The following is more-or-less the original form of the Threefold FAQ. I wrote it in October 1998 as a posting to rec.games.frp.advocacy. The origin was a thread entitled "New FAQ - Simulationist versus Plausible" (October 16, 1998). Following two days of discussion, I came up with the expanded "More General Threefold FAQ" (October 18, 1998). There were two minor revisions to it since then:
The Threefold Model is one way of grouping many aspects of "group contracts" into logical categories. Full group contract includes every facet of how the game is played: not just the mechanical rules, but also how scenarios are constructed, what sort of behavior is expected of PCs, how actions not covered by the rules are resolved, allowance of outside distractions, and so forth. The Threefold divides up many of these into categories known as Drama, Game, and Simulation.
An important part of the model is recognizing that there are valid different goals for gaming. Many models of RPGs or gamers tend to have derogatory categories of "munchkin," "poser," "rules lawyer," etc. which are contrasted with "true role-players". The Threefold model is intended to promote looking at different styles as just other ways of play.
Role-playing games don't simply classify into good and bad. The exact same game which one player enjoys, another might dislike. Rather than say that one or the other has bad taste, it is more useful to try to make sense of patterns of what different players and GMs enjoy. The Threefold is one method of classification, which divides styles up into how much they are Drama-oriented, Game-oriented, and/or Simulation-oriented.
What the Threefold applies to is an open question. It is frequently used to look at GM decisions during a session about what should happen in the game-world, and to a lesser extent at adventure design during a campaign. As for other parts of the "game contract", there are different views about its applicability. It may or may not apply player behavior, out-of-game methods, system design, campaign design, or other topics.
Most likely, none of the above. Your individual style cannot be pidgeonholed into a single word. More to the point, you probably use a mix of different techniques, and work towards more than one goal. You may tend more towards one corner of the triangle, but you probably value a mix.
OK, here is the short definitions:
It is true that these goals are not constantly at odds. On the short term, a given conflict might happen to be both a fair challenge and realistically resolved. However, every game will have problems, including undramatic bits, unrealistic bits, and unbalanced bits. The Threefold asks about how much comparative effort you put into solving these.
Even a perfectly simulationist or gamist campaign will have dramatic bits in them. After all, people will tell stories about things that happened to them in real life, or even about what happened in a chess game they were playing. Similarly, a dramatist campaign will have some conflicts that are a fair challenge for the players, and some events that are realistic. But an equally-skilled gamist GM, who doesn't put excess effort into the quality of the story, will be able to make better challenges. Similarly, a simulationist GM, who focusses only on in-game resolutions, will be able to make things more "realistic" for that game-world.
You can't make it go away entirely. It is an idea which some people are bound to talk about. The way to reduce its usage in discussion is to not talk about it. It existed before this FAQ was around, and will still be around even if it goes away. If you feel it isn't accurate or complete enough (and it isn't), then invent a better model and talk about that instead.
Simulationism is not defined in terms of believability, it is defined in terms of method. For example, you as GM you could have a storyline in mind, and set up the background and characters so well that during the game, the storyline occurs without your having to noticably fudge. A very simulationist player might not notice that you constructed the events to produce that story. However, if she found out, she would feel cheated: you would have violated her preferred contract.
Rightly or wrongly, a pure simulationist isn't simply trying to produce a story that is believable. He is trying to actually find out what would "really" happen by modelling what is in the game world. Of course, it is impossible to perfectly simulate this, but he finds interest and value in the attempt.
For example, say the PCs are know that a target is hiding in one of eight hotels, but they cannot find out which except by searching them. A dramatist GM might decide based on pacing to have the second hotel they search be the right one, so that the game doesn't drag as they go through one after the other. This is perfectly believable, but a pure simulationist GM will refuse to do this. Most likely, she will decide on one in advance and let the players choose what order they search in. The players might find it immediately, or they might have to wade through seven others.
No, those are rabid stereotypes. Even if the stereotypes have some truth to them, the Threefold is not about just the lowest common denominator. There are good and bad examples of each type of game.
A pure dramatist might run a gritty, low-key drama where the PCs are true-to-life characters who perhaps concentrate on their work. In this case, the dramatic story might be framed around how they relate to each other and the tension produced. Dramatist campaigns also include comedic campaigns, where the in-game action is tailored for humorous effect rather than classical "drama". The key is that in-game events are tailored based on how satisfying the storyline of the campaign is.
A gamist could run a mystery game where the PCs are challenged to find the killer based not just on physical clues, but also on the personalities and motivations of the suspects. Note that this is similar on the surface to a dramatic story, but the emphasis is on making it solvable yet challenging to the players. A purely dramatist mystery might make a better story, but a purely gamist mystery will be a fairer test of the player's wits.
Simulationism by definition is going to try to be "realistic" within the game-world, although it may have natural laws different than the real world. However, the players are not neccessarily obsessed with rules or physics. A simulationist game could just as well focus on political discussion between important figures, or rebels fighting a propaganda war to win over the masses. Several posters have run diceless simulationist games.
A purely simulationist mystery would start with determining how the crime was carried out based only on in-game factors. The logical consequences of this might mean that the players can solve it easily, or that they can't solve it at all, or that they can only solve it by turning it over other authorities. An absolutely pure simulationist GM won't go back and change things to make the mystery work better for the PCs.
Gamist was not intended as a catch-all for anything that isn't included in the other two categories. It is specifically about setting up fair challenges for the players to face. The Threefold is not intended as a be-all and end-all of gaming, nor is it neccessarily complete. Several people suggested a fourth group of styles, which was "Social". However, discussion died down as there was no consensus about what that meant in contrast to the other styles, or even whether one could even discuss it on the same level.
Many aspects of gaming are not covered by the Threefold. For example, any of the three can vary from "Light" to "Serious". "Beer-and-pretzels" usually refers to Gamist dungeon-crawls, say, where you are trying to beat the monsters. However, there are also non-serious dramatists, say who run cheesy superhero plots where the hero always beats the villian. Note that this is not gamist since there is no challenge to it -- the hero always wins, it's just fun seeing how she does it.