The principal analysis of Dunbar's Number as it concerns online gaming comes from Christopher Allen, founder of MUD operator Skotos Tech, on his intriguing (if sporadically updated) blog, Life With Alacrity. Starting with a lengthy March 2004 treatise, Allen spent half a dozen posts analyzing the Number's relevance to online gaming:

"I've seen similar limits myself in some of the small online games that Skotos produces. For instance, in Castle Marrach , which is a social-dominant game (i.e. like a MUSH), the game grew quickly until we reached approximately 150-200 active users. However, whenever it grew beyond that number, it always seemed that politics and dissatisfaction would bubble up such that people would drop out, leaving us back close to 150 or 160."

In an August 2005 post, Allen analyzed guild sizes in Ultima Online and World of Warcraft. (MMOG social scientist Nick Yee compiled the data in June 2005 at Allen's request and presented it on the PlayOn blog.) The WoW breakpoint lay close to 50, versus UO's 150, and smaller guilds were far more common in WoW than in UO. "If one-person guilds are excluded, the average guild size was 16.8, the median was 9," Allen wrote. "My guess is that there is something about Worlds [sic] of Warcraft such that even participating in very small groups can be useful, whereas for Ultima Online the utility is gained mainly by sharing the resources earned by larger groups. Thus Worlds of Warcraft has groups that are 'bands' as well as 'tribes,' while with Ultima Online groups are more likely to be just 'tribes.'"

Like Dunbar himself, Allen believes the limit of 150 relationships is practical mainly for groups faced with a strong reason to remain together, such as military units or tribes facing a hardscrabble struggle for survival. In such large groups, each individual member may spend quite a lot of time on "social grooming." Absent this desperate environment, an individual will choose to maintain far fewer close relationships; this "non-survival-oriented" figure, Allen says, "hovers somewhere between 25-80, but is best around 45-50." He also postulated several other breakpoints in group size for practical working teams (five to nine members), small businesses (25-80), and larger businesses with middle management (500+). Above or below these ranges, group efficiency and satisfaction fall sharply. (Allen's analysis gained chilling confirmation from military theorist John Robb. In his blog about "fourth-generation warfare," Global Guerrilas, Robb analyzed the optimal size of an Iraqi terrorist network and a Mafia crime family. Both fit the same breakpoints as an MMOG guild.)

Of course, members of certain professional classes, such as doctors, salespeople and politicians, routinely socialize with hundreds, even thousands of acquaintances. Allen suggests this can work because these professionals practice "Dunbar triage": They spend far more time than average on their relationships, prioritize ruthlessly and meet most of their contacts in highly structured situations. Allen moves into uncharted territory with "Cheers: Belongingness and Para-Social Relationships." He speculates that tracking a character's created life on a TV show is a "junk relationship" that takes up a Dunbar slot. If this unsettling idea is true, it would help explain the social life of, say, a comic book fan, who can instantly recite all the melodramatic events of many hundreds of superheroes but has few friends.


For social network theorists, not to mention guild leaders, this is all interesting. But is there really a Dunbar limit hard-coded into your brain?

Despite the statistics and anecdotal evidence, it's unlikely. Research has yet to determine the human brain's memory capacity. Back in the '70s, memory artist Harry Lorayne, about an hour before his nightly stage act, would go out front and mingle with his audience. He'd learn their names and interesting facts about them, and then during his show he would name each one, every member of an audience of hundreds. (If Lorayne plays World of Warcraft nowadays, he probably knows everyone on his server.) Obviously, just remembering someone's name doesn't mean he's in your monkeysphere. The point is, the human brain doesn't melt down at 150 names.

Comments on