Will The Real Explorer Please Stand Up?

There is an ongoing discussion at another thread about how Thottbot impacts Explorers. While several people identify as being an Explorer, and others talk about what an Explorer is, it's not clear whether everyone is talking about the same thing

Bartle's original definition in his paper was:

[Explorers] try to find out as much as they can about the virtual world. Although initially this means mapping its topology (ie. exploring the MUD's breadth), later it advances to experimentation with its physics (ie. exploring the MUD's depth).

[Aside: Bartle's Test (not made by Bartle) creates the illusion of identifiable types because of the dichotomous forced-choice format and isn't relevant to our discussion here of what an Explorer is. If I ask you whether you like poison or spinach over and over again, just because you keep picking spinach doesn't actually mean you like spinach at all.]

A few people commented on the "cartographer"/"loremaster" aspect of the Explorer:

Some "explorers" come to a game, map it out, put the map on the web, then quit.

- Ola Forsheim

Being able to find rare creatures, or to quickly get from one location to another in short time, is the type of thing that has hitherto been the Explorer player's small prize for his unusual playstyle.

- Grax

But most seemed to agree more with Dave Rickey's "Analyst" portrayal of the Explorer, and that this was the more pure Explorer type:

Most "explorers" are more interested in the underlying rules of the system from what I have seen.

- Dave Rickey

Flatfingers elaborated on the "analyst" motivation, and implied that "cartographers"/"loremasters" weren't the real Explorers:

My personal belief is that Explorers are less interested in cataloging raw data (which anyone can do) than in revealing hidden patterns in data. The developer who interprets "exploring" as some literal walking around physical space to visit a new location has misunderstood what it means to be an Explorer …  Explorers feel satisfaction when they can move from the specific to the general.

- Flatfingers

This got more complicated when others suggested that Explorers do what they do to gain a competitive edge:

When information is a rare commodity, investing time and energy into gathering valuable information (which might give a competitive advantage) would seem to be another player metric of achievement ... It would seem to be harmful to explorers only to the extent that obscure knowledge might give a competitive advantage over those who lack it.

- Barry Kearns

The problem with this portrayal, as Michael Hartman pointed out, is that:

Such a person is *NOT* an explorer then. That is an achiever.

- Michael Hartman

Everyone seems to know what an explorer is. The problem is that they don't agree with each other. Here's what I know from data from online surveys of MMORPG players.

1) "Cartographers" are not correlated with "Analysts". Responses to the following two statements are not strongly correlated. In other words, they are different kinds of people altogether.

- How much do you enjoy knowing as much about the game mechanics and rules as possible?
- How much do you enjoy exploring every map or zone in the world?

2) The "Analyst" type however is highly correlated with the Achievement motivation. So the "game mechanics" question above correlates highly (r = .46) with the aggregate of these statements:

- Leveling up your character as fast as possible.
- Acquiring rare items that most players will never have.
- Becoming powerful.
- Accumulating resources, items or money.

In other words, people who enjoy learning about the game mechanics are largely also your typical Achievers. So the real question seems to be whether Explorers are really a sort of Achievers.

Then the question is whether current MMORPGs have driven away all the real Explorers, or whether game designs always embed exploration within the context of achievement and make it difficult to tease these two motivations apart. Or is it the case that all Achievers are inherently Explorers. After all, no one else is as interested in whether dual-wielding outperforms two-handed weapons. And isn't that what it means to understand the underlying rules, and moving from the specifics to the general?

But if Achievers are usually "Analyst"-style Explorers (the min-maxers), then ... who are the real Explorers? Does it make sense to have a separate Explorer type if it overlaps so much with Achievers? Or maybe the Cartographers are the real Explorers?

If you feel that a pure Explorer type does exist - What are statements I could include in future surveys that would identify Explorers? What are statements that Explorers would agree with that Achievers would never agree with? What sets Explorers apart?

[Another Aside: Bartle doesn't advocate strict "Types" as others pointed out in the other thread, but the Type model assumes a primary motivation. I prefer a Scores model where every player has a score for every motivation and where having a high score in one motivation doesn't mean you also can't score high in another motivation (which Bartle's Test [not created  by Bartle] assumes). This would allow more interesting combinational outcomes - so a guild leader is someone who scores high on Socializing and Achieving, a min-maxer is someone who scores high on Achieving and Exploring. I feel that using the Types framework prevents us from discussing how motivations combine and interact to create the range of roles that people choose in MMORPGs.]


Comments on Will The Real Explorer Please Stand Up?:

Jeff Freeman says:

If you feel that a pure Explorer type does exist - What are statements I could include in future surveys that would identify Explorers? What are statements that Explorers would agree with that Achievers would never agree with? What sets Explorers apart?

"Are you a game designer?"

Posted Jan 27, 2005 7:30:28 PM | link

Nick Yee says:

"Are you a game designer?"

I used to wonder that. But then if only game designers are Explorers, and they're interested in a game in order to design games, then is it a real player motivation?

I mean - there are all these social science researchers creating characters and are interviewing people for an external empirical motive - but we're not going to make a motivation for them, right?

Posted Jan 27, 2005 7:38:21 PM | link

Carl Potts says:

I think the explorer type does live on, although the boundaries between that player type, and the achiever type have blurred somewhat.

Personally I think the 'Maximiser' type identified in the 'Rubicite Breastplate'paper would be a more appropriate definition for the modern MMORPG market. It would certainly explain why at times there are underpopulated tracts present within many virtual worlds, while MMORPG design continues to reward those who choose to camp locations. With my limited knowledge of WOW (waiting on euro release) it seems WOW seems to break this trend.

This is not to say that the Explorer type has dropped of the face of the (virtual) world. A joy in working out the inner dynamics of virtual worlds is still present for many players, notably regarding features such as agro, DPS etc - does this concern for the inner workings have a side affect on factors such as immersion?

Posted Jan 27, 2005 8:40:51 PM | link

Tom Bloodgood says:

Personally, I am an explorer type. Explorers ARE achievers to the extent that instead of trying to beat the game or get the coolest stuff, we are more interested in seeing the whole game, completing our in-game map, and finding out where all the good stuff, like item drops, comes from. Typically, explorers also like to figure out the best way to get somewhere without getting killed.

The best example in my own game play is from original EverQuest: I played a Warrior. I rarely carried any SoW potions or spells, just Jboots, no invis pants or potions and I never carried recall potions. I never asked for ports. Instead, if it was somewhere I had to walk to, I walked there. I would look it up on EQAtlas to find out where I was going, and then found the best way to get there by trial and error and paying attention to my environment.

Probably not too impressive a feat until you consider that whenever players found out I was doing this with the class I played, they were stunned. Routinely, I was told it was too risky and a waste of time and they always asked if I died a lot. This amuses me because I rarely ever died.

To the explorer, being able to go anywhere and know where you are going and what to expect is more important than actually beating the mob or getting the sword. It's the discovery of knowledge that is the focus.

Posted Jan 27, 2005 9:39:58 PM | link

Hellinar says:

I’d count myself an Explorer. For me that means exploring the range of Possibility in a world. Whether it’s the possible types of scenery, beasts, swords or stories depends on your tastes. Where the Analyst aspect kicks in for me is in deciding if I have seen pretty much the range of possibility in that world, and move on. In WoW for example, given my knowledge of computer graphics, I know that by now I have seen pretty much all the beast types in the world, and anything else will be like something I have already seen but with a different skin. I’ve seen a good range of the zones, a lot of the armor available, and have no interest in weapon types.

In my mind, the mark of the Achiever is measuring yourself against some linear scale, and increasing your score. The competitive Achiever then compares that score to others. Though you can be an Achiever who just increases their personal score without regard to others. Exploring, in my definition, is a much more non linear grasping of the pattern and possibilities of the world. A holistic kind of thing. In that definition, its absurd to talk about “competitive exploring” as exploring is a personal relationship between you and a space. Other peoples relationship to the space is not really relevant to your own exploration.

Of course, any real person is not likely to be pure explorer by that definition. How to detect Explorers in my sense of the term? A tough one. Maybe a question along the lines of:

You have enough gold to fund a trip to one of two new zones. One has similar scenery and beasts to those you have seen before, but the beasts are a good match for your new level. The other has scenery and beasts that are new to you, but the beasts are of a level that gives few experience points for you. Which trip do you take?

Taking the second option, at least on your first run through the game, would perhaps indicate an Explorer.

Posted Jan 27, 2005 11:26:08 PM | link

Jeff Freeman says:

I mean - there are all these social science researchers creating characters and are interviewing people for an external empirical motive - but we're not going to make a motivation for them, right?

Yes, actually. I think we did.

Posted Jan 28, 2005 2:09:47 AM | link

Brask Mumei says:

My understanding of "Explorer" is closer to the "Johnny" Magic The Gathering player type.

http://www.wizards.com/default.asp?x=mtgcom/daily/mr11

An explorer seeks to do neat stuff with the game. They don't want the maximum level, or have methodically checked off the entire map. They do want to gain levels in strange ways, or find unusual ways to get to places on the map. An explorer seeks to do things differently and interestingly.

A question may be:
You have a choice of two weapons. Weapon A does more damage and is more reliable than weapon B. But weapon B has a strange and useless power.

The explorer will pick weapon B.

This is why I wouldn't think explorers would be all that invconvienced by Thottbot. Yes, it is macroing the explorer's task. But exploring is all about macroing! The reason why one does things in the interesting way rather than the proper way is because one wants to avoid mundane, repetitious, tedium. (Which may be why these LevelQuest games end up lacking this player type...)

When I wanted to test the random number geneator in UO, I had no compunctions with setting up an EzMacros routine to roll dice overnight, and then parse the resulting text log through automated tools to look for patterns.

Where I think explorers feel cheated is when their personal game style becomes popular. This is usually a result of the development team balancing content to try and move the Achievers away from the Best Item.

"Two paths diverged in the woods. I climbed a tree. That made all the difference."

(Apologies to Frost)

Posted Jan 28, 2005 2:35:26 AM | link

Damion Schubert says:

I think the problem when most people talk about the Bartle's 4 is thinking that each of the four are mutually exclusive. Indeed, when asked what I am, I say I'm a 'completionist' - I'm an achiever, but the achievement is exploration.

There are different shades of all of the groups. Achievers can achieve more than 'seeing everything' - it's clear that the classic Bartle's pure Achiever isn't into 'seeing everything'. He wants to get to level 60, and get there first if at possible.

I think if you scratch the surface, you'll find that recognizable hybrids exist all over the Bartle's Four Landscape.

Posted Jan 28, 2005 2:39:55 AM | link

Cenn says:

does anyone have a url about those 4 types - id like to see more info on them

Posted Jan 28, 2005 2:45:02 AM | link

Mike W says:

Ernest Adams had a designer's notebook column a while back that I think has a better model of player motivation.

http://www.designersnotebook.com/Columns/063_Postmodernism/063_postmodernism.htm

In fact, if you add a fourth type of immersion and call it something like "social immersion" you get a pretty good mapping to Bartle's types.

Killer's are likely interested in tactical immersion, Explorer's in narrative immersion, and Socializers in social immersion (obvioisly). This leaves Achievers with strategic immersion, which many people seem to have advocated so far anyway.

Although I don't really agree with this last one, since I am totally a strategist, and not an acheiver (or any of Bartle's types)

Really I would probably just lump achiever's in with socializers (since it would seem they need a social group to justify their achievements). And add a completely separate strategist type, but now we've pretty much mangled that model beyond recognition.

So under this model you would find achievers by looking for people who want to be considered best/competent by their peers, while explorers want to develop their character's place in the narrative immersion of the world.

Posted Jan 28, 2005 2:51:12 AM | link

Cenn says:

sorry- just read the paper by bartle... makes sense now.

Posted Jan 28, 2005 2:53:58 AM | link

Bruce Woodcock says:

It's my personal curse to be a strange subtype hybrid of the Explorer and the Achiever, which I call the Completeist.

I want to do every quest, discover every location, get all of the "rares", and so on. But my motivation is not Achievement in the competitive sense -- I don't want to do it first, necessarily, or do it best. Nor is my goal the pure joy of discovering something new. I simply want to have done it, and hopefully done it "my way".

So web sites like thotbot are great for me. Because I can be sure to get any content that I might have missed in my random walk through a game. Also, I don't know any explorers who feel thotbot hurts them; most of their joy is in making the discoveries themselves, so they simply don't look at a place that has all the answers. There's some amount of joy lost to those who like to secretly possess obscure knowledge, but that's a minor component and that need can be satisfied in numerable other tidbits that such sites don't record.

Bruce

Posted Jan 28, 2005 2:57:37 AM | link

Michael Steele says:

Before we get too wrapped up in symantics, let us not forget these taxonomies (whether B4 or something else) are really about core human motivations, and not about specific behaviors. As such, they can remain (properly) game in-specific. It is too easy to start making sub-sub-categories to explain all the possible behaviors we observe. Motivations however, are more basic and fundamental, even universal. For instance, the urge to go over the next mountain might be driven by a desire to be the first to conquer the new mountain, or to find knowledge no one else yet has. Or maybe even just to see if you can do it. Those are all Achiever motivations, not Explorer. If the idea of not-knowing what's behind the hill *compels* you to go look, well that's an Explorer motivation.
Similarly, Achievers don't always need to be competing against other people... sometimes they are competing against the game, or in another instance athletes will often compete against themselves, driven by the need to perfect their scores or beat their own best time. We clearly see Killer motivations driving certain gameplay behaviors in both WoW and in Hearts (a cut-throat game, to be sure!). The desire to defeat a human opponent in a competition seems to be the only commonality there.

It is true we frequently need to devise finer-grained taxomonies that expand on the basic B4 set in order to make sure we're covering the core needs of all our players. Thus we can subdevice Socializers into Samaritans/Connectors/Den-Mothers, or Achievers into Mavens/Athletes/Collectors/Solvers, etc. etc. but let's always keep the simplest test in mind: Are we describing a core human motivation, or just some observed secondary behavior that might need be further reduced?

Posted Jan 28, 2005 4:32:50 AM | link

Richard Bartle says:

I suppose I should say a few words here.

The point of the types in the original paper isn't so much as what they do as why they do it. We can, by looking at what players do, hope to figure out why they're doing it. However, this is not always the case.

Example: I had a player who was the friendliest, most helpful guy you could wish for. He'd come to the aid of complete strangers, heal them up, give them useful items, in true Robin Hood style. Over the course of several weeks, he'd become friends with them, join their group, go adventuring with them. Then, when they were the worse for wear and relying on him to fend off attacks, he'd suddenly turn and slaughter the lot of them. He was a killer. For 99% of the time, he looked like some kind of do-gooder achiever, but no, he was a killer.

The important thing is WHY people do what they do, not what they do.

As for whether people can be more than one type, well yes, they can. This is because over the course of time players change type, and sometimes their goals refer to the type they once were or the type they're leaning towards becoming.

Another example: players will often say things like, "Sometimes after a hard day's , you just want to ". This is because, in time, many achievers go on to become socialisers. You rarely see people say "Sometimes, after a hard day's you just want to ". This isn't to say that socialisers can't become achievers, but it does say that it's not a popular path to take. More on this shortly.

OK, so what I've explained here applies to my 4-types model. There are some problems with this, though, one of which is that although it says that players can change types, it doesn't explain why they change types. This is explained by the expanded 8-types version I presented in my book,
Designing Virtual Worlds. Here, I double up all the original types by adding an extra dimension to the model, and therefore have two different kinds of explorer:

1) Scientist.
- Experiments to form a theory.
- Uses theories predictively to test them.
- Methodical acquisition of knowledge.
- Seeks to explain phenomena.
2) Hacker.
- Experiments to reveal meaning.
- Intuitive understanding, no need to test.
- Goes where fancy takes them.
- Seeks to discover new phenomena.

Scientists are aiming to expand their knowledge, but eventually reach a point where they want to apply it. "I'm sick of , I want to ". Hackers are like gurus, who know because they know. They're an end-point of the player types progression.

The other end-point is a version of socialiser. As I said, 4-model types are doubled up in the 8-model version. Although "classic" socialisers (ones who hang around all day chatting simply because they enjoy each other's company) don't become achievers, some socialisers do: these are the people who through early socialising find out enough about the virtual world that they want to apply what they've learned. This kind of player can go on to become an achiever.

I guess what I'm saying here is that you're welcome to discuss the 4-types model if you like, but the 8-types model is the one I advocate myself. Unfortunately (for me!), not enough people seem to have read my book for this to be viable subject matter here. I have a chapter in the up-coming Massively Multiplayer Game Development 2 that consolidates what in my own book is spread over several hundred pages.

When people say they are one or more types, what it means is one of the following:
1) They do things that more than one type would do, although they generally have a single over-arching reason for doing it.
2) They are changing their reasons for doing things. My 8-type model makes a stab at predicting which is the "from" and which is the "to", and where they'll go next.
3) My theory are incorrect and someone needs to come up with a better one.

Richard

Posted Jan 28, 2005 5:21:16 AM | link

Grax says:

Good entry, Nick Yee.

I think there is a (philosophical) point which needs to be stated: That we are all achievers.

Yes, everyone is an achiever, but we are all trying to achieve different things. Now, semantic confusion primarily arises due to the fact that we call some people "Achievers". Why do we call them Achievers, when it is evident (to me, at least) that everyone wants to achieve? I think it would be very useful to flesh out the definition for "Achiever", in light of the fact that everyone is compelled by achievement goals of some sort. So, a more complete description of an Achiever would be: "An individual who perceives winners and losers in a game in a standard way (usually levels/guildstatus/items), and who will then play in such a way that he achieves things that are best rewarded in-game." Whether the Achiever is modifying his playstyle so that he is best rewarded by in-game mechanics, or whether his playstyle just happens to be the one that is best rewarded by the game mechanics, probably does not matter. In the final analysis, he ends up an Achiever, even if he WOULD have played another role in a game that provided more in-game incentives for Killer/Explorer/Socializer playstyles.

I said everyone is an achiever. I'll briefly go through the other Bartle types with this statement in mind:

Killers-
Killers wish to be the best at killing, or they wish to achieve a certain notoriety, or they may even wish to become the most powerful players in game (levels/items/guildstatus). Even if a Killer-type player seeks the more standard achievements on top of his more Killer-specific goals (in other words, even if he wants to be the highest level, best geared, most powerful player on top of his primary concern: that he be known and feared as a killer, etc.), he will primarily stick to his path of killing, even if it is not at all the most effective way of satisfying his secondary, more standard, desires (levels/items/power). Why? Because he is a Killer, and if he chose to give up the life of Killing in order to be better rewarded by in-game mechanics, then he would have become an Achiever; we can perhaps argue here that he wasn't ENOUGH of a Killer to remain a Killer in a game that offered limited in-game incentives for this playstyle.

Socializers-
Socializers seek to 'achieve' a high status within the game, whether it's for their knowledge of the latest rumors, or for their connection to powerful Achievers, etc. It seems that achieving a high socializing status often requires highly-esteemed clothing, so many Socializers may develop Achiever concerns, in order to further their Socializer goals. Socializers wish to be highly-esteemed by their fellow players, and, generally, players will esteem highly anyone who is highly-esteemed by the game (this is essentially the same as saying: Many of the big MMOGs have mostly Achiever/Socializer players*). Therefore, Socializer and Achiever concerns often become strongly tied together to produce the type of individual who: joins a (powerful, often) guild in which he makes friends, then earns items/levels to impress his friends; or the order can be reversed, depending on the player's primary concerns. Either way, in the end he will often 'become' both a Socializer and an Achiever because these types end up being strongly connected in MMOGs today. Analogy: an Achiever taking on Socializer concerns could be likened to a Nouveau Riche buying a piece of modern art**, and a Socializer taking on Achiever concerns can be likened to a woman choosing a husband (ie. join a powerful guild) based upon the status that she will be guaranteed.

Explorers-
Finally, I'd like to talk about Explorers. I'll begin by saying that I think a distinction needs to be made between Explorers and what I would call Game Analysts***. I think that an Explorer ought to refer to an individual who seeks to play within the boundaries of the game, while a Game Analyst would be someone who is not only willing to play outside the boundaries of the game (using Thottbot for analysis), but will mix the two up the way Neo (pardon the reference) slips in and out of the Matrix world. Most Explorers, like anyone else, are very much into achievement. They wish to play the game as an Explorer, and ideally, they would be rewarded in-game for their extensive knowledge (cartography, spell experiments, dungeon traps, whatever).

Many people whose primary concerns distinguish them as Game Analysts will call themselves Explorers (because, most likely, they are Explorers AND Game Analysts; although they, by definition, care about the latter more than the former). What do they wish to achieve? Well, Game Analysts may be the only playertype covered thus far whose Game Analyst 'achievement' concerns are actually outside the game (if people in the game are impressed with his analysis, it does not mean this valuing is not occurring "outside the game"), and so we can say that they are the least affected by whether the game rewards their playstyle or not. Of course, an Achiever, Socializer, Killer, Explorer who takes on Game Analysis (uses Thottbot, sites, etc.) as a minor subprofession will reap enormous in-game rewards from doing so. Game Analysts often circumvent the game mechanisms in place, and this type of ability is, obviously, useful to all of the Bartle playertypes. Similarly, to be able to fly would be useful to pretty much all Olympians. The ones who wouldn't fly because they had a belief that one should walk, or play within the traditional boundaries of the game, are the equivalent of those individuals who have access to Thottbot but choose not to use it.****

What sort of goals fall under the Explorer playertype? The goal to be seen as a fountain of wisdom and knowledge; or, the personal, almost Buddhistic, goal to 'be closer' to the virtual world by knowing it so well (and there are still other goals). The 4 Bartle types can be considered classes, basically. Most people want to be 'seen' a certain way in the game, and this is what dictates their behavior and, therefore, their playstyle/playertype. People are, whether they know it or not, complaining that their playstyle isn't rewarded enough in-game, in the same way that people complain that their class is underpowered. In many cases, I would say these "playstyle X is underpowered" complaints are appropriate: namely, when the playstyle is one that SHOULD exist in a significant way in the game, and when it is actually UNDERPOWERED (because it is circumvented {see BORING} and the advantages it is supposed to have are freely available to all other classes) or BORING (requiring circumvention).

*Whether this is more to do with the players' natural desires, or more to do with the game's use of positive reinforcement to encourage the two popular playstyles, is not relevant here.
**Art is meant to lie in the domain of the real-life Explorer, but, interestingly, it gets 'borrowed' for other purposes.
***or Game Designers, Game Testers, Beta Testers, ...
****Or, some Olympians who do choose to fly might decide to do so because they have little respect for their chosen sport, whether justified or not. They might, for example, think that their sport has some very boring requirements that ought to be circumvented.

Posted Jan 28, 2005 5:55:04 AM | link

Flatfingers says:

I've made noises about discussing how the Bartle types relate to other, more general theories of temperament. Somehow I don't think I'll have a better opportunity than this thread to present the key points of that thesis. ;-) So here it is.

I need to start by agreeing that this kind of thing would probably work better written up as an academic paper, with footnotes and properly cited references and so on. But a reply in a discussion thread is not the right place for a lengthy essay. (Plus I'm not an academic!) Instead, I'll try to summarize the essential ideas as a series of bullet points. That form means I won't be able to back up each statement with the supporting facts and logic I normally try to provide when making strong assertions, which is going to make the whole argument seem weak in places and expose it to objections that a less brief treatment could preemptively address.

But the fact that I don't spend any time backing up the following assertions doesn't mean I'm making them up on the spot -- it just means I'm trying to be succinct to avoid abusing the privilege I've been given to post on this forum. Feel free to question anything that doesn't make sense to you... but try to go after my conclusions first, if you don't mind; those are the bits that matter.

One other note: Any and all references to the work of both Richard Bartle and David Keirsey that I make here are purely my own interpretations, and are not to be considered "official" explications of these authors' ideas. If I fail to correctly or completely capture some aspect of meaning intended by either author, the error is mine and should not be attributed to either Bartle or Keirsey. Basically, if you think any of my ideas are nuts, don't blame Keirsey or Bartle -- blame me. (And this includes the point that Richard makes above, which is that the four-type model I use here isn't his most current and preferred model. Clearly, I have some book-buying to do. Until then, I hope I can be forgiven for using the Web-available four-type version, which I believe still has some useful explanatory power.)

OK, let's get to it!

The Background Observations:

* Type classification systems aren't the only way to to study human personality -- they're just one useful way.

* An example that uses multiple codes (like the Bartle Test) is the Holland Codes assessment for job placement.

* No system for classifying human personality will ever be perfect. People are complex; no one is purely one type only, nor can any system fully describe an actual person.

* A classification system doesn't have to be perfect to be useful; it just has to have some reliable explanatory power.

* "Temperament" is a form of typing that considers fundamental, distinct, and possibly innate personality types.

* Many different theories of temperament have been suggested over thousands of years.

* Most temperament theories (Hippocrates, Plato, Aristotle, Paracelsus, Spranger, etc.) consider four temperaments.

* These models all tend to have certain expressions of human nature in common.

* Jung's theory of archetypes and the Myers-Briggs type system based on Jung's work rationalized these common traits.

* David Keirsey's temperament model is a modern synthesis of past models organized around Myers-Briggs empirical data.

The Big Assertion:

* The four Bartle types are online game player-specific analogues to Keirsey's four general temperaments.

The Details:

* The Myers-Briggs model posits 16 types, based on four dimensions of personality.

* The four dimensions (and the two poles of each dimension, along with the letter abbreviation) are:

+ Need for social interaction: Introversion vs. Extraversion (I/E)
+ Source of truth: iNtuition vs. Sensing (N/S)
+ Decision-making style: Thinking vs. Feeling (T/F)
+ Time focus: Judging vs. Perceiving (J/P)

* Combining all possibilities yields sixteen "types":

ESTP, ISTP, ESFP, ISFP
ESTJ, ISTJ, ESFJ, ISFJ
ENFJ, INFJ, ENFP, INFP
ENTJ, INTJ, ENTP, INTP

* Keirsey's model groups the sixteen types into four groups of four temperaments, and names each temperament:

xSxP: ESTP, ISTP, ESFP, ISFP (Artisan)
xSxJ: ESTJ, ISTJ, ESFJ, ISFJ (Guardian)
xNFx: ENFJ, INFJ, ENFP, INFP (Idealist)
xNTx: ENTJ, INTJ, ENTP, INTP (Rational)

* Excessively brief portraits of Keirsey's temperaments are:

+ Artisan: realistic, tactical, manipulative (of things or people), pragmatic, impulsive, action-focused, experience-seeking
+ Guardian: practical, logistical, hierarchical, organized, detail-oriented, possessive, process-focused, security-seeking
+ Rational: innovative, strategic, logical, scientific/technological, future-oriented, result-focused, knowledge-seeking
+ Idealist: imaginative, visionary, emotional, relationship-oriented, dramatic, person-focused, identity-seeking

* In the second edition of his primary work, _Please Understand Me II_, Keirsey groups his four temperaments according to their use of tools ("cooperative" or "utilitarian") and words ("abstract" or "concrete"):

.....................WORDS........
............ Abstract.....Concrete
Cooperative: Idealist..|..Guardian
TOOLS....... ----------+----------
Utilitarian: Rational..|..Artisan.

* There's nothing wrong with that system, but by the time Keirsey proposed it I had already worked out an arrangement I thought (and still think) better reflects how people interact.

* Rather than two dimensions of tool-use and word-use, I think of the two most important dimensions of behavior as being "internals vs. externals" and "change vs. structure":

................CHANGE.................
..........Idealist.|.Artisan...........
INTERNAL..---------+---------..EXTERNAL
..........Rational.|.Guardian..........
...............STRUCTURE...............

* This structure has an additional feature: Each temperament is most unlike (and usually misunderstands or even opposes) the temperament diagonal to it.

* Thus, Artisans (who seek External Change) tend to perceive Rationals (who seek Internal Structure) as ineffective creators of imaginary and useless ideas, while Rationals see Artisans as ignorant, energy-wasting jocks.

* Similarly, Guardians (who seek External Structure) often regard Idealists as crusading liberal artistes, while Idealists (who seek Internal Change) see Guardians as boring, bourgois reactionaries.

* Next, here is the same kind of diagram, except that this one shows the four Bartle types (rotated clockwise 90 degrees):

.....................PLAYERS...............
.............Socializer.|.Killer...........
INTERACTING..-----------+----------..ACTING
...............Explorer.|.Achiever.........
......................WORLD................

* Considering Bartle's descriptions of Achievers, Explorers, Socializers and Killers, and comparing these descriptions against Keirsey's descriptions of the four temperaments in his model, and further noting the congruence between the 2x2 structure I saw in Keirsey's model versus Bartle's diagram, it seemed natural to explore the possibility that the four Bartle types are context-specific analogues to Keirsey's four temperaments.

* First, let's consider the dimensions of behavior that determine the 2x2 organizations of player types and temperaments.

* In "Players Who Suit MUDs", I interpret Bartle's use of the terms ACTING and INTERACTING as "doing" and "learning about," respectively. (Bartle uses the terms "doing-to" and "doing-with" to describe ACTING and INTERACTING, but he also seems to suggest that INTERACTING is more about understanding the properties of things than actually using those things.) INTERACTING can correspond to my notion of INTERNAL in that interacting with a thing is something you do to discover the internal nature of that thing. This contrasts with the concept of ACTING with or on a thing, which is an EXTERNAL usage of a thing.

* The case for PLAYERS and WORLD as analogous to CHANGE and STRUCTURE is less clear, but the key can be found in remembering that PLAYERS/WORLD are concepts appropriate for a game context, while a CHANGE/STRUCTURE model is intended to apply to a larger set of human behaviors. In this case, the concept of PLAYERS is a special case of CHANGE, while WORLD is a special case of STRUCTURE. CHANGE happens to be the word I used for this pole that describes how much or how little order/control a person needs or wants, but FREEDOM and OPPORTUNITY are also words that could convey this meaning... and these are all attributes that are unique to human players in a virtual world. Conversely, STRUCTURE/ORDER/BOUNDARIES are attributes of WORLD objects, which must all predefined by the game's creators. As Artisans (External Change) demand to be free to manipulate the people in their environment as they will, Killers (ACTING on PLAYERS) won't play if they are denied power over other players. Socializers as game-specific cases of Idealists and Achievers as game-specific cases of Guardians follow the same interpretations.

* There's a lot of subsequent analysis that can be done that discusses how each of the four Bartle types corresponds to one of Keirsey's four temperaments, but I'm already running way too long. So let's just consider Explorers as Rationals playing games.

* The natural approach of the Rational to play (as to everything else) is to look for patterns behind the raw data. These can be patterns in space (as in geography) or patterns in time (as in morphology), but ultimately it's all about achieving a strategic understanding of the system as a whole thing.

* There are different types of Rationals (seen in the four Myers-Briggs types that comprise each temperament). The two Extraverted Rational types (ENTP and ENTJ) will find ways to satisfy their need for internal structure (which basically means mental models of everything) that involve other players -- either organizing groups of players (the ENTJ "Field-Marshal") or creating new systems for players to use (the ENTP "Inventor"). Meanwhile the two Introverted Rational types (INTP and INTJ) will seek more personal understandings of the game -- either figuring out complex cause-and-effect relationships (the INTJ "Mastermind") or synthesizing a coherent model of how the entire thing works (the INTP "Architect").

...

There's much more that could be said on this, but I absolutely have to wrap this up now. A final note: I think one of the reasons we perceive Explorers as having Achiever characteristics is because there just aren't many examples of mass-market MMOGs that aren't strongly Achiever-oriented. To have fun in these games, you pretty much have to do Achiever things. I suspect that skews our self-representations. Because we see ourselves engaged in Achiever activities (since those are the only things that can be done), we conclude -- whether we're Explorers, Killers, or Socializers -- that we must be part Achiever.

If most games were Explorer-oriented, wouldn't Achievers, Killers, and Socializers conclude that they must be part Explorer? Is there a way this theory could be tested experientally (rather than through questionnaires)?

To conclude (and about time!), here's a last question: If most designers/developers are Explorers, why do they make games that are so tilted toward Achievers instead of making games that they themselves would enjoy playing? (I have my own theory; I'm just curious to hear what others think.)

Thanks for your patience in wading through this thing. Comments are welcome.

--Flatfingers

QUICK WEB REFERENCES

Bartle's "Hearts, Clubs, Diamonds, Spades: Players Who Suit MUDs": http://www.mud.co.uk/richard/hcds.htm
Keirsey's Temperaments: http://www.keirsey.com/
Holland Codes: http://www.tiu.edu/psychology/Robinson/PSY210/hollandcodes-intro.htm

Posted Jan 28, 2005 6:15:35 AM | link

Flatfingers says:

I should add that I stayed up all night writing this behemoth, and it probably shows. So I'm going to crash now before I do any more damage to this forum. ;-) But I'll definitely be back later to take my licks.

--Flatfingers

Posted Jan 28, 2005 6:33:59 AM | link

Peder Holdgaard Pedersen says:

As I see it, Analysts are to games what Explorers are to role-playing games.

Most MMORPGs can be (and are, often by the same person) played as either pure game or role-playing game.

I see Analysts as both a blessing and a curse. They are a blessing because they are the best play-testers and troubleshooters available - and because of their role as service providers to the rest of the game community. They are a curse because they unravel the illusions put in place to create the virtual worlds - showcasing the machine inside. They peek inside the chinese room, and destroy the suspension of disbelief.

Posted Jan 28, 2005 9:23:32 AM | link

Hellinar says:


Grax> I think there is a (philosophical) point which needs to be stated: That we are all achievers. <

Defining Achiever so broadly seems to take away any utility. It’s a variant on saying “we are all human”. I'd push for a definition of Achiever that revolves around the achiever need to have an external score to measure themselves against. Whether they then compare that score to others i.e. become Competitors, I would call and extra step.

Grax> Most Explorers, like anyone else, are very much into achievement. They wish to play the game as an Explorer, and ideally, they would be rewarded in-game for their extensive knowledge (cartography, spell experiments, dungeon traps, whatever). <

Huh? You are completely missing the point. You, sir, must be an Achiever. :) In game rewards are an external metric, they are not the point to an Explorer, at least in my definition. The payoff for an Explorer is an internal, personal thing. It’s a point the Patent Office seems unable to comprehend. They assume that people design novel things to make big bucks. But that is a very Achiever motivation. Seems to me much of the true novelty in the world has come from Explorers, people who do it because they can.

I’d mark the boundary between Achiever and Explorer very much on the external or internal reward metric. Nobody can be better at exploring WoW for me than I am. They can only get better at exploring WoW for themselves. Which is why Thottbot is not a big issue for what I would consider an Explorer. Its no more than buying a crossword puzzle book with answers in the back. It doesn’t stop you enjoying doing the puzzles, but does provide a useful outlet when its clear the puzzle creator is in a different mental space to you.

Flatfingers> To conclude (and about time!), here's a last question: If most designers/developers are Explorers, why do they make games that are so tilted toward Achievers instead of making games that they themselves would enjoy playing? (I have my own theory; I'm just curious to hear what others think.) <

My perception is that the modern graphical MMOG are built by and for Achievers. They have to be, getting such a huge project out the door is an Achiever kind of thing. I suspect that many of the early MUDs, were built by Explorers, because they could. But if you look at the promoted “features” of new MMOGs they are generally Achiever, better than thou, kind of things. “We have an easier leveling curve than the other guys”. “We have tougher encounters than the other guys”. “We have better tradeskills".

Like Grax, the current designers assume everyone else is an Achiever. As a case in point, there is no Interface button in WoW to turn off the experience bar, which is the first thing I would want to do in a MMOG. I had to wait till I could figure out the UI coding to do it. And, in fact you can’t do it, it’s the one interface element I found that is hard coded. You can only fake it somewhat by setting your experience gain to 100%. That for me speaks volumes about the Achiever mentality of the designers. Perhaps Nick could use a question along the lines of:

“If you could turn off the Experience Bar in your MMOG, would you do so?”


Posted Jan 28, 2005 11:07:44 AM | link

Peder Holdgaard Pedersen says:

EQ2 actually has the options to turn off and/or hide the XP bar - and some people do hide it! :)

Posted Jan 28, 2005 12:05:07 PM | link

Grax says:

Hellinar, I've only read the first part of your post, but I'd like to point out that it is you who have missed the point of my post when you saw my statement "I think there is a (philosophical) point which needs to be stated: That we are all achievers." You interpreted the statement as: "We are all Achievers", when I said "We are all achievers".

It is perhaps my fault for being unclear, but the capitalization of "Achiever" is supposed to distinguish it as the playertype, and the lower case "achievers" is meant to describe (subterranean) psychological desires, which I argue we pretty much all share, whether we fall into the category "Achiever" (which I define as the playertype whose achievements are ALIGNED with the game's reward systems), or not.

Posted Jan 28, 2005 12:16:28 PM | link

Jeff Freeman says:

As for whether people can be more than one type, well yes, they can. This is because over the course of time players change type, and sometimes their goals refer to the type they once were or the type they're leaning towards becoming.

This is a different perspective than Nick Yee's observation that people have multiple motivations simultaneously though, isn't it?

With regard to the four-types model, you were looking at MUD players circa 1990 or so - Do you think the demographic has changed since then? Is it possible that at that time a great many people continued to play MUDs long after they'd have been interested in them otherwise, because they were interested in MUD design and development? Enough for an exporer-type to exist as a type of player in its own right, that is.

> 3) My theory are incorrect and someone needs to come
> up with a better one.

My own sense of it is that Nick Yee's is closer to reality. I much prefer designs based upon the idea that individuals have multiple motivations (and so any discreet system or piece of content needs to appeal to a variety of motivations in order to satisfy any given individual), vs. designs based on the 'types' model which more often only satisfy a single motivation (and therefore no individual, on their own).

I mean, even if you're completely right and he's completely wrong, the content seems to have more depth if we pretend that's not so.

If the 8-types model is better, I hope it catches on soon. I think the 4-types model can be downright toxic.

Man I have so many books to read...

Posted Jan 28, 2005 12:20:17 PM | link

Hellinar says:

Grax> It is perhaps my fault for being unclear, but the capitalization of "Achiever" is supposed to distinguish it as the playertype, and the lower case "achievers" is meant to describe (subterranean) psychological desires, which I argue we pretty much all share, whether we fall into the category "Achiever" (which I define as the playertype whose achievements are ALIGNED with the game's reward systems), or not. <

Hmm. Well, I tend towards the position that all humans follow all the same motivations simultaneously all the time. The question is, is one dominant to the point of being a useful way to distinguish people in a particular context? In my second quote of your piece, it seems to me you are saying the subterranean “achiever” desire is the prime motivation of the Explorer, just well disguised. On the contrary, I am saying there is a quite different desire dominating the behaviour, which quite orthogonal to the scorekeeping/ external metric achiever desire.

Its that subjective, internally driven desire that I think is getting swamped in modern MMOGs. Perhaps because, being computer generated, its so easy to generate objective metrics? Another signpost to the Achiever orientation of current MMOGs is the dominance of acquiring stuff over creating a story for your character. WoW does quite a job on storytelling, but the only way to access it is through killing stuff and getting better equipment. Or the trend towards dropping the Journey for simply jumping to the final confrontation, as in CoH and probably Tabula Rasa. These features push towards an externally measured success/failure, rather than an internally judged subjective experience. That’s where I would make the dividing line between Explorer and Achiever types.

It’s a very hot topic for me, as you can perhaps tell from the length of my postings. My own VW, creeping towards completion, is aimed more at Explorers, than Achievers. It’s a gardening world, with the game objective of “creating moments of beauty”. “Experiencing Beauty” is not something someone else can do for you, so the basic measure of success is internal rather than external. Of course, the world would not be well rounded if it didn’t have some achievement appeal, but I expect it to be a long way from the current Achiever dominated worlds.

Posted Jan 28, 2005 1:22:32 PM | link

Ola Fosheim Grøstad says:

Jeff Freeman> This is a different perspective than Nick Yee's observation that people have multiple motivations simultaneously though, isn't it?

Did he observe this in-game or did he deduce this from a questionare? If the latter, then how? I mean, it is quite obvious that people are motivated by several things when they do stuff, but was this obsevered?

Anyway, it is pretty well known that trying to figure out what people are motivated by is very difficult, even using in-depth interviews and observation. So I think much of the discussion is moot. Do players know what motivates them? If they don't then how can you figure it out by asking them?

Jeff Freeman> My own sense of it is that Nick Yee's is closer to reality. I much prefer designs based upon the idea that individuals have multiple motivations

Yes, and they are also quite complex... But in order to relate different play styles to each-other you would have to simplify a lot. This is an area where Yee's factors fail, "immersion" isn't a playstyle. Of course, if the goal is to go beyond very broad play-styles then you end up with a gazillion factors... Bascially reproducing the entire body of research on motivation and personality traits + all qualities that could be attributed to virtual worlds, if not more... Good luck!

Bartle should have explained better what kind of systems his model was limited to and its application area, specifically what it doesn't do. Another point: the gameplay in the system it was based on isn't properly described which makes it difficult to relate to.

Maybe N-types versions (N>4) are better, if more is merrier, a very dubious position IMO. No models describing complex phenomena are accurate, not even close. That's ok. That's also something authors might want to be explicit about, what's the purpose of the model and what are the trade-offs for keeping it simple?

My main problem with the 8-types version is:

1. I don't understand the explanation of it so it is difficult to say much about it. I simply don't understand why categories are placed the way they are.

2. It lacks the basic qualities of the 4-types model. I.e. the attempt to relate types to each other.

3. It doesn't resolve my main complaints about the 4-types model. Since I still know of types that aren't covered by the 8-types version, does that mean that one should move to the 16-types version, then to the 32-types version, the 64-types version etc? What would you use it for? What's the point? What is it trying to describe?

Another more general point: being able to get players to categorize themselves by a set of categories doesn't really prove anything. People would be happy to categorize themselves based on any set of categories you hand them: black/white, conservative/liberal, red/blue, offensive/defensive etc etc...

The same can be said about quantitative surveys trying to validate models IMO. You don't know the motivation/reasoning for providing a particular answer...

While I am at it: why would anyone want to come up with predictive models in this area? I mean, it's even hard to predict what your own family members are up to... For virtual worlds it is a lot worse, the contexts are different from game to game, from player to player. Different players join the game for different reasons etc.

Posted Jan 28, 2005 3:02:14 PM | link

Richard Bartle says:

Jeff Freeman>This is a different perspective than Nick Yee's observation that people have multiple motivations simultaneously though, isn't it?

Yes, it is. It says that people can have multiple motivations, but only from certain sub-classes (particular pairs, or at most, threesomes).

>With regard to the four-types model, you were looking at MUD players circa 1990 or so - Do you think the demographic has changed since then?

Yes, but I don't think the demographic has anything to do with it.

>Is it possible that at that time a great many people continued to play MUDs long after they'd have been interested in them otherwise, because they were interested in MUD design and development?

That depends on what you mean by "great many people".

It certainly happened then, and it certainly happens now, too. The proportions are almost certainly different, because today's virtual worlds don't usually let players get to the stage where they become hacker types.

>Enough for an exporer-type to exist as a type of player in its own right, that is.

How many do you need to exist "as a type in its own right"?

What do you want a player types model for? I can give you many models - male/female versus old/young, say - but they're not all of use to designers. The 4-types model, love it or loathe it, has (and continues to be) of use to designers. It tells them that they need to think about who will play, why they will play, how they will interact with one another... It doesn't matter whether the model is correct; it's done its job if it causes designers to design virtual worlds that will please people other than the designers themselves.

>My own sense of it is that Nick Yee's is closer to reality.

What theory is this, exactly?

>I much prefer designs based upon the idea that individuals have multiple motivations

Why? And what designs do this?

>If the 8-types model is better, I hope it catches on soon.

If any model is better, I hope it catches on soon. We've had the 4-types model for 10 years; it's about time things moved forward.

>I think the 4-types model can be downright toxic.

It's like household bleach: toxic when used the wrong way, but a useful (if basic) tool when used as intended.

Richard

Posted Jan 28, 2005 4:51:20 PM | link

Richard Bartle says:

Ola Fosheim Grøstad>Bartle should have explained better what kind of systems his model was limited to and its application area

I did say what kind of systems my model was limited to: MUDs.

>specifically what it doesn't do.

It doesn't do anything.

>Another point: the gameplay in the system it was based on isn't properly described which makes it difficult to relate to.

It's talking about MUDs in general, not one specific system. You don't need to know about MUD1 or MUD2, those are only where the theory came from; it's not a theory of MUD1/MUD2, it's a theory of MUDs. Look at the virtual worlds you do know about and see how the theory fits. If it does fit, it can perhaps help. If it doesn't fit, ignore it or come up with a better theory that does fit. Besides, if I'd gone into detail about MUD1 or MUD2 I'd have been criticised for promoting my games.

>1. I don't understand the explanation of it so it is difficult to say much about it. I simply don't understand why categories are placed the way they are.

Well I took the 4-model and added another dimension so it's now an 8-model. The categories are placed the way they are because that's how categories get placed when they're split in two.

>2. It lacks the basic qualities of the 4-types model. I.e. the attempt to relate types to each other.

The 4-model dynamics still apply, it's just that with 8 types each dyanmic be broken up into 4 sub-dynamics. This gives finer control, but to be honest it's probably too detailed to be of much use to designers. Certain types of interaction are more important than others because of the numbers of player involved and the degree of intensity of their feelings, so it would probably be worthwhile enumerating the interactions at some stage and pointing these out. It's not that the 8-types model lacks the basic qualities of the 4-model, it's that I didn't list them in my book. They're there, it's just that the original statement of the model doesn't include them (it just includes an explanation as to why it doesn't include them).

>Since I still know of types that aren't covered by the 8-types version

What types are these?

>does that mean that one should move to the 16-types version, then to the 32-types version, the 64-types version etc?

If your types are of use to designers, yes, it does. If they're useful but not orthogonal or independent, no it doesn't (that means instead that you should create your own formalism).

>What would you use it for?

To help understand the reasons that people play virtual worlds, that you may make a better virtual world as a result.

>What is it trying to describe?

The full 8-types model is trying to describe what people in virtual worlds find "fun" at any particular stage in their playing career. It attempts to predict what they will find fun next, and suggests ways by which they might find more things fun. People can do with this what they will, but I'd hope that it might at least inspire some designers to give thought to what they provide for their players (even if the thought they give concludes that my arguments are specious).

What do you want from a player types model? Or don't you want a player types model at all? If the latter, can you envisage a different kind of theory that attempts to explain what my player types model explains? What would its starting point be? Or is the very notion that such a theory could exist completely preposterous and we're all wasting our time here?

>Another more general point: being able to get players to categorize themselves by a set of categories doesn't really prove anything.

I agree. I'd much rather get the information through data mining, although even that is subject to misinterpretation (you know THAT someone has done something, but not WHY they did it).

>While I am at it: why would anyone want to come up with predictive models in this area?

Well I did it because people kept harping on at me about how my player types model didn't account for why people gradually changed type over time. Now it does, and you're harping on at me for my having done it.

>I mean, it's even hard to predict what your own family members are up to...

On a day-to-day basis, yes, but in the long term you can make some fairly accurate predictions. I don't know what my kids will do tomorrow, but I am fairly confident on some of the things they will have done by the time they're my age.

Richard

Posted Jan 28, 2005 5:23:20 PM | link

Ola Fosheim Grøstad says:

Richard Bartle>I did say what kind of systems my model was limited to: MUDs.

Which doesn't really say much at all.

>It's talking about MUDs in general, not one specific system. You don't need to know about MUD1 or MUD2, those are only where the theory came from; it's not a theory of MUD1/MUD2, it's a theory of MUDs.

I certainly need to know more about the case if it is a study of a case. Otherwise you basically ask me to view it as non-empirical.

Right now it's more like studying a particular vehicle, present a theory that claims that it holds for transportation in general while at the same time refuse to tell whether you've studied a bike, a car, a snow scooter, a plane...

>The categories are placed the way they are because that's how categories get placed when they're split in two.

Eh, yes, but I don't understand the categories and why they are placed the way they are... Or, I don't understand the relationship between the dimensions and the categories and where they came from.

>It's not that the 8-types model lacks the basic qualities of the 4-model, it's that I didn't list them in my book. They're there, it's just that the original statement of the model doesn't include them (it just includes an explanation as to why it doesn't include them).

Uhm. A model doesn't exist without a description. So it is clearly not there?

> To help understand the reasons that people play virtual worlds, that you may make a better virtual world as a result.

Well, but your model doesn't really say anything about the reasons. It might say something about what type of playstyle people rate as most attractive _after_ being exposed to a particular design.

>The full 8-types model is trying to describe what people in virtual worlds find "fun" at any particular stage in their playing career. It attempts to predict what they will find fun next, and suggests ways by which they might find more things fun.

But haven't we discussed this a length on MUD-Dev time and time again and reached the conclusions that there are no clear typical paths for playstyles for the general case?


>People can do with this what they will, but I'd hope that it might at least inspire some designers to give thought to what they provide for their players (even if the thought they give concludes that my arguments are specious).

Yes, but then it would be better for designers to concoct their own models based on how they expect their particular design will be appropriated.

Otherwise your model will be normative (in relation to design), are you prepared to take that responsibility? :-) I.e. take the responsibility for a design that tanks.

> What do you want from a player types model?

Play styles models are ok (player types models are more problematic), but they would have to take the context into account. So yes, if you had limited it to a set of designs; in general, no. You'd be better off with models from psychology IMO.

>Or is the very notion that such a theory could exist completely preposterous and we're all wasting our time here?

If you change that to plural: yes, theories do exist. To what extent they describe reality is another matter, of course, and will vary from case to case (theory A might describe X well, but fail completely for Y), but that's probably not what you meant as it is obvious... :-)

> Now it does, and you're harping on at me for my having done it.

Not really, but it is less useful as a model as you cannot possibly believe it holds? Clearly it depends on the particular player, ingame culture and design?

Mapping out play style progression is a useful technique IMO.

Ola.

Posted Jan 28, 2005 6:23:38 PM | link

Grax says:

Ola, I find your posts extremely uninformative. I get the impression that you are working from a theory and then trying to apply it to a world that may or may not exist. What is the point of such systematizing?

You say things that are obnoxiously false, such as: "Do players know what motivates them? If they don't then how can you figure it out by asking them?" Please take a moment to consider what you have just said here.

I'd like to upgrade my post to a 'flame'*, and would like to point out that I was not surprised when I discovered (on your website) that you are a "Neoquaker" who believes in God "out of necessity".

*This flame may not be 'justified' in the sense that you are unpleasant (you are pretty civil), but you fill the air with mindless twaddle, and I can't help but find this annoying.

Posted Jan 28, 2005 6:35:06 PM | link

Ola Fosheim Grøstad says:

Ah well, Grax. Do people know what motivates them? Will an artist be able to tell you why he did what he did? Or indeed, why he is spending his life such an ungreatful path. Because he "has to" or because it is "important"..?

It is much much easier to say something about what you like, to explain why is so much more difficult, if you really want the REAL why. Why did you fall in love with that girl?

Posted Jan 28, 2005 6:47:37 PM | link

Grax says:

It is up to the interviewer to choose the appropriate question. If we can discover what motivates animals simply by observing them, do you not think that we can discover what motivates human beings by asking them simple, pointed questions and observing them?

If someone does not know why he kills creatures in games, we can ask him, "Do you like receiving new spell lines?", or "Do you enjoy being able to travel to higher level areas?", or "Do you enjoy using spells to kill creatures in this game?", or "Do you like having your in-game actions be strongly rewarded by the in-game mechanisms in place?", etc.

If someone doesn't know why he fell in love with a girl, we can ask him, "Do you like intelligent women, brown eyes, big breasts, manic laughs,...?", etc. Of course, the question is silly since "love" is not well defined. If you come up with some word that is less shamelessly mystical, then you can start to get somewhere.

Did this have to be said?

Posted Jan 28, 2005 7:00:52 PM | link

Flatfingers says:

I've run into this kind of thing before. No model of personality can ever be perfect, so there's always some flaw that can used as a reason to condemn any model as worthless.

If your model has "too few" types, it's accused of being excessively vague to the point of being applicable to anyone (like astrology).

If your model has "too many" types, it's accused of being so specific that no one type can possibly apply to any real person.

If your model is based on a formalism (such as two or more dimensions of personality), it's accused of not corresponding sufficiently to things that motivate all humans.

If your model is based on types corresponding to an array of observerved behaviors (such as the "Big Five" model that Real Psychologists use, or Dr. John Oldham's 14-type model aligned with the disorders described in the DSM-III-R and DSM-IV), then it's accused of simply being arbitrary and failing to encompass some other types.

Basically there's no way to win with these folks. Maybe they don't like being "pigeonholed" and react badly to any suggestion that there's a model into which they can placed; maybe they're just so practical that they can't abide any theory at all; maybe they just enjoy sticking a knife into the creative work of others. Whatever the reason, for these folks, no model will ever be good enough; there'll always be some reason for trying to invalidate it.

My response is simple: I've done a reasonably wide survey of these kinds of things, and this Keirsey/Myers-Briggs model demonstrably works. It shows consistent ability to explain and predict human behavior in numerous contexts. In fact, working out this model has helped me not only to better understand my own family but my coworkers as well, and all my relationships have improved as a result of applying the concepts found in it.

Is it perfect? No. It's just proved to be good enough.

So here's my response to the critics: If you have constructive criticism to offer about these models, I welcome it. If you see no value in them, fine -- don't use them. But your inability to recognize value in them does not imply that they have no value and should therefore not even be proposed in public.

I will continue to describe to other people this model, and Bartle's types, and any other model that seems to have real-world utility, and trust people to decide for themselves whether there's enough utility in a model to try applying it.

I'd be pleased if game designers found something of virtue in these ideas, imperfect as they may be.

--Flatfingers

Posted Jan 28, 2005 7:05:31 PM | link

Jeff Freeman says:

How many do you need to exist "as a type in its own right"?

Enough to be an influence on the rest of the players. More than one per 100,000. :)

What theory is this, exactly?

From here: http://www.nickyee.com/facets/home.html

Primarily:

The important thing to remember is that these are not player types. It is not the case that we have found a way to categorize players into different boxes. Rather, we have found several distinct motivations that are meaningful and empirically tested constructs. Each individual player has a score for each factor, and it is after looking at the scores of all 5 factors that we can get a good idea of what part of an MMORPG appeals to them. Just because a player scores high on Achievement doesn't mean they will score low on all the other factors.

Whereas, the player-types supposes that the more one is an achiever, the less one is a socializer, etc.

>I much prefer designs based upon the idea that individuals have multiple motivations

Why? And what designs do this?

Because if people really are motivated by multiple things simultaneously, then an activity which only satisfies a single motivation is not going to be as appealing to them as an activity which satisfies multiple motivations.

Also, because it means that single activity can appeal to different people for different reasons.

SWG's creature handler is a profession which satisfies a variety of interests, for example. By comparison, other systems are less appealing, or appealing to fewer people.

I think we should have designed every profession and every activity with the idea that it should satisfy multiple motivations. I believe that would have given them all more depth.

In many cases the approach was "This is a social profession, it gives socializers something to do" and "That is an achiever profession, it gives achievers something to do."

Overall, the game has activities for all types of players, but in practice it doesn't much matter that there are other things a person could be doing (if they stop doing this), which would also appeal to them.

Posted Jan 28, 2005 7:13:04 PM | link

Hellinar says:

Grax> You say things that are obnoxiously false, such as: "Do players know what motivates them? If they don't then how can you figure it out by asking them?" Please take a moment to consider what you have just said here. <

I’m not clear what you are saying here. “Players know what motivates them” and “Players don’t know what motivates them” seem equally indefensible absolute statements in my experience. I believe a normal human being is aware of some of their motivations, and not of others.

Seems to me, if either Nick Yee’s factors, or Richard's player types have any basis in actual player motivation, they are useful in VW design. It is another template to put against your world and see how it measures up. For example, I tend to hear Achiever as “needs a trusted external scorekeeper to tell them if they are winning”. My world architecture is open source client, peer to peer server. With both the server and the client “in the hands of the enemy”, its not a VW design conducive to a Achiever playstyle. So its important to me that there are other playstyles and motivations that don’t require a trusted server to host the world. Obviously Socializer/ Relationship motivation works fine in a distributed server world. A client can provide a guaranteed level of Immersion for those who want it. I’m less sure about some of the other types or factors.

One playstyle dimension that isn’t captured in either the Bartle or Yee formulation is Producer/Consumer. Looking at Nick’s questionnaire, there seems to be an inherent assumption that the Player is a consumer of content, not a producer. In my view, a producer oriented game is easier to implement over a distributed server than consumer oriented game. So I’d be interested in what correlation this motivation has with the others mentioned.


Posted Jan 28, 2005 10:02:51 PM | link

Nick Yee says:

Jeff - Thanks for differentiating between Types and Scores. It's a subtle difference that makes a huge difference and a lot of times I'm not sure whether people notice the difference.

The Types model is a classification scheme and assumes that people primarily fit in one box.

The Scores model is an assessment scheme and assumes that everyone has a score for each motivation.

The Scores model allows more interesting combinations and reflections:

High Achievement + Low Socializing = Solo Optimizer
High Achievement + High Socializing = Raid Guild Member
High Achievement + High Exploration = Min-Maxer

The Types model suggests that your non-primary Types are of little interest.

The Scores model suggests that how you score on each of the motivations is EQUALLY revealing. High scores are just as interesting as low scores.

Posted Jan 28, 2005 10:52:01 PM | link

Flatfingers says:

Nick Yee> "The Types model suggests that your non-primary Types are of little interest."

Of little interest to whom?

We need to remember that we're talking about two different subjects for typing: the player, and the game feature. Many (not all, but many) people do seem to have one preferred temperament most of the time. That in no way implies that they can't appreciate some game features aimed at satisfying other temperaments. Nor does it imply that a temperament-based game feature that doesn't interest Player A won't interest Player B.

To put it another way: just because most players tend to express one specific type doesn't mean that game features can only be (or must be) designed to satisfy one specific type.

Suppose some game development house were to embrace the idea of temperament (a Type model) and design a mass-market MMOG whose features were explicitly conceived to satisfy temperament-based desires. What would prevent that developer from creating game features that could be satisfying to more than one temperament at a time?

I don't believe any such impediment exists. So where's the problem with using Type, since nothing whatsoever prevents developers from offering features that satisfy both individual and multiple temperament-based preferences, just as a Scores system could?

In fact, doesn't the simplicity of a Type system give it an advantage over the more complex Scores approach? (This follows the good design rule that systems should be no more complex than they have to be.)

I actually like the Scores approach. If someone finds it useful for keying game content to player desires, I think that's great.

If a Type approach can offer equal levels of player satisfaction with less complexity, why insist that a Scores model is somehow superior?

--Flatfingers

Posted Jan 29, 2005 12:30:07 AM | link

Nick Yee says:

Flatfingers wrote: In fact, doesn't the simplicity of a Type system give it an advantage over the more complex Scores approach?

The fundamental differences is whether you believe that people in general:

1) Fall into distinct and fairly discrete personality types that are dichotomous or bi-modal in distribution.

2) Fall into continuous spectrums that have a normal distribution.

Say we take intelligence. It may seem parsimonious to classify people as "smart" or "stupid" and only have 2 Types, and it may seem cumbersome to use a continuous IQ scale, but the problem is that IQ (along with most human attributes) tend to occur on a normal distribution and your 2 Types model in fact fails to truly understand the bulk of people who fall around the mean.

And once you realize that IQ isn't the only thing in a person you're interested in, then the problems with the Types model are compounded with each new attribute you add.

Posted Jan 29, 2005 3:14:40 AM | link

Abalieno says:

Jeff Freeman: SWG's creature handler is a profession which satisfies a variety of interests, for example. By comparison, other systems are less appealing, or appealing to fewer people.

I think we should have designed every profession and every activity with the idea that it should satisfy multiple motivations. I believe that would have given them all more depth.

Or maybe transform these relatively closed professions into different roles that can be impersonated with the same character. So that I could chose to have one role in the combat, one role in the social aspect, one role in the crafting and so on. Including and delivering all that the game has to offer.

As discussed on Grimwell months ago. Btw, I agree with the approach of what you write here.

Posted Jan 29, 2005 3:35:34 AM | link

Ola Fosheim Grøstad says:

Hellinar> But your inability to recognize value in them does not imply that they have no value and should therefore not even be proposed in public.

I don't think anyone has said that they shouldn't be proposed in public. And I hope that nobody will say that they shouldn't be dicussed and questioned in public either. :-)

However, humans love to stereotype and are very reluctant to give up stereotypes when they first have established them. This is well known.

So when stereotyping is applied to design or proposed to aid design then there are good reasons to take a closer look at them.

Hellinar> I will continue to describe to other people this model, and Bartle's types, and any other model that seems to have real-world utility, and trust people to decide for themselves whether there's enough utility in a model to try applying it.

That's fine if people using the models have access to a wide array of models or use it to get alternative views in an interpretive manner. If designers appropriate one model then they are in trouble, especially if it is a normative one. There is a reason for why designs are being play tested.

Posted Jan 29, 2005 6:15:01 AM | link

Ola Fosheim Grøstad says:

Ack, I misattributed my quotes to Hellinar, they were from Flatfinger.

Regarding categories or factors/scores: with 3 levels and 4 factors you get 81 categories. Trying to relate these to eachother is well... a challenge.

Posted Jan 29, 2005 7:55:51 AM | link

Richard Bartle says:

Me>What theory is this, exactly?
Jeff Freeman>From here: http://www.nickyee.com/facets/home.html

Read through that again and tell me why you think it's a theory.

It's not a theory. It's not even an hypothesis. It's an organised set of useful and significant observations from which a theory could be built, but as yet there's nothing holding it together. Where's the system? Where's the explanation? How can it be used predictively? It can't be.

>Whereas, the player-types supposes that the more one is an achiever, the less one is a socializer, etc.

That's right, it does. It describes a means by which players change from one to the other, too. The same can not be said of the Facets described by Nick Yee (hmm, I should maybe call him "Yee" as he always calls me "Bartle"). To be fair, he makes no such claims himself anyway.

>if people really are motivated by multiple things simultaneously, then an activity which only satisfies a single motivation is not going to be as appealing to them as an activity which satisfies multiple motivations.

Why not? I like chocolate and I like spaghetti bolognese, but I don't like them both together.

>Also, because it means that single activity can appeal to different people for different reasons.

This is something that's a good idea, yes, but it ap[plies whether or not people have one main motivation at any one time or several motivations simultaneously.

>I think we should have designed every profession and every activity with the idea that it should satisfy multiple motivations.

I don't think you need to do that for every profession, because some people like focus. However, it's a good idea to do it for many professions, yes. As I said, though, this is irrespective of whether people hold one motivation or several. I'd also point out that unless you can change your profession, you're stuck if your motivations change so there is no intersection with that of your profession. You might like everything that animal training has to offer now, but after 2 years of playing your probably won't (or if you do, you probably don't now).

>In many cases the approach was "This is a social profession, it gives socializers something to do" and "That is an achiever profession, it gives achievers something to do."

Well, that's not perhaps the best way to do it. People aren't socialisers or achievers all their playing career - they eventually morph into other types. My original model didn't explain how this happened, but my new one does have a stab at it.

Having a class "for achievers" is like having a class "for people in their 30s". People aren't always in their 30s, they're there for at most 10 years. What you want is a class that allows people to develop from the "in their 20s" to "in their 30s" to "in their 40s" without leaving the class. For virtual worlds, you don't want a class for achievers, you want a class that allows people to change from opportunist to scientist to planner to friend, or from griefer to networker to politician to friend, or for any of the other paths that players take. [I believe this is best done by dropping classes altogether, but that's another story].

If you know how people's interests will change, you can build structures that allow them to take their character with them as they do change. It's not enough to think that people have a "main" reason for doing something (as I do) or to think they are always juggling multiple, unrelated reasons (as you do); you need to know why they have such a reason or reasons if you're to make any use of the information at all. That's what a theory gives you that the raw data doesn't.

Richard

Posted Jan 29, 2005 10:00:02 AM | link

Richard Bartle says:

Me>I did say what kind of systems my model was limited to: MUDs.
Ola Fosheim Grøstad>Which doesn't really say much at all.

It did when the paper was written.

I really don't see what you're getting at here. I realise that you feel the need to challenge everything everyone ever says on Terra Nova or MUD-DEV while never offering any alternatives of your own, but really, do you have to be quite so griefing about it?

That paper was written 10 years ago. It refered to MUDs, a term well understood by its intended readership (it was first published in the Journal of MUD Research). Its preface described the kind of systems that counted as MUDs at the time. What would satisfy you as an answer? Should I have enumerated every single MUD I knew about and say "this applies to you, and any entities like these that your players may write"?

If you can't cope with the concept of a MUD as outlined in that paper, why are you even posting here?

>I certainly need to know more about the case if it is a study of a case. Otherwise you basically ask me to view it as non-empirical.

I am asking you to view it as non-empirical. I'm saying that I have experience, I've made some informal observations, I'm putting forward a theory, you're welcome to use it or trash it as you see fit. Enough people have looked at it and found it to be useful that the chances are there's more than a grain of truth to it. It's not perfect, but at least it got designers to think about what they were designing.

By asking about MUD1 and MUD2, you're asking Shrodinger details about his cat. The cat is irrelevant.

>Right now it's more like studying a particular vehicle, present a theory that claims that it holds for transportation in general while at the same time refuse to tell whether you've studied a bike, a car, a snow scooter, a plane...

I think this is a misleading analogy, but even if it were true, so what? Does it matter what vehicle I studied if the theory I developed has indeed been found to have merit in discussing transportation in general?

I made no attempt to "prove" my theory, I merely stated it. If you don't want to use an unproven theory, don't. If you want to disprove it, go ahead. If you find it's flawed but still useful, develop a more refined version. You don't need to know about MUD1 or MUD2 to do any of this - any virtual world will do.

>I don't understand the relationship between the dimensions and the categories and where they came from.

Knowing where the categories came from is not a precondition to understanding what they are, although it could help in the explanation. As for the relationship between the dimensions, there is none: they're orthogonal. As a set, they produce a means for describing player types that I believe has utility.

>Uhm. A model doesn't exist without a >description. So it is clearly not there?

A model can exist without a description if it has a prescription to derive the description. In my book, I do describe my model, so it does have a description, but the underlying dynamics I don't describe; rather, I describe how to dervice them.

Of course, if I had described them you'd just be complaining about how there were too many of them to be of use, how confusing they were, how unlikely or irrelevant some of them were...

>Well, but your model doesn't really say anything about the reasons.

You read my book, yes? Including the whole thing about the hero's journey and how my model maps into it? Yet you blithely assert that my model doesn't say anything about the reasons that people play virtual worlds?

Nothing is going to satisfy you, whatever I say.

>But haven't we discussed this a length on MUD-Dev time and time again and reached the conclusions that there are no clear typical paths for playstyles for the general case?

No, we haven't. No discussions in which you are involved ever have any conclusions.

>Yes, but then it would be better for designers to concoct their own models based on how they expect their particular design will be appropriated.

I'd be happy to see anyone concoct their own models, so long as it led to players having a better time.

>Otherwise your model will be normative (in relation to design), are you prepared to take that responsibility? :-) I.e. take the responsibility for a design that tanks.

I'm responsible for the model, not for the use of the model.

>Not really, but it is less useful as a model as you cannot possibly believe it holds? Clearly it depends on the particular player, ingame culture and design?

I do believe my theory holds to a degree that is useful to designers. If you're saying that designers need to understand individual players and the exact in-game culture as it is or will be so that they can design a better game, well yes, that would mean they could design a better game, but they're never going to get that information are they?

Richard

Posted Jan 29, 2005 10:48:44 AM | link

Ola Fosheim Grøstad says:

Actually, I haven't tried to be "griefing" about it. I am sorry if you feel that way, that was certainly not my intent. But then, you seem pretty good at retaliating so maybe I shouldn't worry too much about it...

I've always seen it as an empirical, and simply think it would be useful to know the characteristics of that system from which it was derived. To create a theory for all muds is pretty ambitious, and if it holds for a subset then why not go for it? But then one will need to take the interaction between design and playstyles into account, which the 4-types model does mention.

Regarding the 8-types model. You asked why it isn't mentioned. I simply stated that I found it more difficult to understand than the 4-types model. In particular the reasoning behind the placement of the various categories. I am not saying that you are wrong, I say that I don't understand it as well as the 4-types model.

Posted Jan 29, 2005 11:33:06 AM | link

Abalieno says:

> Richard Bartle:
I don't think you need to do that for every profession, because some people like focus.

So? Peoples like to choose their focus if they feel so. Having possibilities is never felt as something negative, having those possibilities removed is.

The point is about offering those possibilities because the game should offer to the player ALL its resources. It's about taking advantage of your FULL potential instead of slices of it.

Then it's the player to choose where to go, what to do and what will be its focus.

Posted Jan 29, 2005 5:01:30 PM | link

Flatfingers says:

Nick Yee> "1) ... distinct and fairly discrete personality types that are dichotomous or bi-modal in distribution.
2) ... continuous spectrums that have a normal distribution."

I guess I'm in trouble, then, because I've come to see types as both of these things.

That is, any one type in any personality model doesn't capture the totality of any individual's personality -- people display behaviors within a continuum of behavioral preferences. At the same time, however, most people do seem to have one general style that they prefer over other styles most of the time.

Accordingly, I see personality as a continuum with N local maxima. Yes, personality is a spectrum, but there are peaks in that spectrum that we can think of as "types" with uniquely identifiable traits.

So a Scores approach isn't wrong if it's constructed on empirical data to broadly capture behavioral styles. (In fact, that's a good description of how the Myers-Briggs model works, which combines four roughly independent dimensions of personality based on self-reported behavioral styles.) At the same time, a Types approach isn't wrong as a convenient shorthand to say, "a lot of people have one of these N styles much of the time."

I conclude that there's nothing inherently "better" or "worse" about either a Scores or a Types approach, as both -- so long as each is based on good empirical data -- are capable of saying something useful about human behavior. I further suspect that the real difference between these two approaches is a tradeoff of completeness for simplicity: Scores tell you more about the whole person but are more complex than, say, a four-temperament Types model, while Types are easy to use but don't describe an individual's preferred behavioral style as completely.

If that's the case, then the decision of whether to use a Scores or a Types model ought to be based on the application. If you need to predict behavior to a fairly high degree of accuracy, choose a Scores model; if you just need a system that's "good enough" but simple enough to be easily implemented, choose a Types model.

Ola> "when stereotyping is applied to design or proposed to aid design then there are good reasons to take a closer look at them."

I'd say that any theory deserves scrutiny. But there's a real difference between pecking at the work of others just to find cracks, and offering constructive observations intended to help the model-builder improve his model.

As for stereotyping: it's usually a mistake to predict an individual's beliefs or behaviors based on a perception that that individual is a member of some (stereotypical) group. But that doesn't mean that any particular stereotype is wrong in its description of a group of people as a whole, who may in fact share particular behaviors most of the time.

The fault lies in assuming that every individual within a group expresses the characteristics of the average group member. In that sense, there's definitely some amount of error built into any model of personality types.

But the alternative -- treating every individual as a unique individual -- is not possible in a system intended to service large numbers of people (such as a MMOG). There aren't enough developers to create specialized content for every player; some "chunking" system is necessary to achieve economies of scale.

Ergo, we need a typing system of some kind to try to understand players as members of group. Will any such system be flawed because it treats individuals as average members of some group? Yes. Does that utterly invalidate the utility of such models?

I don't think so... not as long as people actually do behave in similar ways. You're not required to believe that they do, of course, but nothing you've said so far leads me to believe that they don't.

--Flatfingers

Posted Jan 29, 2005 5:03:52 PM | link

Flatfingers says:

Richard Bartle> "People aren't socialisers or achievers all their playing career - they eventually morph into other types."

If this is true, it invalidates my perception that the player styles you describe are context-specific subsets of the four general behavior styles described by David Keirsey... which typically *don't* change over time.

> "My original model didn't explain how this happened, but my new one does have a stab at it."

I definitely need to have to have a look at this new model, don't I? Meanwhile, I'd be very interested to know your reaction to the "congruence" notion I've proposed here as it stands currently.

--Flatfingers

Posted Jan 29, 2005 5:10:47 PM | link

Ola Fosheim Grøstad says:

Flatfingers> we need a typing system of some kind to try to understand players as members of group.

((Warning: I am getting totally off-topic here.))

In relation to design? We need to consider effect of the system when it is in use. That's true. Having a generic model might be one way of approaching this challenge. Involving representative users might be another way of doing it. Analyzing similar systems with a similar demographic would be yet another approach.

I am not saying that one should avoid stereotypes altogether. I am saying that designers should not get stuck on one particular "true" model. If you can get new perspectives by moving between models and look at a design proposal from different perspectives, great. But then you are most likely using the model in an interpretative manner... I would think? Structuring your design around a model or deducing a design from a model is something else entirely.

User modelling is a tricky field, and I think most of the issues that field experience would be relevant for those approaches to design you allude to. Predictive models in relation to design is and will always be tricky. That's why the current "ideal" software design paradigme is user-centered and iterative development (or evolutionary). If it is desirable to challenge that paradigme, then that is ok too, of course.

Most games are similar for a reason, I think that reason is the risk related to how users react to a new design? Would you say that using generic "psychological models" would mitigate that risk?

Posted Jan 29, 2005 6:18:00 PM | link

Abalieno says:

"Models" to design directly a game? Beware of this.

Those models are surely useful but *after*. Those models are useful to understand and reveal important elements but they shouldn't be design structures to use directly. The infinite struggle to create better and more inclusive models is just an exercize of the mind to control and master everything but it won't bring anywhere but an illusion.

Wittgenstein described this in the best way possible, no matter of the context or the subject. They are like ladders that can help you to climb up. But when you are at the top you take the ladder and throw it away.

When we were discussing exactly this point Raph Koster added:

I don't know anyone who only designs to fit the playstyles. Usually, people make the game they want to make, then afterwards start thinking about questions like "what about people who just want to chat?"

Do not put the cart ahead of the horses.

Posted Jan 29, 2005 8:04:38 PM | link

Jeff Freeman says:

It's not a theory. It's not even an hypothesis. It's an organised set of useful and significant observations from which a theory could be built, but as yet there's nothing holding it together. Where's the system? Where's the explanation? How can it be used predictively? It can't be.

Fair enough.

>Whereas, the player-types supposes that the more one is an achiever, the less one is a socializer, etc.

That's right, it does.

Don't you feel that is contradicted by the observations Yee (heh) has organized? Or is the point that it doesn't matter?

Posted Jan 30, 2005 4:22:51 AM | link

Richard Bartle says:

>Actually, I haven't tried to be "griefing" about it. I am sorry if you feel that way, that was certainly not my intent.

Oh, I didn't think it was your intent; I didn't even think it was griefing. I used the word so you'd maybe pause and think about what you're trying to get out of this discussion.

>I've always seen it as an empirical, and simply think it would be useful to know the characteristics of that system from which it was derived.

Well you can play MUD1 and MUD2 right now, because they both still run. You can't play MUD2 as it was in 1990, though. We didn't have an anthropologist looking at it, and we don't have an historian looking at it. It may be useful to know how MUD2 was back then, just as it may be useful to know the background of the various people who took part in the debate, or the background of the UK telecommunications industry at the time. These things are all useful to know if you want a clear understanding of the context in which the theory was developed. They are, however, irrelevant to the theory itself. I'm saving you from having to look at all this background data by presenting the theory without it. It stands on its own merits. You don't have to look in the past and decide "he would think that because he was dealing with these kinds of people". You just have to look at existing virtual worlds and decide "does this theory apply here or doesn't it?".

>To create a theory for all muds is pretty ambitious, and if it holds for a subset then why not go for it?

Because it is a theory for all MUDs. I do make some qualifications (in the section on balancing, for example, I show how something that starts off as a MUD can become something else if pushed too far); I consider these as being supportive of the theory, though, rather than undermining of it.

>But then one will need to take the interaction between design and playstyles into account, which the 4-types model does mention.

The 8-types model is a refinement of the 4-types model. The dynamics of the 4-type model still apply. It may be that by delving deeply into the interactions between the 8 types some new and interesting dynamics are unearthed, so it probably is worthwhile doing this; if I thought I could get an enumeration of all 64 permutations published, I might even do it myself. The current lack of such a dynamics for the 8-types model doesn't render it invalid, though.

>I simply stated that I found it more difficult to understand than the 4-types model. In particular the reasoning behind the placement of the various categories.

I'm having problems figuring out what you're asking, here. What do you mean by "placement of the various categories"?

Richard

Posted Jan 30, 2005 6:12:43 AM | link

Ola Fosheim Grøstad says:

Abalieno> Those models are surely useful but *after*. Those models are useful to understand and reveal important elements but they shouldn't be design structures to use directly.

That's perhaps a bit strong claim, but such models are clearly political statements and one will have to accept that they are treated as such.

If one presents Meyer-Briggs as a perspective on personality then that is one thing. If the model is either used in decision making processes (design) or being touted as useful for such then it can no longer be seen as an innocent bystander, it is now a political agent arguing a particular design perspective. One cannot then demand that users or others who have concerns about that perspective, dislike the pigeon-holing or the resulting design should either improve the model or shut up.

Posted Jan 30, 2005 6:21:26 AM | link

Richard Bartle says:

Jeff Freeman>Don't you feel that is contradicted by the observations Yee (heh) has organized? Or is the point that it doesn't matter?

Oh, it does matter!

Nick's results give 5 facets. Three of these map reasonably uncontroversially to my system (Relationship, Grief and Achievement). One of the other two, Immersion, is explained by my 8-types model (basically, as a kind of meta-motivation). The final facet, Leadership, can in its pure form also be explained by my 8-types model (it maps to the politician type), but it seems a little broader than that as Nick has it in that it overlaps with some of the other categories: you can have Leadership as an expression of the Achievement, Griefing or Relationship facets. Interestingly, you can't have it as an expression of the Immersion facet, a fact that highlights how Immersion is somehow different from the other facets.

The main issue concerning facets is not that the ones discovered don't fit into my player types system (because they do); rather, it's that my player types system predicts that there is an explorer type, yet Nick didn't find any evidence for this even when he re-engineered his survey to look for it.

Does this contradict my theory? No (because you can't prove non-existence). It does, however, give my theory a problem: why didn't Nick find any explorers?

There are many possible explanations, but the most likely ones are:
1) My theory is wrong.
2) Explorers don't play the kind of virtual worlds Nick surveyed.
3) Explorers do not answer questionnaires (or they do answer but subvert them).
4) Explorers never exist in statistically significant numbers.
5) Nick's surveys are invalid.

My feeling is that 2) is correct. In the same way that you won't find many achievers in a MOO, there may be something about EQ-like virtual worlds that serves to disguise or eliminate explorers. What's more, I have some ideas as to why this may be. I'll give these now, but bear in mind that this is all fairly speculative.

In my 8-types model, there are two types of explorer. One type, the scientist, arises fairly early in the player experience as individuals attempt to learn more about their environment through exploration and experimentation. The other type, the hacker, comes much later as particular individuals internalise the workings of the virtual world so much that they attain a guru-like understanding of it.

In order to get become a hacker-type explorer, the player basically has to give up achieving (this is a hero's journey thing, so feel free to disagree). Some people will do this of their own volition, but most people will only do it by "winning" the virtual world and being released from the treadmill. If a virtual world has a long treadmill and no winning condition, then this will greatly reduce the number of players who become hackers through their playing experience - almost entirely eliminating them, in fact. The only hackers around are those who self-actualised as hackers without following the hero's journey to its completion (which I have actually seen happen).

As for the other type of explorer, the scientist, again EQ-like games conspire against them. When players need to acquire knowledge, they can do so either by finding it out themselves or by finding out from others. In a virtual world that promotes grouping, it's very hard to be a lone explorer - you have to be with a group. When you're with a group, you get your game world knowledge from the other members of the group whether you like it or not. [Aside: some players do like it, and these would be categorised as networkers in my 8-type system. They'll look at web sites for information if no-one around knows it]. Thus, players who want to explore through experimentation have less opportunity to do so than they would in a world where they could extensively play solo.

It may be wrong, but that's how I currently reconcile my theory with Nick's data.

So does it matter that Nick's data shows a problem with my theory? Yes, it does, although not because I'm precious about my work. It's important because it raises a point that requires an explanation we don't currently have, therefore if we can explain it we'll have advanced our knowledge of why players play virtual worlds. I don't particularly care whether my theory is the one that explains it or whether it's better explained by some yet-to-be-formulated theory. All I want is for there to be a coherent, useful explanation. Then, we can look forward to better designs that employ this new theory, all the while looking for explanations for the new flaws that will surely be present in it.

Richard

Posted Jan 30, 2005 7:12:20 AM | link

Richard Bartle says:

Me> "People aren't socialisers or achievers all their playing career - they eventually morph into other types."
Flatfingers>If this is true, it invalidates my perception that the player styles you describe are context-specific subsets of the four general behavior styles described by David Keirsey... which typically *don't* change over time.

I don't think it does. People do change over time - are your grandparents the same people they were as teenagers? It's just that in virtual worlds they get to change more quickly. By "change" here, I don't mean their personality changes; rather, their understanding of their personality changes. Few people who are introverts are going to become extrovert through playing a virtual world; however, by passing through a phase of doing what extroverts do, they come to understand their own introversion more. Whether they regard this as a flaw to be corrected or a feature to be celebrated depends on the player's underlying "real" identity. [Note: post-modernists won't like the suggestion that there's such a thing as a "real" identity at the core of an individual, rather than just a construction of disparate reflections. Well tra-la-la.]

>I definitely need to have to have a look at this new model, don't I?

If it means I sell another copy of my book, yes!

>Meanwhile, I'd be very interested to know your reaction to the "congruence" notion I've proposed here as it stands currently.

I think you're on to something. I just wish I could find that MUD-DEV posting from last year that was along the same lines (and drew in the Medieval notion of the 4 humours, too, I seem to recall).

It's pleasing to find that there may be some deep-rooted psychological justification for my player types model, if only because it adds credence. However, the proof of the pudding is in the eating: it doesn't matter how much evidence there is to suggest that the model is correct if the model itself isn't fit for its intended purpose.

That said, if it is trusted as a model then the underlying psychology can perhaps be used to make further predictions beyond those that the model itself suggests. This may help designers (and others, eg. marketers) get an even better handle on what the players want. Thus, although the linking of the player types model into a wider theory of identity doesn't help validate the player types model, it does open up new avenues for those researchers seeking deeper or more appropriate models to explain or understand player behaviour.

Richard

Posted Jan 30, 2005 7:35:36 AM | link

Ola Fosheim Grøstad says:

Richard Bartle> You just have to look at existing virtual worlds and decide "does this theory apply here or doesn't it?".

Yes, that's sufficient for designers. For thinkers who want to build on it more transparency is always helpful. I believe that all models provide perspectives of which some are explicit and some are implicit. To get at the implicit ones you might need to understand how it was conceived and from what it was derived. It's not a requirement of course, but helpful. *shrugs*

Richard Bartle> Because it is a theory for all MUDs.

Yes, but as it appears to be primarly embraced by "RPG-mechanics-world people" maybe the 4-type model best fit systems dominated by certain designs. Of course, one will have to follow one's own goals, but I think one could say a lot more about dynamics, and with greater certainty, by narrowing down the scope.

Regarding "explorers", the criteria for including them in a model with a focus on dynamics and design should be the effect of their presence, not how many they are. With the WWW even a few publishing explorers can have a tremendous effect on the system. Of course, designers will ask themselves questions such as: "what happens to the entertainment factor when our players start to map out our game?" and "do we have to write help-docs for this or can we let the players do that themselves?".

Richard Bartle> In a virtual world that promotes grouping, it's very hard to be a lone explorer - you have to be with a group.

There has been plenty of explorers in AO. Of course, players will alternate between achievement and exploration. Achievement is a great way to validate your exploration efforts.

(I'll have to postpone the comments on the 8-type model, it is too complicated to get into here.)

Posted Jan 30, 2005 8:39:51 AM | link

Tess says:

In my experience from my MUSH-wizarding days, the number crunchers were always unabashed powergamers... er, I mean, uh, "achiever-types." Explorers are something entirely different. I see Explorers as people who are chasing a high called "Wonder."

What on earth do I mean by that?

An explorer is the person who plays a completely under-developed class that's no good in combat, because it gets an invisibility spell at a really low level, and then trucks off into super-dangerous wilderness she's not ready for, and then completely falls out of her seat in awe when she sees something amazing and special in some faraway place, and then chases that high for the rest of her gameplaying lifetime. She is forever looking for the next amazing sunrise, the face carved in the wall of a cliff, the trees taller than God, the first glimpse of a titan she has never seen, the rumored seamonster, and the enemy cities where she should not be.

When she peers over a hillock with her friends, and suddenly sees the baby monsters eating each other and growing bigger, all her hair stands on end! She has seen wonder, and it will sustain her.

For a little while.

Posted Jan 30, 2005 9:35:24 AM | link

Jeff Freeman says:

The main issue concerning facets is not that the ones discovered don't fit into my player types system (because they do); rather, it's that my player types system predicts that there is an explorer type, yet Nick didn't find any evidence for this even when he re-engineered his survey to look for it.

Isn't there also an issue, though, with Nick Yee's observations contradicting the idea that (for example) the more one is interested in Socialization, the less one is interested in Achievement, etc.?

As I understand it, the types-model predicts that would be the case, but Nick Yee's data doesn't support it.

Regarding the lack of explorers, Nick Yee was unable to isolate 'to explore' as a motivation distinct from all others, but while that could mean that no one is interested in exploration in the games surveyed, couldn't it also mean that everyone is?

I doubt you could isolate 'to have fun' as a motivation for the same reason, but I wouldn't take that to mean that people in EQ-type games don't like to have fun.

Posted Jan 30, 2005 2:20:30 PM | link

Hellinar says:

Tess> She has seen wonder, and it will sustain her.

For a little while. <

What a beautiful post. Sounds like a much more poetic version of my conception of an Explorer. My first EQ character was a level 5 Enchanter who used the invisibility spell to stick her nose into pretty much every zone.

Richard> In order to get become a hacker-type explorer, the player basically has to give up achieving (this is a hero's journey thing, so feel free to disagree). Some people will do this of their own volition, but most people will only do it by "winning" the virtual world and being released from the treadmill. If a virtual world has a long treadmill and no winning condition, then this will greatly reduce the number of players who become hackers through their playing experience - almost entirely eliminating them, in fact. The only hackers around are those who self-actualised as hackers without following the hero's journey to its completion (which I have actually seen happen). <

In my experience, the hero’s journey is not difficult to complete in most MMOGs. The only requirement is that the Player sets there own goal for completion, rather than following the annoying experience grind that the designer set. And it that self-actualised goal setting that I see as the major mark of the Explorer, as opposed to the Achiever, Socializer and Killer who all require an external validation.

As examples, my gnome wizard in EQ had his heart set on making a wondrous device called a compass. Which involved journeys through dangerous lands, across the seas by ship, gathering materials, and negotiations with skilled tradesmen for parts. A few weeks play at my casual playstyle, and a very satisfying character story. In WoW, my Tauren hunter was determined to hunt all the creatures of The Barrens, even the most dangerous. A life story that was helped by a Blizzard built quest series along just those lines. He was a hunter of the open plains, and had no time for the forests and ash pits that framed the lives of more powerful characters. Again, that journey took a few weeks at my casual playstyle, with other characters intertwined.

I’d count myself an Explorer, and note that my typical play history with a “kill MOBs and get a better sword” type MMOG is to explore for a couple of months then leave. So its quite plausible to me that Nicks demographic is the long term players, who are likely not Explorers. “How long have you been playing your current MMOG” might be a useful question to add to the survey to explore this angle.


Posted Jan 30, 2005 2:25:12 PM | link

Nick Yee says:

Richard wrote: Does this contradict my theory? No (because you can't prove non-existence). It does, however, give my theory a problem: why didn't Nick find any explorers?

I do agree with Richard's observation that this may be an artifact of current MMORPG design, but I also feel it may be because I haven't asked the right questions. And I'm trying in the current survey to isolate exploration motives that provide no real functional value, and also looking at "knowledge seeking / discovery" rather than "mapping / analyzing".

Jeff wrote: Isn't there also an issue, though, with Nick Yee's observations contradicting the idea that (for example) the more one is interested in Socialization, the less one is interested in Achievement, etc.?

Yes - this hits at an important underlying distinction between the two models.

Richard's model seems to imply that you are primarily driven by one Type, and that your scores on the 4 Types add up to 100% in some way - in that the more of an Achiever you are, the less of the other Types you are. Richard - do you agree with this?

Whereas a factor model doesn't make this assumption. Just because you like ice cream doesn't mean you don't like spinach. And it's possible you love both (or hate both) at the same time.

If Richard's model were true, then we'd see negative inter-correlations among the factors (because high Achievement -> low other Types). This is not the case. Most factors are in fact mildy to moderately positively-correlated, which suggests that motivations do not suppress each other.

In fact, I would argue that this is the main theoretical difference between the two models - not in the the number or labeling of motivations (which is a more superficial difference), but how we propose the motivations of individuals should be understood or conceptualized.

Flatfingers wrote: Accordingly, I see personality as a continuum with N local maxima. Yes, personality is a spectrum, but there are peaks in that spectrum that we can think of as "types" with uniquely identifiable traits.

The MBTI creates the appearance of "peaks" because it uses forced-dichotomy questions. There is a strange tautology going on. If you assume bi-modal distributions and ask forced-choice questions, then the data will appear bi-modal. The problem is that most human attributes fall on a normal distribution. The peaks fall on the mean, not on the sides.

The MBTI assumes that in the Extraversion-Introversion scale that there are 2 peaks. All personality tests that test for that trait using a full scale have found that it lies on a normal distribution. There is just one peak on the mean. The illusion of Types is an artifact of a forced-dichotomy measurement tool.

Posted Jan 30, 2005 5:57:08 PM | link

Dave Rickey says:

I've got a thought kicking around in my head that someday will probably get turned into an essay, but I wanted to put it forward here for people to kick around: In A Theory of Fun, Raph proposed, in essence, that "fun" was the result of a fundamental brain-reward mechanism that evolution has developed to goad us into learning. Now, I think that he's fundamentally right, but the particular mechanism he proposes is incomplete, and applies to only one of several brain-reward mechanisms that we interpret as "fun", specifically "neophilia", the urge to discover new things and be exposed to new stimuli.

If I'm right, then there are other brain reward mechanisms that account for other "play" activities not accounted for by Raph's theory. And this should map well to Nick's surveys, and at a minimum could be proved or disproved by his methods.

--Dave

Posted Jan 30, 2005 6:15:38 PM | link

Grax says:

I like to take a more behavioral* view of Bartle's 4 playertypes. As far as the strict system is concerned, I consider each player's playertype to be decided by actions rather than motivations. So a player who is motivated one way, but not sufficiently rewarded or compelled to act upon this motivation (in a given Virtual World), will not be labelled according to this motivation.

An example, which I've mentioned earlier, would be a player who has strong Killer motivations, but who exists in a VW that, for whatever reason, influences him in such a way that he seemingly does not act in accordance with whatever primary motivations he may have 'in theory'. I believe it is true that we are all killers, socializers, achievers, explorers (lowercase) to a greater or lesser degree**, but that certain in-game encouragements, or peer pressures, etc. can convince us to stifle some of our primary motivations in favor of other motivations that are more 'useful' in a given situation.

How useful is an action-based, rather than motivation-based, playertype categorizing system? Actually, I don't know exactly how useful any of these systems are. I think that systems are a useful way of making broad (and broadly true) statements, but I also think that once one has played or designed enough MMOGs, and observed player behavior and motivations, and thought about all of this together, one doesn't really require a system to think about what is actually happening.

Gonna cut myself short here, as I haven't said anything especially noteworthy in this post, though I suspect it will come.

*The player motivations are obviously very useful and important, but they are perhaps not the best way of categorizing people; a system based on viewable actions would be a better 'base system'
**some may prefer killing killers, but they are still killers, and some may be Gandhi-esque, but they will kill in certain circumstances...

Posted Jan 30, 2005 6:55:38 PM | link

Richard Bartle says:

Nick Yee>Richard's model seems to imply that you are primarily driven by one Type, and that your scores on the 4 Types add up to 100% in some way - in that the more of an Achiever you are, the less of the other Types you are. Richard - do you agree with this?

Yes, but there's more to it than that (in a way which probably makes it easier to disprove my theory, but what they hey!).

As it stands, my theory implies that at any moment players will be somewhere along a journey of personal discovery that maps fairly well onto Campbell's Hero's Journey. There are players who aren't on this path, and there are players who were on it but who have become derailed, but on the whole players do follow it.

In terms of type, this means that players start as one of two types, then move on to one of another two, then to another two, then they wind up as one of the final two. At each stage, they may still have some leanings to the type they came from, or may have some leanings towards the type they're going to (or possibly two, if it's not clear to them which path to take).

Thus, if you have a "classic" achiever, you might see them still doing some exploring or networking, but it will gradually diminish, later to be replaced by more friend-like or hacker-like aims. From what I've observed, it's possible to predict reasonably well which of the two directions they're more likely to go if you know which path they've taken so far, but they can switch between paths so it's not guaranteed. Also, it's just my personal observation: the theory doesn't actually have any need a list of common sequences of player types.

So yes, I would expect to see some correlation between player types, but very specific ones. I would not, for example, expect to see any correlation between hacker types and griefer types: hackers may occasionally look to others like they're griefing (indeed, in real life the word "hacker" has been co-opted by the general public in this way; call them "gurus" if you find this confusing), but griefing is not their goal. I'd expect to see a higher correlation between scientists and planners, because lots of people move on to planning (achieving) once they have enough knowledge to use (from exploring), yet if they find new areas they want to explore they may find the urge to do so remains.

Richard

Posted Jan 31, 2005 3:32:57 AM | link

Richard Bartle says:

Don't like my player typology? The following appeared on the Digiplay mailing list 3 days ago (thanks to Aaron Delwiche for the heads up):

We are putting the finishing touches to a book entitled '21st Century Game Design', due to be published by Charles River Media later this
year. The book is an all-ability-level introduction to what we have termed demographic game design (which can also be considered to be market-oriented game design).

In brief, the first five chapters discuss audience models for video games, and the mechanisms by which awareness of games propogate (market vectors). More specifically, there are detailed chapters discussing the DGD1 audience model (which some of you will be aware of from the brochure we published last year). This model uses Myers-Briggs typology as its basis, plus detailed surveys and case studies, to identify four play styles - Type 1 Conqueror, Type 2 Manager, Type 3 Wanderer and Type 4 Participant. These play styles are discussed in great detail.

In other news: I've had a great idea for an object I'm calling a "wheel".

Richard

Posted Jan 31, 2005 10:37:47 AM | link

Thabor says:


From my personally standpoint I don't consider any of the styles exclusive to each other. I'm typically rated as a EASK or an AESK interchangably. Really what that should probably map to is on a scale of one to 10 about 8 Explorer 8 Achiever 4 Social and 2 Killer.

From my standpoint the trait which seperates Explorer behavior from Achiever behavior is throwing away the map. Both types might be interested in discovering underlying rules but the pure Explorer would have a preference for doing it without spoilers or hints. Whereas an Achiver would be content to use spoilers in reaching their goal.

The reality is though its all the categories overlap, and depend on each other to a greater or lesser degree.

Posted Jan 31, 2005 1:01:43 PM | link

Bart Stewart says:

Michael,

Your suggestion for a correspondence between Keirsey's more detailed model and Richard's more detailed model that equates Implicit with Introversion and Explicit with Extraversion (in the What Do You Want To Know thread) was pretty clever.

But I have to admit that after thinking about it and banging it around in my head for a while, it just doesn't seem to hold together. For one thing, that's not how Richard defines "Implicit" and "Explicit." As he puts it: "Explicit wizzes favored the known over the unknown, realism over idealism, the overt over the covert, expectation over surprise, and order over disorder." (Designing Virtual Worlds, p. 166.) That description sounds to me very much like the vertical axis in my variation on Keirsey's model.

I agree with Keirsey that the primary differentiator in people is their preference for iNtuition or Sensing -- the view that it's what's inside that matters versus the conviction that external realities are what really matter. We both use this preference as a horizontal axis on our 2x2 matrix diagram.

And I further agree with Keirsey that the next most important differentiator depends on the first one. If your preference is for Sensing, then the next most important preference is for judging versus perceiving; if your primary preference is for iNuiting, then the next most important preference is for thinking versus feeling.

But how do you represent that as an axis? It doesn't correspond to any single Myers-Briggs factor.

So we pick something that splits the difference... but that's where I disagree mildly with Keirsey's more detailed model. Where Keirsey for his vertical axis uses a "cooperative vs. utilitarian" concept, leading him to draw up his 2x2 matrix with Idealists opposite Artisans and Rationals opposite Guardians, I'm not convinced that this is the next most important axis of differentiation. To my mind, after concluding how to view reality (iNtuition vs. Sensing), the next most important aspect of personality is a person's need for freedom vs. stability, for change vs. structure... which I think maps very nicely indeed onto Richard's definition of "implicit vs. explicit." (And I should mention that I had not heard of Richard's initial player types model, much less his enhanced model, when I developed my own.)

Working out this notion of change vs. structure gave me a vertical axis that corresponds to a kind of merger of two (out of four) Myers-Briggs factors: the preference for change corresponds to xxFP, while the preference for structure corresponds to xxTJ. (Interestingly, the TJ combination is nearly distinct enough to be a type in itself, and shows up in most Myers-Briggs-related works -- Jean Kummerow's description in WorkTypes of TJ managers as able to "organize and structure the work to achieve goals" is typical.)

I think this FP/TJ distinction maps onto Richard's extended model as his Implicit/Explicit axis. The problem is figuring out how to apply this axis.

My solution is "it's a tertiary preference that, like the secondary preference, is determined by the primary preference."

When the primary preference is for Acting, the secondary preference of Player vs. World corresponds to Perceiving vs. Judging and the tertiary preference of Implicit vs. Explicit corresponds to Feeling vs. Thinking. When the primary preference is for Interacting, the secondary preference of Player vs. World corresponds to Feeling vs. Thinking and the tertiary preference of Implicit vs. Explicit corresponds to Perceiving vs. Judging.

In other words:

SECONDARY PREFERENCES:
Acting on Players (Killer) = Sensing + Perceiving (SP Artisan)
Acting on World (Achiever) = Sensing + Judging (SJ Guardian)
Interacting with World (Explorer) = iNtuiting + Thinking (NT Rational)
Interacting with Players (Socializer) = iNtuiting + Feeling (NF Idealist)

TERTIARY PREFERENCES:
Explicit Action = Sensing + Thinking (J/P is defined by secondary preference)
Implicit Action = Sensing + Feeling (J/P is defined by secondary preference)
Explicit iNteraction = iNtuiting + Judging (T/F is defined by secondary preference)
Implicit iNteraction = iNtuiting + Perceiving (T/F is defined by secondary preference)

So, rebuilding the list of correspondences according to this model, we get:

Politician (Killer; APe) => Promoter (eStP); Crafter (iStP) [Operators]
Griefer (Killer; APi) => Performer (eSfP); Composer (iSfP) [Entertainers]
Planner (Achiever; AWe) => Supervisor (eStJ); Inspector (iStJ) [Administrators]
Opportunist (Achiever; AWi) => Provider (eSfJ); Protector (iSfJ) [Conservators]
Scientist (Explorer; NWe) => Fieldmarshal (eNTj); Mastermind (iNTj) [Coordinators]
Hacker (Explorer; NWi) => Inventor (eNTp); Architect (iNTp) [Engineers]
Networker (Socializer; NPe) => Teacher (eNFj); Counselor (iNFj) [Mentors]
Friend (Socializer; NPi) => Champion (eNFp); Healer (iNFp) [Advocates]

Basically this is saying that Richard's eight subtypes actually do map onto Keirsey's eight subtemperaments.

The Politician type of Killer (I really would prefer to use "Manipulator" instead of "Killer"!) corresponds to the Thinking Artisan who consciously manipulates people toward some specific goal, while the Griefer corresponds to the Feeling Artisan who manipulates others for the emotional rush of the experience.

The Planner corresponds to the Thinking Guardian (there's that TJ subtype!) who deliberately and carefully organizes action in the virtual world toward achieving a specific goal, while the Opportunist corresponds to the Feeling Guardian who deals with things as they come up (usually not so much trying to make gains as to prevent losses). [Note: I'd place more value on the validity of this correspondence if there were evidence that Opportunists were the type most likely to spend time insuring the security (esp. physical and financial) of others in the virtual world.]

The Scientist corresponds beautifully to the Mastermind, and in an extraverted form can be considered to experiment with people by creating and leading purposeful organizations (as the Fieldmarshal must).

And Richard's description of the Hacker subtype as intuitively exploring the virtual world sounds very much to me like both Inventors and Architects, who share a need to understand at a deep level how systems work.

Teachers and Counselors in Keirseian terms both rely on creating networks of social relationships in order to help individuals realize their value to the community.

And both Champions and Healers feel their greatest satisfaction when they are able to establish and demonstrate deep levels of commitment to special individuals -- their friends.

...

Now, having said all this, there are problems with this attempt to establish correspondences. In particular, while Rationals and Idealists match up very well with Explorers and Socializers respectively, Artisans and Guardians don't mesh quite so cleanly with the Killer and Achiever types. It's possible to say that Artisans like to manipulate people and things, often destructively (like Killers), but they also are known as skilled performers and crafters... and how does "Killer" encompass that aspect of the Artisan temperament?

Similarly, the primary interest of Guardians is to achieve security against bad times. They are often compared to the Ant in "The Ant and the Grasshopper" (the latter being the classic Artisan) in their need to accumulate objects "just in case"... but how does the Achiever description also capture the Guardian's need to feel responsible for others, to serve and protect others?

Since the four player types and four temperaments don't match perfectly, we shouldn't be surprised that the subtypes and subtemperaments don't fit together perfectly, either.

Is perfection required? Or is "better than any other model we're currently aware of" good enough?

It's possible that some of the expressions of behavior in Keirsey's four-temperament model aren't captured in Richard's descriptions of his four primary types because behavior in virtual worlds is a subset of behavior in the universe. (Maybe it shouldn't be, but for now, action in virtual worlds is more highly constrained by the physics of those worlds than action in the real world is constrained by real-world rules.) Since Keirsey's temperaments are designed to apply to the amazing range of real-world behavior, it's not unreasonable that, if Richard's types apply best within a virtual world-specific context, those types don't match up perfectly with Keirsey's temperaments.

What I find interesting is how well they do match up despite some disparities. Artisans are more like the Killer type than they are like any other type. Achievers more closely resemble the security-through-personal-possessions Guardian temperament than they do any other temperament, and so on. Yes, there are gaps in the correspondences between Richard's four player types and Keirsey's four temperaments, and those gaps (not surprisingly) also show up in looking for correspondences between subtypes and subtemperaments.

But there seem to be a lot of congruences, too. In fact, it's the numerous congruences that create the appearance of "gaps."

So I find the mapping given above plausible. Maybe that's simply gullibility or naivete on my part. Maybe... but where's a better theory explaining why players behave the way they do? Let's see some other theory of player behavior -- not merely data or analyses, but an actual theory -- that has some explanatory and predictive value. If it's better than this one, I'll happily switch.

Meanwhile, I'm satisfied that the temperament-level aspect of this theory works, that there's something real to it. It has successfully explained aspects of my personal and professional life, increasing my happiness in both realms -- something I can't say about other models of personality I've considered. (Which I believe obliquely answers the question that Mike asks about whether having a taxonomy of behavior has had any real application.) If, as I suspect, Richard's player types model is a proper subset of Keirsey's model, I would expect it to be equally effective in its virtual world context by suggesting how to key world features to player desires.

Richard has said he believes it has been effective. And so far I know of no reason to dispute him about that. There are some extended features of his model that I think are unnecessary -- in particular, I don't agree that there is a "development track" or a "Path to Ascension" -- but the basic model seems to be sound.

And if anyone wants to give me a bunch of money to help me implement my in-progress design for a MMORPG that explicitly uses this model for its skills system and gameplay, I'll prove it to you. *grin*

To anyone who's waded through all this: thanks!

--Bart

Posted Jul 17, 2005 5:51:57 AM | link