Pokémon Go Is All the Rage, But Where's the Science?

Image: Ivana Li, Shelomi et al/AIR
Image: Ivana Li, Shelomi et al/AIR

The Pokémon Go craze is still going strong, as players scamper about the great outdoors trying desperately to catch ‘em all. And amid all the accompanying media frenzy, there’s one burning question on everyone’s mind: where’s the science?


Yes, there is a science of Pokémon—specifically, a 2012 satirical study that appeared in the humorous journal, Annals of Improbable Research entitled, “A Phylogeny and Evolutionary History of the Pokémon.” The authors—then a group of budding young entomologists at the University of Caifornia Davis—traced the evolutionary history of the 646 fictional species of “pocket monsters,” even creating an impressive 16-generation phylogenetic or evolutionary tree.

Why do this at all? Well, the paper itself cites the need to document all the species because they are “threatened by the Pokémon fighting rings that are growing rapidly in popularity, particularly among urban youth.” But really, it was just a fun side project.

“I had a lull in my dissertation research and decided to spend the weekends and downtime making this phylogeny,” lead author Matan Shelomi said at the time. “It took at least a month to actually collect all the data, which I did manually by scrolling through Pokémon websites.”

Considering the creatures are entirely made up, the tree turned out surprisingly well, although a fierce debate over possible errors raged on Reddit when the tree was first published. It was also a fitting tribute to Pokémon’s creator: Japanese video game developer Satoshi Tajiri, who collected insects in his childhood and once considered becoming an entomologist.

You can see the full-scale, high resolution image of the evolutionary tree here.

Illustration for article titled Pokémon Go Is All the Rage, But Where's the Science?

[Annals of Improbable Research]


I was interested in seeing how this tree was assembled because I assumed there wasn’t any molecular data for fictional creatures. So thank you for supplying the paper.

our cladistic analysis used the following shared characters: Type (e.g. Fire, Water, Bug, Ghost, Fighting, Steel, etc.), Egg group (16 categories limiting which hybridizations are possible), Body Style (14 categories describing general morphology), and moves and abilities. Over 700 moves and abilities are known, and whether a given taxon was capable of learning a move via natural development was the main synapomorphy (shared, derived characteristic) used in our phylogeny.

So all of these were being treated as being analogous to morphological characters.

The software MrBayes was used to run a Bayesian MCMC (Markov chain Monte Carlo) analysis of phylogeny.

I’m not entirely sure why a Bayesian approach was used here; it’s not a bad idea but since there’s no molecular data included a Bayesian analysis wouldn’t necessarily be better than a parsimony analysis.

The Grass-type Pokémon, which spurred furious debate over whether they should be considered Plants or Animals, are shown to be a monophyletic group that evolved from a clade of Normal-type quadrupeds; the half-plant, half-reptile Venusaur appears as the transitional species between these two stages.

That is a really awkward way of saying that Grass-type Pokémon form a clade that evolved from Normal-types and that Venusaur is the earliest-diverging member of the clade.