temporary foraging parties whose composition changes repeatedly (see for example Wrangham 1986). Nor does it follow that a species' social system consists only of a single type of group: it is now clear that most primate species live in complex multi-tiered social systems in which different layers are functional responses to different environmental problems (e.g. the gelada and hamadryas baboons: see Dunbar 1988, 1989a). Rather, the neocortical constraint seems to be on the number of relationships that an animal can keep track of in a complex, continuously changing social world: the function subserved by that level of grouping will depend on the individual species' ecological and social context.
We, human beings, cannot maintain more than 150 social relationships.
Guilds in online games are smaller than this because their members have other important social relationships that they must maintain. Don’t you have any friends or family offline? No individual member of the group can maintain more than 150 social relationships, total.
So let’s say, offline, you have a 100 social relationships. You could go online, join a guild of 50 people, and maintain all of your social relationships, both online and off. The optimal guild size for you is 50. If your guild is larger than that, then it ceases to be ‘your online community’.
More likely, you will have 150 offline social relationships, and as you invest yourself in the online community you begin to sever ties with your old offline friends. Ever heard of that happening? ‘Course.
Different people have needs (or perhaps I should say, “available capacities”) to form different numbers of relationships online, on a case-by-case, person-by-person basis. This is determined by how many relationships they are maintaining in other groups, including (especially) offline ones.
You, being so popular with your rugged good-looks and smooth-talking ways, might be maintaining 145 social relations, straining your little brain to the very limit just to maintain those. And then you join an online game. You could form some new relationships with a tight knit little crew of six people. If all six of your posse are in the same situation, then the ‘optimal group size’ for this hypothetical happy band is isn’t 250 or even 150. It’s six! Even if they aren’t all in the same situation, since you are, then the optimal group size is still six. If those other members want more relationships, then they need to join other groups (of various sizes, up to their limit).
This may even dictate what sorts of games a person wants to join. If the game is especially geared toward very large groups then it might even be prohibitive for the average person to form the sorts of meaningful relationships which lead to cohesive online communities (which lead to greater retention).
Again, not due to time-limitations, but because our brains cannot handle maintaining more than 150 total relationships simultaneously.
My conclusion to all this leads to requirements for MMORPGs to support multi-guild, micro-guild and persistent-partner relationships. To stop thinking of guilds as 'online communities' and to instead recognize that online communities exist within subsets of even relatively small guilds, communities which should be supported and encouraged to form via system mechanics (similar to the system mechanics which support guilds, only different).
But I’m going to leave that essay for another day.
Update: Previously I had incorrectly attributed a quote to the PWoT article. I dont know where it came from, but it didn't come from there, so. Fixed now.
