Before I forget and lose them (since The Forge doesn't have the neat RPG.net feature of being able to subscribe to threads)

The Raid on the Bank Nationale de Paris, started by Gareth (contracycle). Cited as an example of Sim in a recent RPG.net thread, in answer to the general problem of "find an example of a game which is identified as Fun Sim by GNS proponents".

[Middle Earth - home brew] 2nd day in July 4th '06 week of play, started by Jay (Silmenume). Culminates in a very harsh judgment of Jay by Ron Edwards, which I don't think is entirely justified, and which also IMO terribly confounds a value judgment of Jay's play (i.e., RE's distaste is evident) with an analysis of how Jay's play works. About the latter, I see some conclusions being leapt to--especially the notion that the GM is masking his control over the game behind a mass of dice rolls, acting as a sort of Maxwell's Daemon. (Cf. also The Sixth Sally from Lem's Cyberiad, if you care to.) I suspect this mixes a distaste for and incomprehension of bricolage ("tinkering" with game-reality) with assumptions about power/authority distribution.

Note that Jay considers his play to be Sim. I can't tell what RE and Ralph (who also commented on the thread) consider it to be. Likely Sim, but unlike Jay they seem to think it's a product of GM-centric illusionism, while Jay sees it more as collaborative bricolage.

The first example is interesting because it'd take some effort to see it as Nar or Gam under GNS and therefore it does look like a fun example of Sim. But it raises some questions. What would the group do if the dice came up "wrong" such that one of the PCs was killed? Did the players really believe that the SAMs might take out their getaway helicopter? Were they worried about the consequences if this did happen? If the answer to these are "fudge/manipulate", "no", and/or "no", did those factors bother anyone? And also, if all that was the case, why did anyone bother to roll dice at all--that is, what added value is achieved through use of (in Forge terms) Task Resolution?

Furthermore in connection to Jay's post--is it necessary that the manipulation (either open or hidden) be restricted to the GM, or could it be group-consensual and still support the kind of fun shown here? Does group-consensual manipulation--player empowerment--necessarily bring about Nar? (I think not, especially if Nar is defined in terms of Egrian moral Premise, and even if "player empowerment" is facilitative of Nar.)

(Disclaimer: I'm very skeptical of GNS, partly because I think it's poorly defined, but when evaluating its utility as an analytical tool and looking at it as a social phenomenon, I try to take its framework at face value.)

I'm probably going to ramble a bit. Sorry.

The idea for this post starts with an RPG.net thread titled Thoughts on Simulationism, where Jay (Silmenume) is continuing his efforts at explaining his idea of Sim. A number of others jumped in, and I was interested to see them lining up along several positions. Jay & Thanuir seemed to be trying to find ways to refine Sim within GNS. Other posters (including me, but I was a latecomer) basically critiqued GNS by saying that its system of categorization is incomplete: the things excluded by Nar & Gam aren't effectively described by the remaining category of Sim. At least one poster insisted both that GNS Sim does adequately describe a real style of play and that the things the other posters felt weren't included in GNS-Sim were actually forms of Nar--i.e., the "Shit! I'm playing Narrativist" thing from Simulationism: The Right to Dream.

Some additional discussion clarified that Thanuir, at least, isn't trying to define all Sim, but is trying to come up with, basically, a constructive model for a focused type of play which he thinks isn't covered by GNS Nar or Gam. Meanwhile, other posters including Eric Brennan & John Morrow engaged in some description of their own existing play styles and preferences, explaining why they didn't see them included in GNS. I contributed observations and links to Chris Lehrich's writings about bricolage in RPGs.

The key point that arose from this for me is that the type of gaming which most (?) self-identified Simulationists say they prefer is something that, as Chris wrote in one of the links I provided, is resistant to Forge-style mechanical engineering, and which appears to be "intrinsically incoherent" from a Narrativist or Gamist perspective.

I also see ties to Jonathan Walton's proposal of low-impact games designed to work within "communities of practice", although I may be mistaken or premature in seeing a connnection. The most important similarity is the fact that Jonathan's proposal appears to rely very strongly on the particulars of each individual group and therefore defies a common precept of Forge-influenced design & analysis: that whatever it is that your group does to make a game great can be encapsulated in a rulebook. (Like the idea of storing a successful post-industrial business model in a three-ring binder, or of storing a successful Mesopotamian social system on clay tablets, from Snowcrash.) Where I think I see a bit of divergence between Jonathan's ideas and some of the stuff I've been writing about is that I can't tell exactly where his "low-impact gaming" obtains its initial group-focus for play; in the bricolage model, it seems to me, the initial focus can be provided by a very concrete way of playing which is limited but not restrictive. As play continues, this "seed" grows outward. In D&D, what happened was that people discovered new ways of applying the rules to handle situations outside the dungeon.

I made a rather long comment in Merten's blog so rather than reproduce it here, I'm just going to give a link to his entry. Here it is. (May take a little while to show up as Merten moderates his comments.)

I get a little polemical. But actually I think we may be approaching a new level of understanding across the RPG theory world, with at least some acknowledgment of immersion as a genuine, worthwhile aesthetic goal--not just a form of design conservatism--albeit not one that everyone shares.

By way of background, these threads:

http://www.indie-rpgs.com/forum/index.php?topic=17612.0
http://www.indie-rpgs.com/forum/index.php?topic=17651.0

What I'd like to know is: does anyone know of any AP threads on the Forge which are (identified as) functional expressions of Simulationism?

Yes, I know it's kind of beside the point if "GNS is over".

This thread concerning '"Channeling" and GNS' on the Forge ended badly. Perhaps its successor here will be more productive. (Note: "channeling" refers to the practice of "channeling character"--seeing and doing things entirely from the perspective of a character. "Drama", I suppose, refers to the desire for thematic developments in the course of play.)

One thing that came up in the doomed thread was a comment from Ron Edwards that "channeling/drama" conflicts are likely to arise when Narrativist and Simulationist CA's clash, but that a "channeling/drama" conflict doesn't mean that a CA clash is occurring.

It turns out that this is an unbelievably complex assertion. It can be approached from many angles, but before you can even start, you have to grapple with the fundamental limitations of normal speech and how people tend to think about words like "correlation" and "typical" when they aren't being careful.

Suppose you have 100 cases where games went wrong due to either "channel/drama" conflicts or N/S clash. They might break down as:

40 "channel/drama" clash and N/S clash
20 N/S clash only
40 "channel/drama" clash only

With these numbers you could reasonably say that an N/S clash typically entails a "channel/drama" clash, since 67% of all N/S clashes are also "channel/drama" clashes. Conversely you might say that a "channel/drama" clash is just as likely not to involve an N/S clash as it is to have one, so there's nothing "typically N/S" about "channel/drama" clashes.

But let's look at things a little differently. Suppose that in addition to the 100 cases we just talked about, we've also seen some number of clashes which involved neither "channel/drama" nor N/S clash. And now I say that one type of clash is typical of the other, or that the two types of problems are correlated. What does that mean? This opens up an alternate view of "typical", which happens to be the strict definition of "correlation" in mathematical terms. ("Positive correlation", if you want to be picky.) Instead of saying that "X is typical of Y" when X shows up more than 50% of the time that Y is present, we might say that "X is typical of (correlated with) Y" when the presence of Y makes us more likely to believe that X will show up than we would think otherwise.

Suppose that, other than the 100 cases, there are 10 more clashes that fall into the category "other". Now, if I make a slip of paper for each case, throw them all in a hat, and draw one at random, what's the chance that it will be an N/S clash? Answer: 60/110 = 55%. But what if you draw one at random and tell me that it's a "channel/drama" clash--what is the chance that it's also N/S? Answer: 40/80 = 50%.

Let's look at it from the other direction. Draw a random case: what's the chance it will be a "channeling/drama" clash? Answer: 80/110 = 73%. Now draw a random case and see that it's an N/S clash. The chance that it's also a channeling clash goes down to 40/60 = 67%

In other words, if these 110 cases are representative of all your gaming experience, then seeing one type of clash should make you less likely to suspect that the other type is present. The two types of clash really aren't "typical" of each other at all, from that perspective. At best, they're just "typical" of the sorts of clashes that turn up in gaming.

But wait a second. Suppose that, instead of 10 "other" cases, there are 100. If you draw a random case now, the chance it'll be "channeling/drama" is 80/200 = 40%. But if you know it's N/S, the chance it's also "channeling/drama" goes up, to 67%. Conversely, the chance that a random case will be N/S is 60/200 = 30% if you don't know anything else about it. But if you know it entails "channeling/drama", there's a 50% chance that it's also N/S.

Now, even though seeing a "channeling/drama" clash shouldn't make you think that there's a better-than-even chance that the other type is also present, we might still say that they're "typical" of each other--certainly, they're related enough that seeing one of them should make you more likely to suspect that the other is present.

Well, I have to run. This entry was originally going to be "Notes on Correlation and Causation", but I'll have to take that up in the comments.

So, there was a thread which discussed how The Riddle of Steel should be classified in GNS terms, and then a followup thread that also got into the same discussion.

TROS & GNS: NS? NG!
TROS & GNS (split)

My last comment boiled down to this question: Does TROS support a kind of play in which players gain personal satisfaction through application of "personal strategy and guts"? Or does TROS support a kind of play in which players gain personal satisfaction through Addressing Premise?

Implicitly: if TROS supports both kinds of play, does it instead promote Incoherence in GNS terms? If not, how? If it does promote Incoherence, how does it avoid being dysfunctional?

And explicitly: if TROS only supports a kind of play where players gain personal satisfaction through Addressing Premise, then what is the combat system doing? (Note: I only know TROS second hand and through skimming the preview rules; everyone describes the combat system as intricate and gamey; player skill will win out over character ability in a fair fight without SA's involved.) Why not just toss all that out?

You'd think all this would be highly relevant when discussing a theory that's supposed to help you design and play RPG's. But no--it's all bogus.

(Please don't post in those Forge threads.)

As GNS further evolves (or as my understanding evolves--since I've been catching up starting with the essays and then reading the Forge forums) it has become more psychological and behaviorist, and more descriptive than prescriptive. Thus the concept of Rewards moves from explicit mechanics to the effect of those mechanics on the players. However, there's a problem with the concept of Rewards as shown in the last few posts of this thread, starting with Vaxalon's.

Basically, RE and Vaxalon are correctly drawing a distinction between character-development mechanics and player-gratification. Changes to characters through experience or whatever aren't always an important source of gratification and may even be counterproductive. For example, a campaign can be blown wide open by "character advancement" above a certain level of power, particularly if "advancement" is purely in terms of personal abilities.

Conversely player-gratification may occur without any overt mechanical support. However, recognizing these facts causes problems for GNS. Why? Because the "instance of play" concept, which currently underlies the idea of a coherent CA, depends on the existence of Reward Cycles. But how can you identify a Reward Cycle in the absence of concrete Rewards? This is a particularly vexing problem when it relates to identifying Simulationism, since in searching the Forge, I find that no one really has a good grasp of what a Simulationist Reward looks like. So maybe Simulationism is (once again) defined negatively, as the CA which doesn't have Rewards?

I think there are several ways to conceptualize Simulationist Rewards within the GNS framework, but I also think they might tend to undermine the framework from within. That is, it is hard to conceive of a Simulationist reward which isn't actually Narrativist. About the only one I can think of that doesn't fall into this category is the kind of reward that might go with the "leftover Sim" I described in my last post. I.e., the Reward is when the GM says, "Okay the scenario is over. Next time, I'll have a new scenario for you to explore." Which may entail new or changed characters, or a new or changed setting/situation.

Anyway, I don't have any solid answers to this question, and it's particularly hard for me since I'm not really a GNS advocate. Here are a couple of interesting threads that touch on the issue, though:

Need for Experience/Improvement
Simulationist Reward Systems

There's also Dealing With Simulationism, in which, interestingly, several of the solutions to providing Simulationist motivations (which are sorta related to rewards) are identified as Narrativist by Ron Edwards.

Under the current formulation of GNS, what does a functional Simulationist instance of play look like?

Much or all of what's left for Simulationism is very Force-heavy, in Forge terms. That is, the players don't get to make "thematically-significant" decisions for their characters. Ron provided some examples in a thread I started over at the Forge (the interesting part is here). Essentially, the GM provides a scenario which implicates the characters. The players "buy in" to the scenario and they work through it under the GM's guidance. Clues and hooks may be triggered either by truly objective criteria or by GM-manipulation (the Moving Clue), and the expectation, the only way the game is really going to work, is for the players to take the bait as instructions to guide the characters onto the next scene. At the end you have a complete "adventure".

Non-Force Sim play consists of situations where the players make no effort to do anything "significant", and instead engage in things like minor hijinks and "personality costumery". The characters interact with each other and the environment in a manner that highlights the character conception, the setting, etc. This may be in the context of a fixed chronology known to the players (a metaplot), in which case the events of the chronology allow for an "underbelly" of incidental activity where the characters can interact with whatever is going on, but are guided by players who understand that they can't change the chronology. I suppose in a sense, this might be viewed as form of Force exerted by the chronology-text.

When thinking about all this, it's important to bear in mind that Forge-theory treats CAs as self-reinforcing modes, not as goals or prescriptions for play. If the players wander off the reservation and decide they want to control where the campaign is going, you no longer have Simulationism. Instead you probably have Incoherence (particularly if the GM doesn't go along) or Narrativism.( Read on for examples and further analysis... )

Half the struggle of coming to terms with GNS is understanding what the vocabulary means. At the moment there are several Forge participants, with long posting records, who've suddenly discovered that their understanding of the term "Creative Agenda" doesn't match what Ron Edwards is saying. Whatever other factors may be involved, it's obvious that a great deal of difficulty, particularly for newcomers, is the disconnect between everyday meanings of key terms and the way they're used in GNS. Some see this as due to the inadequacy of language for describing complex concepts. That may be. However, in my brief time on the Forge, I've come up with some alternate vocabulary that works well for me.

I find the term "agenda" is terribly misleading for defining the class of -isms including Gamism, Narrativism, and Simulationism. You can refer to whatever dictionary you like; if from "agenda" you get anything having to do with a goal, a decisionmaking criterion, a plan, or a personal preference--forget it. Those interpretations of CA keep cropping up and Ron keeps shooting them down. The Forge Glossary itself offers, as a definition of Creative Agenda, "The aesthetic priorities and any matters of imaginative interest regarding role-playing." "Aesthetic priority" may be a better term but it's a bit vague, not to mention the fact that it can be construed as referring to consciously-determined priorities--again, something that Ron rejects. So what is he talking about when he says "Creative Agenda"?

Based on my reading, a CA is actually a mode or state of a dynamic system. The system in question is a group of RPGers playing a roleplaying game. To be a little more precise, there are some modes which Ron does not consider to be CAs. The first is Incoherence, and the second is Zilchplay (apparently a recent development in GNS theory). What distinguishes a CA from other modes is the presence of a positive feedback through the game as well as the players which makes the mode coherent and self-sustaining. The stronger and more focused this feedback, the more clearly a CA is expressed.

(In the GNS essays and probably elsewhere, "expressed" is used in a manner akin to how it's used in genetics, to mean "manifested". This isn't surprising given Ron's profession.)

The nature of the feedback, or the part of the game on which it focuses, defines the CA. If the group consistently takes interest in approaching and dealing with moral issues in the game, then the CA is Narrativism. If they consistently take interest in strategizing to overcome challanges, then the CA is Gamism. If the group isn't focusing on those activities, and instead focuses only on portraying the fictional elements of the game-world ("exploration on purpose"), then the CA is Simulationism. "Incoherent" play literally doesn't cohere--it's chaotic and subject to collapse. Zilchplay, on the other hand, isn't really focused on the game at all--it's just "being there".

In order to account for the fact that the presence of an overt CA may not be immediately apparent, as well as for the fact that different modes may occur at different times, GNS theory posits that a CA exists only during an "instance of play" as opposed to an "instant" (such as a decision point) or an entire game. And recently the "instance of play" has been defined in a manner which is consistent with the notion of CAs as stable feedback modes within a system. Specifically, an "instance of play" has been defined as "a complete reward cycle". I gather this refers to the time from establishing what we're going to do, doing it, then getting some sort of positive reinforcement out of the content of play. And I suspect that the positive reinforcement necessarily relates back to the CA--i.e., it enhances and provides the opportunity for doing more of that CA. So a Gamist reward would both give you more tools to strategize with and offer you the chance to strategize against new challenges. A Narrativist reward would give you the chance to further address premise (possibly by complicating your character or the situation in interesting ways). A Simulationist reward would establish certain facts or memes about the imagined reality as a platform for further dynamic, focused exploration ("what if?").

Now I would note that the implementation of the "reward cycle as instance of play" concept hasn't completely resolved all ambiguity. For one thing, there's a bit of confusion between "reward" referring to a thing which occurs in play (gaining a level, completely addressing a Premise, finishing a "chapter" of exploration) and the emotional response to in-play activities. If a "reward" is the former, wouldn't that mean that a game system could be inherently Gamist, etc., rather than "Gamist-supporting", which is the usual formulation? But if a "reward" is an emotional response, there may be dozens of rewards in a single session. What if they vary--a sense of achievement from intelligently overcoming a challenge, which sets up an opportunity to learn more about the game world, which in turn leads to an interesting moral decision? The answer which has been advanced is that you have to see if these small reward cycles are contained in a larger cycle.

Whether that is satisfactory either from the standpoint of theory construction or validation will have to be the subject of further discussion. But in the meantime, I'd be interested in any comments. How does my description accord with your understanding of GNS? If you've been mystified by GNS, does this description clarify things?

(Note that you don't have to agree with GNS to understand it. But you have to understand it in order to agree or disagree with it. At the moment I'm understanding more and disagreeing less, but you might guess from my first post, I doubt that full understanding is going to lead to full agreement on my part.)

Over at his design blog, Matt Wilson has a post where he gives his take on GNS. I think it's a good summary, and it concludes,

The trouble with S and N is that there's kind of a gray area regarding how much of the character traits are assumed up front. You could play Buffy as S, basing the characters' decisions on everything we know about them, or you could play Buffy as N, with the belief that the characters are dynamic, and there may well be a time when Buffy would choose to kill a human being, for example, even though she never does otherwise (I think. If I'm wrong, replace this with a different example).

I like it that folks like John Kim and that Marco dude are picking and poking at it, questioning various boundaries (that gray area between N and S is worth a lot of poking), but if you're one of those boobs calling the whole thing crap, chances are you just fear what you don't understand, like being the only person in the room who can't see those magic eye pictures.

I'm another one of those people who's poking at the GNS boundary, though I certainly haven't been doing it as long as John or Marco. The problems at the boundary have several dimensions but many of them center around the fact that GNS is presented as both a descriptive tool and a prescription for good play. And I think the implications of those problems go a long way to explaining why GNS is often dismissed or attacked.

At the descriptive level GNS contains an "exclusivity rule" that identifies the Creative Agendas based on the kind of fun being prioritized (for which, Matt's post is a good summary which also agrees with my understanding). Where one of the elements is prioritized, GNS says that a CA is present; the GNS term for play where there's no clear priority is "Incoherent". Furthermore, GNS defines the Sim-Nar border in such a way that you can never speak of Simulationist play supported by narrativist techniques: if there is ever a "human-centric" conflict in play, then the Creative Agenda is Narrativism.

If that was all there was to it, there wouldn't be much of a problem. One could argue with the terms, but if you change them the categories are still well-defined according to their descriptive characteristics and the "exclusivity rule".

However, as a prescription for enjoyable play (and as a design tool for creating game systems), GNS runs into problems. First, the categories themselves become controversial since they do not distinguish between play where the human conflict is strongly prioritized and play where the human conflict is buried deeply in the situation or is emergent. Second, GNS claims both that mixed ("Incoherent") play is particularly vulnerable to dysfunction (i.e., it's likely to be dissatisfying) and that games which focus on a specific, identifiable CA are more likely to provide consistent fun. By "focus", what is apparently meant is that elements of the non-prioritized CAs are treated, as much as possible, only as tools to support the prioritized CA.

By "deeply buried" conflict, I mean conflict that provides the stakes for evaluating the outcome of a situation. Compare a boardgame, particularly a wargame, which has clear win/lose conditions. Often such games hinge on an artificiality such as maintaining a single combat unit in a certain town; if you "win" by meeting the condition but you otherwise suffer enormous losses relative to the enemy, did you really "win"? To the extent that such a question is meaningful, one would have to look at the larger context and subjectively evaluate the overall priorities: just how important is maintaining the position relative to sustaining casualties? If you maintain the position and suffer light casualties, but the enemy advances well into your rear in full strength, does it matter? If you fail to maintain the position but you inflict lopsided damage on the enemy, might that aid your cause?

One of the attractions of RPGs is the way that such evaluations are often the only way of determining success or failure. Thus a situation where a SWAT team has clearly defined goals and interests, seemingly a straightforwardly Sim-ish hostage rescue mission, nevertheless contains conflicting priorities: save the hostages; capture the bad guys alive if possible, dead if necessary; minimize own casualties. Grouping the play of such a scenario under Nar just because there are embedded "human-centric" conflicts may completely overlook the fact that the group mainly gets its kicks from simulating the logistical, command, communications, and intelligence problems associated with carrying out the mission. The only reason the group chooses not to use explicit "victory conditions" to precisely define the various priorities (and thus answer all questions beforehand about what to do in any situation) is that such a thing is practically impossible. Therefore it does not seem useful to make a distinction that separates this kind of play from more straightforward Simulationism.

Similarly, GNS says that Narrativism encompass the address of "emergent human-centric conflict", no matter how small or how quickly dealt with. Consider a putatively Simulationist situation, with high emphasis on, say, portraying the struggles of a jungle expedition. At some point, an accident might present a player-character with a decision regarding how far he should go risking his life to recover someone who fell into a ravine. The definitional reason for classifying this as Narrativism is clear; what is not clear is the utility of such a classification.

In fact, due to certain prescriptive elements of GNS, the classification may not only be of limited utility--it may be positively harmful. As noted above, GNS claims that Incoherent play is particularly vulnerable to dysfunction and that games which focus on a specific, identifiable Creative Agenda are more likely to provide consistent fun. It naturally follows that even if some of the borderline Nar/Sim cases were classified as Sim, GNS would prescribe reducing or eliminating the "human-centric conflict" elements, perhaps (in the extreme case) yielding a kind of simulation wargame devoid of the subjective/imponderable evaluations that go into decisionmaking. On the other hand if the borderline cases are classified as Nar, GNS would recommend de-emphasizing the highly simulative elements of play in favor of focusing on and heightening the "human-centric conflicts". This is one aspect of GNS which I think tends to annoy roleplayers who enjoy precisely the type of play that emphasizes simulation but uses "human conflict" as a way to frame the stakes and subjectively evaluate outcomes. When GNS theory is qualified by saying that "unfocused" or "Incoherent" gaming isn't necessarily dysfunctional, merely "more vulnerable to dysfunction" or "less likely to provide consistent fun", the best these "borderline players" can conclude is that GNS is irrelevent. Without those qualifications, the theory is patronizing or even offensive ("badwrongfun") to "borderline players".

Now, I know that Ron bases GNS on a great deal of observation--so perhaps his experience does show that focus within the CAs as defined by GNS is the surest way to producing enjoyable gaming. There are even people who ascribe improvements in their gaming to learning about GNS focus. But I'll bet there are a lot of people out there who are enjoying the heck out of Incoherent designs and Incoherent, unfocused play. So at this stage, I view GNS more as a manifesto or a craft-manual than as a rigorously tested theory. The prescriptions may show you ways to make great games or run great game sessions, but to me it's not clear at all that they won't also steer you away from other great gaming experiencees. In fact, by admitting that Incoherent and unfocused gaming can potentially be just as enjoyable as CA-focused gaming, GNS implicitly contains a challenge to designers and gamers: develop the tools to reproduce this kind of enjoyment.