Your collection of "guys" is a good, qualified overview of the field, but at the moment it seems to lack the foundations which show us how each differ on specific topics (such as the goals of Bartle's four types). As an example, both "Sight See'er" and "Alt-Guy" share a desire to experience all of the game. What else might they have in common? How could the game be shaped to fulfill their desires, and how would this affect the other profiles?

This could possibly overlap with the "BrainHex" quiz devised by International Hobo. I've found them to be quite adept at creating profiles of gamers based on personality types. It would be interesting to start with your profiles, and "reverse engineer" them with the quantitative tools BrainHex (or some other non-MMO specific personality test, such as Myers-Briggs) provides.

link: http://blog.brainhex.com/

Interesting.

Have you looked into what Bartle has been writing since he wrote his first paper? I think he would agree with you that the original player types are overused (in fact, he has).

Might I suggest you look at:

Richard Bartle. Understanding the limits of theory. In Beyond Game Design Nine Steps toward creating
better video games, Chris Bateman Editor, 2009.

and

Nichole Lazzaro. Why we play games: Four keys to more emotion without story. White Paper: XEODesign, 2004.

http://www.xeodesign.com/xeodesign_whyweplaygames.pdf

Good, thought-provoking post. I'm all for more efforts to develop additional tools that help us understand what gamers want.

I've written about the Bartle types myself; specifically, I suspect they're a gamer-specific subset of David Keirsey's four general "temperaments" (http://flatfingers-theory.blogspot.com/2005/01/bartles-player-types-and-keirseys
.html). In fact, I think we can see correspondences among several of the better-known gamer personality style models (http://flatfingers-theory.blogspot.com/2005/01/styles-of-play-full-chart.html), all of which seem to be recognizing the same four fundamental motivations for what we enjoy when we play games.

On Nick Yee's dismissal of Richard's original four-style model... Nick has held for some time the opinion that his preferred independent-factor model is the only acceptable kind, and that the four-factor model derived from two axes of observed behavior is merely "speculation" and therefore illegitimate as a tool for understanding gamer motivations and predicting gamer behaviors. (http://terranova.blogs.com/terra_nova/2003/11/empirical_frame.html) Richard, meanwhile, has responded to Nick's criticisms. (http://terranova.blogs.com/terra_nova/2007/10/the-hidden-bart/comments/page/1/ and http://terranova.blogs.com/terra_nova/2005/01/will_the_real_e/comments/page/1/)

As to "Know where to find things" versus "Know how to get things," I haven't seen the internals of the old Bartle Test but I think I can say with some assurance that the first one reveals an Achiever motivation while the second is what an Explorer would tend to say. The location of an item is usually a static fact allowing a direct-line acquisition of that item, while knowing how to get things is dynamic process-knowledge. The former is more attractive to Achievers as a simple, concrete, collectible item (e.g., "the Pixie Crown is at X=134, Y=2203 on the Hoek Woods map"), while the latter (e.g., "the aggro radius of the NPCs in this area is much larger when all three moons are in the sky") feels more satisfying to Explorers who enjoy understanding complex systems without feeling a need to profit in some tangible way from that knowledge.

Finally, I'd suggest that the suggested player types overlap somewhat, particularly if we're trying to build a model that explains the internal motivations for the behaviors expressed. Predictability, I'd think, depends on the internal logical consistency of the model. Still, as quick-and-dirty profiles, there could be some utility in the suggested styles. I hope they'll get plenty of constructive attention in other comments on this excellent blog post.

@ Bart

Try reading the Bartle paper, "Understanding the limits of theory". He says himself that his player types have nothing to do with motivation, they're simply categories of players based off of their behaviours. That's why it's misused and abused so often.

Andre, I also respectfully disagree with Richard regarding his theory that individuals change their core styles of play, and that they do so in one or more specific sequences of styles. Whether he feels that his original four-type system reflected something internal motivations for styles of play or not, I believe there's reason to think it actually does.

Of course I'm just some guy, while Richard is... well, he's Richard Bartle. :) But while his authority should be respected, I think he'd agree that authority should never be the final word when it comes to trying to understand reality. People should be free to review the arguments made -- playstyles as expressions of innate and generally unchanging personal temperament, or playstyles as expressions of behavior that change in one of two specific sequences for every gamer -- and decide for themselves which explanation seems to fit better with observed reality.

Labels are getting in the way, I think.

Rather than saying "player killer" (which brings to mind a max-level ganker lurking at a newbie spawn point) or something like that, if we go back to Bartle's original works we see that it's more complex than that.

The two-axis system he used has one axis as "compete / cooperate" and the other axis is "with other players / with the game-world". This gives rise to the four player types for this particular analysis:

Compete / with other players : PvP'er. The "killer". In Bartle's writings, this isn't just who pwns who. It's also political, in which the competition is for social power. Many guild leaders and forum trolls fall into this category (for good or ill).

Compete / with the environment : the "achiever". This is the person who wants to do it all, and who wants a scoreboard to see how he's doing. An achiever might not care what YOUR score is, just caring about "29/30 quests done" and tearing his hair out over the last. Thottbot-addicts and people who min-max just for the sake of being able to generate the big numbers fall into this category.

Cooperate / with other players : the "socializer". This is someone who is online for the sense of camaraderie. It's less important what the socializer is doing and more important that they are doing it with others. Some "real" roleplayers (the in-character types) fall into this category.

Cooperate / with the environment : the "explorer". This is the person who wants to see it all, who is the lore-junkie, who doesn't like spoilers. In online games they are often primarily soloists.

If you look at Bartle's "Designing Virtual Worlds", you will find extensive work in this area. This particular two-axis classification system is but one of many which Bartle has explored; he covers others which may be of use. It is also worth remembering that at the time there really was very little difference between the major games out, and there were few of them. However, for games of that type which still exist today (and I include WoW in that), it is still relevant. Anyone who is serious about being an MMO designer needs to have a well-worn copy of "Designing Virtual Worlds" on their desk and refer to it often. It may not be perfect, it may be showing its age, but it is still a good read and is still relevant... and if you're going to take your own direction, it makes a good reference-point to start from.

As the original poster has.

Lance, I agree regarding the "Killer" label. From my perspective (whatever that's worth), this kind of player can be a PKer or griefer, but I see those more as negative-behavior subsets of the underlying motivation, which can also have positive forms of behavioral expression.

That's why I started thinking of this type as the "Manipulator."

That term, I think, pretty neatly captures both the positive and negative senses of sensation-oriented play. Sometimes I wonder whether gameplay features based on skillfully manipulating things (like playing with the edge cases of the gameworld to see what breaks) and manipulating people (which encompasses not only PKing and griefing but the more constructive forms of political play as well) have been given short shrift in online games just because of the stigma attached to "Killer" as a type of player.

...

On reflection, I feel like I might be yapping too much in someone else's sandbox. :) I'll hush now to make sure Dan has a chance to direct the flow of this conversation as he wishes.

In other entertainment industries they call this approach "formulaic".

It's true you might find a "recipe" which pushes all the "correct buttons". But invariably it will lead to a product that is flashy and slick, but ultimately dead and soulless. It's manipulation, and people will see it for what it is.

Games are an entertainment medium. The best way to make them is through authentic connection and originality. People aren't machines. They're... um... people.

I tackled this subject a bit in my thesis work. Although I won't bore you with the brutally long details, my conclusion was that a "Competitor" player type was one of the biggest misses in the Bartle list. It's understandable because the core of his extensive expertise stem from the earlier days of MMOs such as with MUDs.

Competitors are similar to achievers and killers in many respects, but differ in one important way: intent. The intent of a competitor is, of course, to compete - particularly in a skilled way. This does not need to involve killing or achieving anything substantial, though both are common results of competition.

You may ask, how can a player defeat another without competing? The immediate answer to that is kills that are considered "griefing," in which a player preys on weaker opponents, but there's more to it than this. The gameplay of many games is not complex enough to allow for skilled competition. In these games, victory is determined by a player's gear and other similar factors. "Killers" of this game seek to dominate, not to compete. Semantically, one can say that players are always in competition with one another for power, but are they after the end result or the competition itself? This is what separates competitors from other player types.

Thanks for the comments guys.

My list of Player Characters is just meant as a pencil sketch to get the conversation going.

My desire is to develop a Player Type tool that will have real world applications in game design and development. Let me sketch out what I want to use a Predictive Player Type test for. To save me some typing let's call this new test PPT.

Assumptions. I believe that World of Warcraft, much like the Old Country Buffet restaurant, offers something for everyone at a reasonable price. I believe that most new games will draw primarily from players that have or are playing a Massive. Therefore to bring a viable game to market we must define a niche that will intice players from their current game.

PPT & Game Design. Let's say I want to do a game World of Duel-Hammer. This will be aimed Primarily at Duel-Guy, so I want to find out as much about him as I can. I want to find if there is a strong correlation between Duel-Guy and another Player Type. Let's say the PPT shows that fifty percent of Duel-Guys also rate high on Verbal Griefing Guy (aka Barrens Chat). This may or may not be a problem, but something like this needs to be addressed early in game design.

PPT & Money People. The next step is taking our game before the Money People. I need to use the PPT to show the money people that Duel-Guy is a viable portion of the general player population. I would also like to do a market survey of Duel-Guys we have identified with the PPT, and get their reaction to the general feature set of our new game, to bring to the Money People.

PPT & Beta. I would use the PPT to select my beta population. If I'm making a game for Duel-Guy I want a least 50 percent of my beta population to be Duel-guys.

Additionally the PPT would be used in Marketing, helping a player a the character selection screen and allocating live team resources among other things.

To me, a PPT would be just one tool to use in game design and think about Player Types.

OK, well since my name seems to be appearing here with some regularity, I guess I should post...

First, some general points:

1) I see at least one article like this every year, some years two or three. This suggests that my player types model does indeed need revisiting, although the fact that none of these alternatives have yet taken off indicates it's not as easy as it might first appear. I've been longing for years for someone to produce a better theory, but to date no-one has. This is an endless source of frustration for me; I dearly want articles like this one to come up with a model that supersedes mine, because a better model means better virtual worlds. Nothing has turned up yet, though.

2) I am alive. I can easily be found on Google. You can even guess a web site and email address for me that will work. It's simple to get in touch with me. So why do people who write articles about my player types model never email me beforehand? If they did, then I could comment in advance on some of the misunderstandings that inevitably arise, and then wouldn't look the jerk I do for pointing them out in public.

3) So often do I see articles like this one that (as Andre Gagne mentions) I wrote a piece explaining the basics, which people thinking of penning their own can read before they make a start. I did this in the hope that by pointing out some of the common issues in advance, they can hit the ground running and we might see some progress. It only recently appeared, though, so I guess it didn't happen in this case (it's Chapter 4 in http://www.amazon.com/Beyond-Game-Design-Creating-Videogames/dp/1584506717).

Second, some misconceptions I often see:

1) I wrote the player types paper. I didn't write the Bartle Test. The latter is based on the former, but that doesn't mean you can criticise the former because of perceived failings in the latter. That would be like criticising a song because you didn't like the way it was being sung.

2) Both my work and Nick Yee's have advanced. Since 2003 I have had 8 player types and since 2007 Nick has had 10 facets. These are actually pretty well in accord with one another (Nick's extra facets come from features that are emergent consequences of my model, to do with immersion).

3) My player types model is just that: a model. It's not a categorisation system. This isn't to dismiss categorisation systems, but it is to say that there is more to a successful tool than merely pointing out a bunch of stereotypes you've seen. That's the easy part: the hard part is understanding why those categories exist. You need to know WHY they're there in order to make use of them; merely knowing THAT they're there is only of limited use.

4) The utility of a model depends on the target audience. My player types model is for designers; if you want a model for players to use for selecting raid make-up, or for customer service reps to use for adjudicating complaints, or for virtual store owners to use for luring customers, well, my player types may or may not be of much use to you.

5) My model applies only to people who play virtual worlds for fun. It applies to all virtual worlds, whether social or game or anything else, but only for people who play for fun. If you play for other reasons (teacher, journalist, designer, gold farmer) it doesn't apply to you. You may still find it's appropriate, but I don't make any claims that it is.

OK, so having got all that out of the way...

Looking at the categorisation we see in this article, the first question I'd ask is: whom is it for? It looks to be for designers, although it seems to have a fairly player-experiential focus. It's also MMO-centric (and a particular kind of MMO, too - you won't find Auction-House-Guy in an MMO with no auction house, for example). This isn't necessarily a bad thing, of course, but it does raise the supposition that there's a reason why the categorisation only applies to modern MMOs - that there's some underlying system which the current paradigm uses that this categorisation somehow exploits. We don't get any inkling of what this might be from the article, though. My own guess is that the article is about addressing the needs of people who are experienced MMO players, who will have certain expectations about what MMOs should be like, who have found particular features of these MMOs that they like having, and therefore if you want to attract them to your new MMO then you should provide those features. This is all well and good, but you can get that from a straight features list - you don't have to personify it by inventing player types to embody the features. If you do want to categorise players, rather than features, you need to look deeper into why it's important you do that. This would deliver advantages: if you know WHY someone likes the AH, then you should be able to think of some other things they would like, which might possibly be nothing like the AH. This is where the power of having a model, rather than an arbitrary and incomplete categorisation system, comes into play.

The next thing I'd be looking at is what we can use the model for; the answer here is predictiveness. OK, well this is a pretty good reason. It's perfectly reasonable to say that if we have a bunch of people who are fairly set in their ways, we should give them what they want. Again, though, it would be more useful if we knew why they liked what they like, so we can give them something they want that they perhaps don't know they want. Bob the duel-guy wants professions that allow him to create gear for duelling. Well, he does, yes, but what he really wants is to win the duels. If he could get better gear from reputation or questing or some other approach we haven't thought of yet, he perhaps would. On the other hand, perhaps he likes to make an investment which pays dividends in the longer term.? The more we think about him, the more the different possibilities open up. Suppose we decide that he likes to work for long-term results; this suggests other attitudes that could sit alongside his. Someone who likes to work for short-term results might be challenge-guy; someone who wants long-term goals handed to them on a plate merely for having been around could be raid-guy; someone who wants instant gratification for little effort could be sight-seer-guy.

See how if you start looking at the reasons behind why people do things, you end up with a more constructive model? In this example, what we got isn't a bad fit with some of the types the article suggests (I wouldn't suggest using it, though - it's just something I made up on the fly without any analysis), but it only covers 4 of them; what about the others? Well, that's the kind of thing I'd like to see: some justification that the categories you have produced are consistent, coherent and complete. The more foundations you can build, the better chance your typology has of being able to stand on them. If it does stand, then the final question is whether it's going to be useful or not, however we can't really say much about that until we know what those foundations are.

So what we have here is a start, but it's only a start. It needs more work for it to be of practical use. The article says as much itself at the end, of course, but it doesn't indicate how this should proceed - just that it will need a team of people. Well yes, but what are they going to do, and to what end? And how will you know when they've done it?

Ultimately, you need a model, not a checklist, because a model is a system, not data. You can get a long way with data, but a system will get you further.

Richard

PS: Bart Stewart's comment about the Keirsey tempraments. These (and Myers-Briggs, from which they derive) have been linked to my player types theory but don't match up experimentally. See the piece by Chris Bateman in the book I cited earlier, or in http://www.amazon.com/exec/obidos/ASIN/1584504293/1n9867a-20 .

I've talked with Chris before; he seems quite sharp. Books are on order.

In the meantime, there are some experimental data that do appear to support the alignment of the four Keirsey temperaments and the four original player types: http://finegamedesign.com/personality.html . I wasn't aware of Ethan Kennerly's site until a little while ago, so our analyses were independent of each other. Thus it's at least a little interesting that Ethan's conclusions line up exactly with mine. His assignment of types with temperaments, based on analyzing the numeric data from Andreasen's test (http://www.andreasen.org/bartle/mb.html), exactly matches mine which (I freely acknowledge) was based on perceiving similarities between the descriptions of the types and temperaments as well as of the two dimensions of personality that underlie both models.

Definitely not conclusive, of course; lots of people could "prove" that the world must be flat. (Plus there's the weirdness of some Rationals showing up as Killers in the Andreasen data.) Still, as a working theory, is it so unlikely that a model that generates four fundamental play styles might turn out to be a special case of a similar model in a larger context that generates four fundamental personality styles?

But to return to the question of the original post: the author's emphasis on features rather than motivations might actually mean an independent-factor approach like Nick Yee's could be more useful. Rather than a list of "X or Y but not both" choices, why not provide a multiple-choice list of features (i.e., checkboxes rather than radio buttons)? (See "approval voting": http://en.wikipedia.org/wiki/Approval_voting .) Let people select all the features they enjoy, rather than asking them to select one thing they enjoy most, then total up the times each feature is selected by the responding group. The result will be a multi-modal distribution, but that will be a more useful representation of the specific features the test group enjoys than extrapolating features from any one style of play.

Of course, that will only measure interest in known features. Unlike a model, it won't tell you anything about other (unlisted) features that potential players might like even more, because (as Richard implied) it doesn't speak to why some people prefer those things over other things. But perhaps that's OK for the author's purposes.

Bart>I've talked with Chris before; he seems quite sharp. Books are on order.

Damn, I should have asked him for a kickback before recommending them...

>In the meantime, there are some experimental data that do appear to support the alignment of the four Keirsey temperaments and the four original player types:

Yes, I know about Ethan Kennerly's work. I'd previously wondered if there were a connection between the four medieval "humours" (sanguine, phlegmatic, choleric, melancholy) and my player types, because, well, there were four of each, but I didn't know that these humours were just a medieval expression of the older notion of the four elements, nor how this led via Jung to Myers-Briggs. It was exciting to see a link made, because if it did prove to be strong then that would support my player types theory (ie. we'd know we were on the right track with it) and it would mean that all the theory that had been developed in this area could be directly applied to MMO design. However, I did have lingering doubts, in that I knew that players changed type over time (something we knew even before we had a formal notion of player types!), but these temperaments are supposed to be fairly fixed in adulthood. Ethan showed a correlation between Bartle Test results and Keirsey types, which suggests there could well be a causal link between the two; however, there's no indication as to what that causal link may be, just its possible existence.

>Still, as a working theory, is it so unlikely that a model that generates four fundamental play styles might turn out to be a special case of a similar model in a larger context that generates four fundamental personality styles?

No, it's not unlikely at all. It would be strange indeed if the player types I came up with had not been noticed before in other fields, if they in some way reflect elements of common human psychological experiences. Indeed, I myself link them to the Hero's Journey, so I'm not averse to this kind of thing!

The next step, then, is to ask what making this connection buys you? If we beg the question, and accept that my player types are just a particular incarnation of Keirsey/MB/Jung/humours/elements, how does this wealth of new knowledge help virtual world designers? Given that the player types are one riff of a much larger body of work, what does the full force of the full theory tell us beyond what the player types stuff does?

>But to return to the question of the original post: the author's emphasis on features rather than motivations might actually mean an independent-factor approach like Nick Yee's could be more useful.

If it were factored on features, well yes, that could be very useful. It doesn't need player types, it just needs a feature list. As with all correlations, though, it wouldn't tell you where the connections were between the elements, so it might be that a designer could decide to cut a small feature which, while seemingly unimportant and unregarded in itself, has important consequences. For example, in Asheron's Call 2 the design cut out the "make a trip to town" feature, because players didn't like having to go back and put their stuff in the bank and so on. However, this undermined a feature they really did like, which was bumping into other people by accident, making friends and forming groups. This is why you need a model, not just a categorisation.

>Let people select all the features they enjoy, rather than asking them to select one thing they enjoy most, then total up the times each feature is selected by the responding group. The result will be a multi-modal distribution, but that will be a more useful representation of the specific features the test group enjoys than extrapolating features from any one style of play.

This would be fine if the players knew what they liked. Often, they don't - or, worse, they think they like something they don't like. See my 2004 Gamasutra soapbox piece for a rant on this...

Richard

"The next step, then, is to ask what making this connection buys you? If we beg the question, and accept that my player types are just a particular incarnation of Keirsey/MB/Jung/humours/elements, how does this wealth of new knowledge help virtual world designers?"

I hadn't actually taken on such a constructive project yet!

At this time, the value of showing a connection between your player types model and an independently-developed broader model of personality isn't (to my mind) about generating new information as it is about serving as a possible confirming instance that both models are saying something truthful about reality. The claims that some people make that your original player types model is completely bogus become harder to sustain if there's evidence that it's one window among several others onto some deep structural pattern in the human psyche.

That's why what particularly excites me is the possibility that you and Keirsey and Caillois and Lazzaro and Edwards (as I try to document at http://flatfingers-theory.blogspot.com/2005/01/styles-of-play-full-chart.html) and others have all been seeing aspects of some kind of universal truth about human personality. It's a bit odd to see that fourfold model pop up again and again (albeit tweaked in the case of Ron Edwards's original GNS model) unless there is some kind of underlying reality. There's no guarantee that such a deep reality exists. But the possibility that there might be something real there, that there truly are four basic innate motivations for the behavioral choices we make... I find that such a fascinating possibility that it seems worthwhile to explore it.

Of course it's important to bear in mind the potential danger of trying to wedge any observed phenomenon into a beloved model. ("Four cardinal points on a compass -- huzzah, another confirming instance!") I also acknowledge that there's nothing magical about the number four -- useful models could have eight factors (like your updated 3D model) or sixteen (like the Myers-Briggs) or some other number. If those are refinements of a simpler and more fundamental system of innate styles, however, then I think it's reasonable to hold that a basic four-fold model can have utility (including predictive utility), too.

At some point I'd like to try developing a few constructive tools based on this theory of some deep reality of four basic motivations, to see how well such tools might work. (I certainly want to develop my own survey instruments; the various "tests" I've seen over the years have seemed very poor to me.) But for now, I'm still looking for flaws in the Big Correspondence theory; if someone can make a good objective case that it's fatally broken, there's no point in trying to develop tools based on that theory.

"This would be fine if the players knew what they liked. Often, they don't - or, worse, they think they like something they don't like."

Very true, which I suspect is why it's popular to suggest assessing styles by watching what people actually do when they play a game, on the theory that people wouldn't actually do things in a game that they don't truly enjoy doing.

The problem with that approach (versus a text-based, pick-an-answer test, which has its own problems) is twofold. First, there still has to be a model in order to associate the actions that an individual takes with some style or behavioral interest... so how do you decide which in-game actions go with which style? And second, if you're going to use an existing game or virtual world for your test instrument, rather than building one expressly designed to expose preferred styles of play, how do you insure that every player at any moment can choose to express any of the possible styles that are defined in the model of choice? If you drop me into some zone of some existing game, and I immediately start reducing the local rabbit population to collect their pelts, how can you be sure that I'm really an Achiever if most of the gameplay activities available in that zone (or that game) are things that Achievers typically prefer to do?

Alternately, would a game that's tailor-made for testing playstyles be enough fun to play that a statistically sufficient number of gamers would finish the test?

...

If nothing else, it appears that the field of "understanding player motivations" will remain fertile ground for study for years to come. :)

Thanks once again for the comments guys. This has helped me clarify my thoughts and pointed out where I have been less than clear in my original post. Not to mention just generally interesting.

I am primarily interested in tools to help me design Massives and get Massives to market. I think I can illustrate what kinds of tools I am looking for by describing three potential problem areas in a Massive I am currently working on. In this case by potential problem I mean core game features that are not "main stream" nor anticipated by players.

1. For player vs. player reasons I would prefer my servers to be up only during prime time. This is going to cause conflict with players who have been trained by beginner Massives, such as WoW, that servers are up 24/7. I don't want to lose potential players over non 24-7 servers, so the question is how can I mitigate the fallout from it. My solution is to let players log into major cities, but not go out into the countryside. After reading the comments on this thread, I realized what I had was a general issue, one that affects all players, therefore the tool to use should be a general tool. I revisited non 24-7 servers using Richard Bartle's 4-axis model. I found ideas just popping.

As long as I let Socializers log in I will mitigate some of their concerns. Perhaps I can sweeten the deal by adding some perks for Socializers during down time. Perhaps a guild recruitment expo area that only opens during non prime time. In a similar fashion Explorers could get content that only opened up in non-prime-time, such as dungeons that unlock in the cities, quests, or hopefully something more creative. Non 24-7 servers will affect Achievers the most. The first idea that come to mind are bonuses to trade skills during non-prime hours. I know I have to do more for Achievers, but I can put this aside and wait for inspiration later on. Finally Killers on non 24-7 servers will understand it is necessary for good player vs player gameplay. That said I will try to include duel ladders only during non prime time.

Bottom line, I have a general problem and I was able to focus and get help from a general tool, Bartle's 4-axis model.

2. I can not make Raids work in my game world. It's hard to define what a raid is other than I know one when I see it. Let's call it ten or more players cooperating in a player vs. environment mini-game, either player-scaled such as original EverQuest or designer-scaled such as WoW. I have nothing against Raids, and I certainly want to get Raid-guy to subscribe to my game, but I can't make it work. So I want a tool to identify what not having Raids is going to do to my potential player pool. Is this going to be a game-breaker? And what can I do to minimize the effect of not having Raids? This is a fairly specific issue. This is where I want what I am calling a Predictive Player Type or PPT test.

First I need to know what percentage of the general Massive player population are Raid guys. All sorts of player types Raid, but not all of them are Raid guys. Enrique's Competitors may be in a Raid, a Sight-See'er Guy might be in a Raid, a Socializer or Achiever may be in a Raid, but all these guys will happily find some other area of a game to participate in if there are no Raids. Raid-guy is a unique guy. If Raid-guy is ten percent of the general Massive player population then I have no problems. If Raid-guy is fifty percent of the population I might have to rethink. If Raid-guy is one percent of the general population, I'm not going to lose any sleep. My gut is that Raid-guy is closer to twenty percent of the population and thus must be addressed.

Richard makes the point above that "we need to know why they liked what they like, so we can give them something they want that they perhaps don't know they want." I completely agree. For specific issues we first need to define specific player types and then find out what makes them tick. When I first looked at this problem I noticed that Raiding is the most Social player vs. environment mini-game. Perhaps I can make Social player vs. player more attractive in some fashion to these Social player vs. environment Raid guys. At this point I just don't know enough about Raid-guy to know for sure. I also want to know more about other player types that raid but aren't Raid guy.

So I want a tool that: identifies unique player types, quantifies them, and finally analyzes them.

3. WoW has two sides, Horde and Alliance, which are all but identical except in art. In my game I want the two sides to be different. One side will be weaker and require more co-operative play than the other. I want the player communities on both sides to be different, not just their toons. If I can make this work it will be more art than science, and I can't envision any tools that would help. What I will do is plan a large enough and long enough beta that I can scrap the idea if it doesn't work.


I hope the above has clarified my thinking a bit. I want to identify player types on a more granular level than has been done to date, I want to quantify them, I want to know the correlation between the player types and finally, I want to know everything I can about each player type. I realize that is a lot of work.

I will be writing in more detail about player types I have identified in the weeks ahead. Additionally I am willing to put in one day a week on this project. My BA in Psychology and BA in Sociology have taught me I don't have the chops to design questionnaires, tests and the like. So we need a real sociologist on board. We need a pool of players we can question, perhaps one of the gamer sites like Massively.com. And we need the thoughts and comments of people in the industry.

Step One is a discussion here on player types we can identify.

Step Two is to query Massive players about their own self-reported Player Type. Additionally query them about other player types they see. After all one guys self-reported Friendly Social Chat Guy, is someone else's annoying Barren's chat guy.

Step Three is distill the two sources of Player Types and come up with our unique Player Types, with the knowledge that since we are going for specifics, this is a moving target.

Step Four is to find the correlations between our unique Player Types.

Step Five is to understand the motivations of each unique Player Type.


-------------

About the time Richard Bartle first wrote about player types I brought a PC-based six-degree-of-freedom motion-platform arcade game to market. My team was composed of designers/gamers constantly fighting to get High Score on our game. When we got to Friends and Family testing, either they weren't really arcade game players or they were Competitor-types like us. When we first introduced the game to the public at an industry show I got a huge bite in the ass. We used self-adjusting-difficulty to soften the noob experience. Even with that some people were getting frustrated and even mad at us for how hard the game was. As someone who had fed an Asteroid game a literal mountain of quarters to master it, I just did not understand the reaction we were getting. So we placed some machines in a local arcade and watched people play. We also spent a lot of time at Dave & Buster watching how people play. We modified game-play so that early levels were much easier and adjusted the self-adjusting difficulty to ramp the difficulty up quicker for good play, so a competitive player wasn't stuck doing easy levels for very long. What had happened is that the Achievers (high score guys) and Killers no longer went to arcades, they were at home playing console games. The arcades were now filled with Social players.

So every time I sit down the bite on my ass reminds me that understanding your player type is one of the most important elements of game design.