The Calculus Diaries - Jennifer Ouellette [65]
In one study, Eagle and his colleagues provided cell phones to ninety-four test subjects—all students or faculty at MIT—loaded with special software that kept track of their location and logged all calls made and received among those phones. As a control, the study also included self-reportage, with subjects identifying which of the other subjects were friends, acquaintances, or strangers. Based on the calling patterns, the researchers were able to identify correctly whether two given subjects were friends or strangers more than 95 percent of the time.
That is just one study. Based on three basic parameters—a user’s activity, location, and proximity to other users—Eagle says it is possible to accurately predict someone’s future behavior based on limited observation of their current behavior. And because transmission of disease is strongly correlated to social networks and proximity to others, his method is extremely useful for epidemiological modeling.
According to Eagle, the typical epidemiological model rests on an erroneous assumption: that the probability of infection is equal for all, that is, the population is well mixed. But social networks are much more complex than that; there is noticeable clustering of social contacts, and people in those clusters would have a higher probability of becoming infected. It makes a strong case for being a hermit. The Bennett family fears social shunning when their younger daughter, Lydia, scandalously elopes with Wickham; had this transpired, they could have taken comfort in the fact that they would have been far less likely to be bitten by zombies than their more socially active neighbors.
Eagle believes that the data captured by his cell-phone software application gives a much more realistic picture of the dynamics of human social networks,42 thereby arming epidemiologists with “more information to make predictions about our vulnerability to the next SARS, as well as greater insight into preventing future epidemics.” One can only wonder what further insights might be gleaned if Eagle outfitted zombies with cell phones.
How human beings react to these kinds of threats is another factor that can be tough to calculate. That’s why other researchers are looking to online virtual worlds for modeling the spread of infectious disease, much like the collapse of a virtual bank in Second Life might shed light on economic models. Most epidemiological models use mathematical rules to approximate human behavior, but the modelers must make certain assumptions about how humans are likely to behave— and those assumptions can be inaccurate. Deliberately introducing a deadly pathogen into a controlled population to study the outcome would be immoral, but what if it were possible to design a “disease” specifically for a virtual online community?
Game designers were a little ahead of the scientists on that front. Blizzard Entertainment, the makers of World of Warcraft—a hugely popular multiplayer game—deliberately introduced a zombie plague into the game to promote World of Warcraft: Wrath of the Lich King. But a far more interesting development was the virtual Corrupted Blood epidemic that broke out in 2005. Blizzard added a new dungeon called Zul’Gurub, controlled by an “end boss” named Hakkar. Only highly advanced players could find Zul’Gurub, where the objective was to kill the end boss. Among the creature’s weapons was a spell called Corrupted Blood, which inflicted damage on infected players at regular, repeating intervals, slowly draining away their vitality until their avatars “died.” Killing Hakkar was the only cure.
The spell was designed to infect only nearby players, and to remain confined to the Zul’Gurub game space.