In Chapter 3 of Designing Virtual Worlds, Richard Bartle describes a system he devised to break players of Virtual Worlds in classifications based on what they consider "fun" within the virtual world. Admittedly any such system is prone to exceptions, but in thinking about it I found one telling exception that to me says that Bartle's two-dimensional system doesn't have enough depth to it... the fact that Bartle excludes himself from the categorizations. Bartle explains that because he analyzes virtual worlds he doesn't see them as fun.
First of all, let me admit that I don't have the academic depth that Bartle has, and have not subjected these thoughts to as rigorous approach, as it is only meant as a thought experiment.
What I believe the primary lack in Bartle's system is is that it doesn't take into account all of the depth of a virtual world. Bartle's categorizations break a virtual world into a main dichotomy of "Players" and "Game". I think that in any virtual world there are often much more layers of gameplay than just those two. The reason for this is that whereas Bartle admits to not "having fun" in Virtual Worlds, I can admit to having, from experience, fun in my own ways in Virtual Worlds. (Barring pure egotistical debate that I "don't know as many of the magician's tricks" as Mr. Bartle, it does lead me to suspect that the system could use widening.) Extending it can only really add those layerss which I find fun that are currently outside the existing chart... Bartle himself might well still remain off the chart, but I think that the added depth will explain things better.
What is needed is a good general framework to represent the major such layers in a Virtual World without dealing with all specific systems (or variations of systems such as codebases, turn-based rpg vs. mostly-real-time rpg vs. puzzle-based). The system I think works best for this is actually modified from an older "English class" system for describing literary conflicts, that some may or may not remember from earlier school years. I think this makes a large amount of sense when taken from the point of view that a player sees his journey/path through the Virtual World as a personal narrative.
The system in relation to Virtual Worlds:
- Society/Clan/Guild (groups of players)
- Designers (Gods)
Self is included for completeness, and arguably is still important in Virtual World design.
What I find most interesting is that the two larger categories effectively mirror the lower categories. Just placing these two new categorizations I've added in their own graph yields:
Here are some initial thoughts on the new classifications exposed by this graph:
First off, splitting the "Players" from Bartle's graph into "Player" and "Society" makes some sense and here breaks apart things into a major "micro" and "macro" level. In this particular case the interesting thing is that in seperating the Meta-Killer from the Killer we notice that a Killer (who enjoys conflicts with other individual players) may not be so preferable to a designer (just as thieves aren't the best element to fill a city with), whereas the Meta-Killer who prefers conflict with groups of players may in fact be an asset (where games prefer Realm vs. Realm, Guild vs. Guild combat forms these people can be good Guild/Realm members and leaders).
There isn't as much difference between the Socializer and Meta-Socializer as there is between the Killer and Meta-Killer. Here it seems that the only real factor is scale.
I believe that every Virtual World designer has had more dealings with meta-achievers than they would care to have had... The meta-acheiver would be the type of person to spend lots of time in the game's forum trying to achieve certain goals such as annoying the designers, or influencing the designers. I would guess, simply from his position and interest in these topics (as well as his admitted lack of "fun" within virtual worlds, and I would assume the "fun" he has from dealing with meta-virtual world objects such as these book itself) that Bartle would be a Meta-Acheiver. (Obviously not one of the annoying newbie meta-achievers you see on forums... a well-read textbook can be much more influential and cause much more achievement than can individual forum posts.)
This is where I would place myself. I enjoy interacting with a good world design: learning it's limits, it's boundaries. From personal experience I would say that Meta-Explorers are much less like than Meta-Acheivers to enter the forums, and when they do it is largely to "lurk" (to browse, listen, learn). I got bored with Cryptic's City of Heroes once I had run out of Meta-Exploration and realized that the game had little content for my (Meta-)Socialization side to fall back on. On the other hand I believe that I may be spending a long time in Three Rings' Puzzle Pirates because it has some real neat design features that constantly feed my Meta-Exploration side, whereas the Pirate-y theme engages my (Meta-)Socialization side. (Furthermore, the puzzle-based nature engages my self-competition.)
Spinning the wheel...
The last paragraph alludes to how I think the two graphs can best be utilized. If you recombine the two graphs into the following (ultimately widening the dichotomy between players and world, and perhaps leading toward a further more generalized system that could deal with further layers of a virtual world):
I believe that these two graphs much better represent the types of players in a Virtual World and that you can represent a person better by their categorization on both charts. I think that you can also notice a much closer relationship to the quadrants of these two graphs than in Bartle's. For instance, I would classify myself as a Meta-Socializer and a Meta-Explorer. I would say that I lean more towards Socialization and Meta-Killing than I do towards Killing itself. Similarly I feel that I lean more towards Exploration and Meta-Achievement than Achievement. I think that this then is a much richer platform to make judgements and estimates of how I might feel in a given situation as "fun" than Bartle's system and would perhaps better explain some of the prevalent Bartle Quotient scores.