Player Types: Watch for Moving Targets
A Deeper Look at Richard Bartle’s Player Types, Part II
Since the Bartle Test of Gamer Psychology was created in 1996, more than 740,000 people have taken it. Gaming circles have seen the old saw “What’s your sign?” transformed into “I’m an KEAS. What’s your type?” For example, I found this comment on a gaming blog:
“So I took the test again today.
Apparently I’m a little bit more into PVP these days and a little bit less in socializing.
Killer 93%, Explorer 60%, Achiever 33%, Socializer 13%”
—Loregy.com
This comment hints at a two interesting subtleties about Bartle’s player type model. First, most players exhibit a combination of all four player types, and second, and just as important, players may change their type from time to time. In fact, as we will see, players will often move through a predictable progression of types over the course of playing any given game.
If you’re not familiar with Richard Bartle’s Player Type model, my last post delved into the definitions of each player type, how players of different types interacted with each other and amongst themselves, and how multi-user games need to achieve a balance between types. The fact that players often exhibit behaviors of all four types provides another reason to avoid designing applications that don’t cater in some way to all of the player types. In this post, we’ll take a deeper look at how Bartle’s full model explains the movement between types.
Expanding Bartle’s Original Model
Bartle’s original model mapped players on a two-dimensional grid with the two axes expressing each player’s degree of preference for acting on or interacting with the game world itself or its players.[1] In his 2005 paper “Virtual Worlds: Why People Play,” Bartle notes that there were several flaws in this model:
“Although this model has been generally accepted as a useful tool among designers, it
nevertheless has flaws. Two are of particular importance. Firstly, it suggests that players
change type over time, but it doesn’t suggest how or why they might do so. Secondly, all
of the types to some degree, but especially the one for acting on players (that is, Killers),
seem to have sub-types that the model doesn’t predict.”[2] (more…)