Mad Science Monday, 8/31/2009

I live in Austin, TX. Not only do we have confirmed record-setting zombie hordes, but we also have a populace ready to warn one another of outbreaks of zombiism. So of course I'm interested in knowing whether we (or anyone else) will be able to survive an actual bout with zombies. Thankfully, a group of certified mad scientists have figured this out for us.

Mad Reference: Philip Munz, Ioan Hudea, Joe Imad, and Robert J. Smith?. (2009) "When Zombies Attack!: Mathematical Modelling of an Outbreak of Zombie Infection." Infectious Disease Modelling Research Progress. (full text available free online [PDF])

Mad Background: First off, that isn't a typo in the name of the lead researcher. His last name is "Smith?" with a question mark. In addition, from his homepage at the University of Ottawa (emphasis added), "People kept asking if I'll be getting US citizenship next and I kept laughing at that. Somewhat hysterically, it must be said." Your last name is "Smith?", and you talk about your hysterical laughter on your homepage? You are an inspiration for all would-be mad scientists, Dr. Smith?!

As far as the science, the background you need is that mathematical models are used in fields like epidemiology to help predict the spread of diseases under various conditions, and thus to plan out the best way to combat those diseases. For example, mathematical models can help predict what will happen if only a limited number of vaccine doses are available for a disease, or what will happen if people infected with a disease are quarantined. But can they predict the outcome of a hypothetical disease that follows a pattern very different from known real-world diseases?

Mad Observations: Especially in modern movies and video games, zombiism spreads like a disease. If it spreads like a disease, it should be possible to model it the same way we model diseases.

Mad Hypothesis: If a zombie outbreak occurs, mankind can survive. At least, that's what they're pretending to test. What they're really testing is whether mathematical models can be put together for a "disease" as strange as zombiism, in particular the strain of zombiism in which the dead can become "infected" with the disease and come back to terrorize the living.

Mad Experiment: The researchers built five mathematical models for zombie outbreaks: a basic model, a model with an incubation period, a model in which the unaffected attempt to quarantine the infected, a model in which a treatment for zombiism is available, and a model in which humanity fights back. They then used each model to predict the equilibrium; in other words, to predict whether humanity would survive. Each model had some assumptions in common:

• The particular form of zombiism being modeled is the "slow zombie" style. "Fast zombies," like the things in 28 Days Later, were not studied. I'd be interested to see what would change in such a model, but, alas, that will require further research.
• As I mentioned in the Mad Hypothesis section, the strain of zombiism being modeled also infects the dead (including dead zombies), allowing the dead to join the population of zombies. The whole point was to model something far from known diseases to see how the models held up, so it made sense to include the truly undead in the model.

They All Laughed, But: We are all screwed. Unless we get infrastructure in place to quarantine zombies and zombies-to-be, or are able to quickly develop a cure when an outbreak occurs, or are able to successfully coordinate zombie-eradication attacks, zombies eventually wipe us all out. The eradication model was the only one in which we eventually won, and it seems likely to me that this would require military involvement. If you've ever seen a zombie movie, you know that involving the military is a terrible, terrible idea, so our best hope is also the one that, the "literature" shows us, is empirically shown to lead to a society in which the living envy the dead.

The treatment model was also unique in that, at equilibrium, a large zombie population survived in addition to a small human population. Note that this human population would remain at a certain size, but would not always contain the same individuals; you might become a zombie for a while, then get treated, then die, then rise as a zombie, then get treated again and rejoin the human population. This wouldn't necessarily be a fun existence, although it would definitely be interesting. This model is the only one in which pet zombies, like in Fido and Shaun of the Dead, are even slightly possible. And it looks far more likely that zombies would have pet humans (for a few minutes, before eating their brains and/or infecting them).

The quarantine model seems like our best bet, but, alas, assuming we don't have a massive infrastructure already in place for such quarantine, even then zombies eventually kill us all off.

Mad Engineering Applications: As it turns out, Dr. Smith?'s page indicates that a mad engineer has already expressed interest in this research, in that someone wrote to Dr. Smith? asking for help engineering a zombiism virus. Presumably that evil genius plans to control the treatment of his strain of zombiism, thus ensuring that he is (at least occasionally) a member of the small surviving human population. In case that nutjob is able to design such a virus, I guess the rest of us need to be ready to work with the military to make sure his plan isn't successful.

Of course, the other point of all of this was that the models seemed to work. The real application is to, essentially, not be afraid to try to model things that don't follow traditional disease models. The paper mentions the examples of allegiance to political parties or diseases with dormant infection, but there are definitely other things that can be modeled much like diseases.

It's hard to pick a favorite part of all of this, but, if you get a chance, I strongly recommend at least reading the last two pages of the PDF (the references). I couldn't stop laughing (maniacally, of course), seeing things like "Capcom, Shinji Mikami (creator), 1996-2007 Resident Evil" listed alongside "van den Driessche, P., Watmough, J. (2002) Reproduction numbers and sub-threshold endemic equilibria for compartmental models of disease transmission. Math. Biosci. 180, 29-48."

If you'd like to continue to study zombie survival tips, I recommend my friend Jon's weekly Zombie Friday! You can probably guess on which day you should check his site for said column, unless, of course, you're already safe from zombies.