Control and Restrictions in RPGs

[M. J. Young]

Welcome to The Forge, Lucian. I'm pleased my article was helpful.

I'm jumping in while only half finished reading this, so I won't forget what I'm thinking (I do that sometimes, particularly reading long threads); so forgive me if I duplicate someone.

I'm looking at your exposition of plot and character for Cinderella, and it seems to me that your view of role playing is always either participationist or trailblazing; you're so locked into this, that you fail to see how very little control over "character" you're really allowing the character player when you give control of "plot" to the referee player.

Let's take the plot of Cinderella:

Setup:  Cinderella lives with her stepmother and stepsisters.  They make her work like a maid and do little work themselves.
Conflict:  There's a ball, and they don't let Cinderella go.
Rising action:  Cinderella tries to get to the ball, and is variously thwarted and helped along the way.
Climax:  The prince find out who she really is
Denoument:  They get married; the stepmothers/sisters get their just deserts.

In contrast, here's the character of Cinderella:

Who she is:  human, female, beautiful
What she can do:  work really hard, dream, dance
What she does:  Makes a dress in her spare time, dances with the prince, leaves the ball at midnight.

The problem is that What she does has nothing to do with her character; it is entirely about the plot.

Thus your referee has set up

• Cinderella lives with her stepmother and stepsisters.
• They make her work like a maid and do little work themselves.
• There's a ball, and they don't let Cinderella go.

Now he expects the player to decide

• Cinderella tries to get to the ball, and keeps trying until she succeeds.

Everything else is contingent on the player's decision that his character will do this; but that means it is not the player's decision at all--it is the referee's decision dictated to the player.

In this exact same situation, we could have:

• Cinderella sits by the fire crying, and commits suicide.
• Cinderella kills her stepmother and stepsisters in their sleep, takes their good clothes and jewelry, and goes to the ball.
• Cinderella finds some friends who are equally put out by this oppressive class system, and they assault the castle during the ball, taking the prince prisoner, killing chief nobles, and establishing a socialist state.
• Cinderella runs away from home and starts a life as a gypsy fortune teller. She marries a traveling bard, and they live happily ever after.

All of that would be things the character could choose to do, and thus would be about controling the character; however, your design dictates that the referee controls the character at the critical points--the only critical points--of deciding what the story is going to be about.

The story shouldn't be about what the referee's plot plans are unless that's already agreed in advance by the players (participationism or trailblazing). It should be about what the character chooses to do. Thus if the player truly does control the character's choices, then the player controls the plot, and there's nothing the referee can do about it but throw a fit and refuse to allow actions the player wants to pursue.

A few years back we were playing Multiverser in our online forum, and someone suggested that I demonstrate the system's flexibility by running several players at once independently through the Kevin Costner film, The Postman. I started each of them by "versing them in" (characters who have died in one world suddenly find themselves alive in the next, just as they were) to an abandoned gas station where some mule had gotten its harness caught trying to get water from the rain barrel. From there, I cut them loose.

• Eric "Tadeusz" Ashley gathered information about what was happening, then proceeded to harrass the Army of Eight as a one man commando unit, adopting the name The Ghost until he had them so terrified that one night he walked into their camp unopposed and killed the general in single combat, and claiming that he was now in charge.
• Graeme "Blue Canary" Comyn put up no resistance when conscripted into the Army of Eight, but once inside he began quietly fomenting rebellion. Using his engineering skills and manipulating public opinion within the army, he managed to get the general to ask him to build a public shower and laundry, within which he hid a printing press and paper maker. He began recycling paper and printing one-page newspapers opposing the Army, which his sympathizers smuggled out of camp to the villages only to have them brought into camp by soldiers to further raise internal opposition to the way things were run. Ultimately, the soldiers rose up and overthrew the general, and started rebuilding society.
• David Marcoe realized what was happening, and went as quickly as he could away from the territory of the Army of Eight toward Washington, DC. Along the way he gathered talented people and scavenged equipment with a view to creating his own small army. He was killed in a fire fight about half way there, but before that he built a tideswell for the reestablishment of a constitutional democracy in the United States.

There were one or two other players who took different approaches, each of which resulted in an interesting game, but each of which had an entirely different plot.

The plot was determined entirely by the choices made by the principle characters.

That is always what determines plot. If the referee controls plot, he is de facto controling the choices made by the characters, whether doing so directly through illusionist techniques (e.g., it doesn't matter which way the character goes, the Army of Eight will catch him) or through social contract (e.g., the point of this game is for us to be captured by the Army of Eight to start, so let's go get captured so the story can begin).

So you think you're giving control of character to the character players, but you're reserving the most important aspect of that control to the referee, the decisions that determine what is going to happen.

It's a good start, by the way. Thinking about the elements and trying to isolate them can only lead to a better understanding of the process.

--M. J. Young

[lpsmith]

First off, let me say this has been very helpful in clarifying my thoughts.  Then let me disagree with you again ;-)

I'm definitely *not* saying that roleplaying is always either participationist or trailblazing.  That would be kind of a dumb thing to say, since clear examples to the contrary abound.  What I'm trying to say is that participationist and trailblazing play *can* happen when the GM is solely in control of what-I'm-calling-plot.  Does it help if I say that I believe 'bass playing' can *also* happen when the GM is solely in control of the plot?  And that if both players and GM are given control of the plot, you can have 'bass playing' but you can't have participationist or trailblazing play?

That is always what determines plot. If the referee controls plot, he is de facto controling the choices made by the characters, whether doing so directly through illusionist techniques (e.g., it doesn't matter which way the character goes, the Army of Eight will catch him) or through social contract (e.g., the point of this game is for us to be captured by the Army of Eight to start, so let's go get captured so the story can begin).

Well... no.  If the GM controls plot, he is controlling the *outcomes* of the choices made by the characters.  This is a crucial difference!  Let's say we're playing a traditional (classic?) system where the players have absolute control of my 'character' axis, and the GM has absolute control of my 'plot' axis.  If I'm playing Cinderella and I decide to try to commit suicide, that *intent* cannot be overruled by the GM.  I'm in control of Cinderella!  In this story, Cinderella is *the kind of person who tries to commit suicide*.  That's tremendously important!  I'm crafting a character here, and this is 'My Gal'.

Now, the success or failure of my suicide attempt is up to the GM, with the caveat that he has certain restrictions placed on him in the form of 'you must play by the rules'.  Surprise!  The gun wasn't loaded.  There was a pile of hay at the base of the tower.  The fairy godmother shows up (a blatant GMC if I ever saw one) and hits you with her wand.  Or not!  Cinderella succeeds, and dies tragically alone and forgotten in the east wing.  Alternatively, she succeeds, and the tale of her death reaches the prince, who is so moved that he outlaws evil stepmothers right then and there.  Whichever.  My point is that if the GM is in control, the GM is in control, to do with the plot what he wishes.

I'm starting to think I shoud re-name my 'Character' axis 'Characterization'.  Intent is characterization.  Outcome is plot.  Good intentions are character, hell is plot.

There's an improv game/exercise called "Neutral Bench".  The rules are simple:  Two players start off sitting on a bench.  One has the 'OOC' goal of getting off the bench.  The other has the goal of keeping the two of them in the same place.  Everything you say is true, and must not contradict anything that was already said.  Go.

With a bit of tweaking, you could play a form of this game in roleplaying.  One person wants to commit suicide, and the other person wants the first to marry the prince.  Everything the first person says about what they do is true.  Everything the second person says about everything else is true.  Go.

(Hmm, that could actually be fun.  Anyone up for a Railroading Invitational?  You try to derail the plot, I try to keep things on task.  Pure Gamism ;-)

You listed three different plots that stemmed from the same intial setup (I haven't seen the actual movie), and they all sound like a lot of fun.  They also sound like the GM (you) decided to use the players intentions and actions form the basis of a series of situations that moved closer to and further away from a climax that was dreamed up on the spot.  I don't know why you decided to to this.  Perhaps the system demanded that you do so (I'm unfamiliar with Multiverser).  In that case, control of the plot was at least partially granted to the players by that system.  Perhaps the Social Contract demanded that you do so.  In that case, the control of the plot was either granted by you to the players, or you simply agreed to use your powers for good (Social contracts are slippery beasts, which is one reason I left 'em out of my analysis).  Perhaps you're just That Kind Of Guy, and that sort of play is the most fun for you.

If the system did not demand that you give control to your players, though, you could have decided to railroad 'em all.  You would perhaps have had to be very creative in your railroading, and the less creative you were the more ham-fisted you would perhaps have had to be, but it *could be done*.  You control the plot.

But at the same time, you do not control the characters.  If Eric wants to gather information, you can tell him 'you find nothing', but that doesn't change the fact that Eric has created a character who tries to gather information before taking action.  If Graeme wants to surrender, you can make a rival group suddenly appear on the horizon and rescue him, but that doesn't change the fact that Grame has created a character who will surrender to a superior force.  If David wants to run away, you can have him hunted down, but that doesn't change the fact that David has created a character who runs away when he hears the first hint of bad mojo in the area, with the plan to come back when better prepared.

Now, having your character affect the plot is clearly very important to you:

So you think you're giving control of character to the character players, but you're reserving the most important aspect of that control to the referee, the decisions that determine what is going to happen.

You used words like 'most important' a lot in that essay, and it finally grated on my nerves enough that I wrote my essay in response.  If it's important *to you*, you need to be in a gaming group where your input directly affects the plot, either through the mechanics of the system, or through the largesse of the GM.  But if the system grants that power to the GM, your only recourse is the Social Contract.  In a straight system analysis (as I attempted to structure my essay to be), the GM simply controls the outcomes in many (not all!) systems.

Let me bare my soul a bit here.  In my actual life, character is more important to me than plot.  Who I am is tremendously important to me--fortunately, it's what I can control.  What happens to me I can *hope* to influence by who I am, but in the end, the race is not to the swift, nor the battle to the strong, nor bread to the wise, nor riches to the intelligent, nor favor to the men of skill; but time and chance happen to them all.  I can *be* swift.  I can *be* strong.  Whether I win the race or finish the battle is outside of my purview.

After a long, painful, 10-year process, I got a PhD.  Did I get a PhD because I'm a tenacious SOB who finishes what he starts?  Or did I get a PhD because I'm an idiot who almost destroyed his life because he wasn't strong enough to walk away from a bad situation and cut his losses?  Those questions still haunt me.  I could hardly care less about actually having a PhD.  I am who I am by the choices I make.  For the plot, I have to trust the GM.

Hmm, this is turning into a testimonial.  Better wrap things up.

My intent with my essay was not to say, "This is what roleplaying must be."  It was to say, "These are aspects of control that shape and limit the range of possibilities for what roleplaying in a particular system can be."  It doesn't completely answer that question, but it was a start.

[Josh Roby]

I'm starting to think I shoud re-name my 'Character' axis 'Characterization'.  Intent is characterization.  Outcome is plot.  Good intentions are character, hell is plot.

I'd go a step further and say that that axis that you want to give to players is 'Intent' and relabel the axis normally controlled by the GM as 'Outcome'.  At which point, you're in IIEE territory.  I think you're getting closer to what you're trying for, but personally, I think you're kind of painting yourself into a corner, describing how gaming could work, under a single, given (if common) social contract.  I'm not sure you're going to be able to draw any useful conclusions from this.

Lastly, I'd challenge -- rather stridently, in fact -- your assertation that Intent is Characterization.  Intent is a pretty pale thing next to Characterization.  Character involves a thousand things outside of the individual's intentions -- accidentally killing a friend, for instance, has huge impact on characterization.  Trying and failing to stand up for one's principles (failed Will check) says more about who the character really is than whatever it was he "intended" to do but didn't have the internal resources to actually be who he wanted to be.  Character is entirely about what you do.  Good intentions are fluff.

[Marco]

I'm starting to think I shoud re-name my 'Character' axis 'Characterization'.  Intent is characterization.  Outcome is plot.  Good intentions are character, hell is plot.

I'd go a step further and say that that axis that you want to give to players is 'Intent' and relabel the axis normally controlled by the GM as 'Outcome'.  At which point, you're in IIEE territory.  I think you're getting closer to what you're trying for, but personally, I think you're kind of painting yourself into a corner, describing how gaming could work, under a single, given (if common) social contract.  I'm not sure you're going to be able to draw any useful conclusions from this.

The (traditional) GM-Player interaction set is very complex. Even in a group that has a good history of functional gaming, the actions and interactions between GM and Players may change a lot depending on circumstances and edge conditions during play (i.e. a GM may normally be very pin-ball sim, to use a term I think may be depricated, but might someitmes take on a heavier hand depending on energy level of the group, present mix of players, real-life conditions, etc.)

As such, any generalization is going to have problems. For example Ron's recent assertion that two CA's are as different as one guy who rides motorcycles and one guy who uses them in his abstract art is an example of this: the two CAs look almost exactly alike and the theoritically, in this frame work, most cycle racres do, at times, enjoy turning their bikes into abstract art. NOTE: while I can rarely figure out exactly what Ron is talking about without a scorecard, I'm pretty sure that his example was for a very constrained set of circumstances and conditions rather than a total overarching statement. Even so: any attempt to be specific usually results in someone painting themself into some kind of corner.

Lastly, I'd challenge -- rather stridently, in fact -- your assertation that Intent is Characterization.  Intent is a pretty pale thing next to Characterization.  Character involves a thousand things outside of the individual's intentions -- accidentally killing a friend, for instance, has huge impact on characterization.  Trying and failing to stand up for one's principles (failed Will check) says more about who the character really is than whatever it was he "intended" to do but didn't have the internal resources to actually be who he wanted to be.  Character is entirely about what you do.  Good intentions are fluff.

I think this is an example of splitting hairs a bit too finely. I think intent is one of the forces behind player Input. A player's mental model of characterization (i.e. how I precieve the character) is part of this as well. What the character "actually does" is the result of a lot of factors (and, indeed, may be vague in a lot of cases or imagined differently by different people).

If you are not the player then characterization is about what the character does and what you precieve the player thinks. If you are the player you may have  more data to work with and see things differently (and things like a failed will check can result in a disconnect between the mental model and the SIS rendering of the character--I don't think it's any more proper to say that one is more about who the character "is" in a general sense since both are imaginary constructs).

-Marco

[Josh Roby]

I don't think it's any more proper to say that one is more about who the character "is" in a general sense since both are imaginary constructs.

One conception of the character is given credibility by the system; the other conception is not.  Characterization is something that is irretrievably fractured by point of view.  Player A will think of Character A in one way, Player B will think of Character A in another way.  While those characterizations can be reconciled, the way that happens is through the game system (which may or may not be accurately described in the written rules) and the way that it apportions credibility.  Under this model, the control of the plot -- the outcome, the effect, the "what happens" -- gives absolute credibility over that reconciliation process.

Is a character a hero if he really wants to stand up and fight for the innocent, or is a character a hero when he does stand up and fight for the innocent?  Heroism is doing, not wanting to do.  The same thing applies to the rest of characterization; the why is irrelevant without the what.  To say that the credible statements made during play are meaningless to characterization begs the question why the system that provides those credible statements is being used at all.  Why play with other people and roll dice?  Just write out the character profile as a short story.

To drag this closer to topic, I think the "typical" railroading/illusionist/participationist play that is described by the model assumes that the GM will provide the what while the players provide the why.  That isn't the same, however, as saying that the GM provides the plot and the players provide the characterization.  Character and plot are inextricably linked, and absolute control over one means near-absolute control over the other -- you can't give absolute control of characterization to one person and absolute control of plot to another person.  Their spheres of control will immediately be in conflict.  You can, however, give absolute control of intent to one person and absolute control of outcomes to another.  No conflict arises there.

[Darren Hill]

You can, however, give absolute control of intent to one person and absolute control of outcomes to another.  No conflict arises there.

Going further, I'd argue that if you give control of outcomes to one player, then - over time - that means no other player can have absolute control of intents.
Players learn that X will always fail, so they stop intending to do X. Then they learn Y won't work, so they cut that out of their possible options. Gradually, the range of possible intents shrinks to those the players have learned can lead to (or think might lead to) achievable outcomes.
One player having absolute control of oitcomes will shape the kind of intents allowed to everyone else.

[Josh Roby]

Control of outcomes does imply control of context, true, but this is at best an indirect control -- potentially powerful, sure, but by no means 'absolute'.  Players can always continue to beat their heads against the wall, after all.

[lpsmith]

Hmm, we're veering into philosophy here.  But anyway.

My little "Intent is Characterization" aphorism seems to be causing some confusion from what I meant--'intent' probably isn't the right word.  I didn't mean 'stuff you think about doing', I meant, 'stuff you try to do'.

Which changes the question to "Is a character any less of a hero if he stands up for the innocent and fails, or must he succeed?"  And you will actually get different answers to that question depending on who you ask, but *my* answer (and, I suspect, the answer of others who 'don't care about' plot) is 'No'.  And that gets at the difference between what I meant by character and what I meant by plot.  And I continue to assert that they are separable.

However, I'm clearly not communicating something, because while I see what I've written as a perfectly extensible way to dissect a wide variety of roleplaying models (participationism, trailblazing, actual module play, bass playing, battlegrounding, timetabling, relationship mapping, etc. etc.) you all keep thinking I'm only talking about participationism.  Why is this?  What did I say?  Is 'control' overloaded?

To clarify:  I'm trying to talk about *systems*, not Social Contracts.  Any social contract must work with what it is given by the system.  If a system gives control of 'plot' (/outcomes?) to the GM, the GM can then bequeath control to the players through the social contract.  But you cannot keep for yourself something you do not have, and if the system gives control to the players, the GM no longer has the power to run a Participationist game--if that's what everyone wants to play, the players would have to give control *back to the GM*.

Maybe this is enlightening:

Going further, I'd argue that if you give control of outcomes to one player, then - over time - that means no other player can have absolute control of intents.
Players learn that X will always fail, so they stop intending to do X.

Why do you think that my saying 'the GM has control' means 'the players will always fail'?  What's up with that?  A perfectly valid mode for a GM to follow is to make the players always *succeed*.  I'm only saying that it's the GM's choice which mode to follow.

And this is also a sort of red herring.  Success or failure, does the action result in a change in the situation that is closer to or further away from the ultimate climax?  (Keep in mind that the climax may itself be shifting or not even exist yet.)  That's even more up to the GM than success or failure of particular actions are (often because success/failure of 'small' actions are determined by die rolls).  This is the foundation of why participationism and trailblazing work at all, but (I contend) is *also* at the foundation of certain flavors of 'bass playing'.  The GM is making different types of decisions, that result in different play experiences.  The players don't have to do anything different at all.

I contend that a sufficiently creative and/or ham-fisted GM, if granted control of the 'plot' by the system, could make player action result in situation change towards a pre-defined climax even if the players were intentionally trying to avoid that climax and/or stall.  Whether this would be any fun is an open question.

[Marco]

Which changes the question to "Is a character any less of a hero if he stands up for the innocent and fails, or must he succeed?"  And you will actually get different answers to that question depending on who you ask, but *my* answer (and, I suspect, the answer of others who 'don't care about' plot) is 'No'.  And that gets at the difference between what I meant by character and what I meant by plot.  And I continue to assert that they are separable.

Right--exactly. And while there's a great deal of credit given to the idea that the game system "resolves" or "reconciles" player's ideas about what their character's characterization is, I don't think this is as true as the theory would like it to be. If my character is a doctor with a decent skill roll but the three or four times he tries to save someone the dice come up really unlucky, some people will say he's an incompetent doctor despite what my internal characterization is and (correctly, IMO) continues to be.

The theory sounds good--it doesn't apply to reality: if you ask the players at the table they may, rightly, feel nothing has been reconciled through mechanics or system (in fact, if the player(s) decide that the mechanics are out of line with system then the two characterizations remain fractured).

This is why saying "Just write out the character profile as a short story" is missing the point: a lot of things, not just system and not just mechanics go into the in-game representation of the character (one of these is 'situation'--which is a major component of what, I think, becomes 'plot' in this construct--and may or may not be counted as a part of 'system' depending on who you are and how far you want to take it).

Ultimately there isn't a universal concept of "characterization" in an RPG (or, perhaps, anywhere else--but very strongly in an RPG). You have to specify a person holding that view. As noted, I might very much differ from you about who is a hero and where heroism resides in a given set of character interactions. The idea that one of us is objectively wrong about that, across a broad spectrum of possibilities, is hard to legitimately hold.

-Marco

[Josh Roby]

My little "Intent is Characterization" aphorism seems to be causing some confusion from what I meant--'intent' probably isn't the right word.  I didn't mean 'stuff you think about doing', I meant, 'stuff you try to do'.

That's all well and good when the System operates after Initiation in the IIEE sequence, but this isn't always the case.  There are a fat lot of Systems out there where player intent does not necessarily equal character attempt.  The failed Will check is a prime example.  If in GURPS I fail my Will roll against Cowardice, I didn't "try" to save the people in the burning building and decided not to, I cowered in the corner.  You're assuming that the System provides this divide between "character" attempt and "plot" outcome, when this is not necessarily the case.

I'm clearly not communicating something, because while I see what I've written as a perfectly extensible way to dissect a wide variety of roleplaying models ... you all keep thinking I'm only talking about participationism.  Why is this?  What did I say?  Is 'control' overloaded?
...
To clarify:  I'm trying to talk about *systems*, not Social Contracts.  Any social contract must work with what it is given by the system.  If a system gives control of 'plot' (/outcomes?) to the GM, the GM can then bequeath control to the players through the social contract.

You attempt to show that control of character and control of plot can be separated (the Impossible Thing Before Breakfast), which is pretty much assumed as an untenable position around here.  Railroading/Illusionism/Participationism is where that breakdown occurs, which is why most of the comments are focused on it.    Additionally, Social Contract is typically portrayed as the level "above" System.  In fact, Social Contract determines System in the usual configuration -- your group of players decides what game to play, after all.  So the Social Contract doesn't work with what the System provides, the System works with what the Social Contract allows.  Also, when you say "System" are you using it as Lumpley Principle defines it, or as a one-word shorthand for "The rules as presented in the published material"?

Why do you think that my saying 'the GM has control' means 'the players will always fail'?  What's up with that?

When you say there is control in two places, that is true right up until those two places disagree -- there is always some rule (explicit or not) that says when one takes precedence over the other, when one gets to say they're right and the other one is wrong.  Otherwise the players would just sit there staring at each other.  When the GM exerts his control to stymie player initiative, you have the Impossible Thing Before Breakfast.

Here's my take on your model as presented -- you say that Character and Plot are two separate axes and you can successfully grant control of each to different people.  When you say 'Character', though, you mean 'Player Intention expressed through a character' which you are used to being the same thing as Character Intention.  This is a big assumption that is not universal.  If you model is built on assumptions that are not universal, then your model is not universal.  Additionally, when you say 'Plot' you mean 'Events that happen in the game' and perhaps 'An overarching sense of coherent motion towards a climax'.

What you have is a model of how roleplaying usually works in most cases in traditional games, which may or may not be a useful tool for understanding roleplay that works outside of this scheme (in games like Capes, Polaris, or Under the Bed, all of which mix these elements up).  On the other hand, I can easily see your model reconfigured to talk about 'Roleplaying Tasks' in general, and who has responsibility for what.  Those outlier games would simply organize those tasks in different schemes than your four axes.

[Blankshield]

Maybe this is enlightening:

Going further, I'd argue that if you give control of outcomes to one player, then - over time - that means no other player can have absolute control of intents.
Players learn that X will always fail, so they stop intending to do X.

Why do you think that my saying 'the GM has control' means 'the players will always fail'?  What's up with that?  A perfectly valid mode for a GM to follow is to make the players always *succeed*.  I'm only saying that it's the GM's choice which mode to follow.

And this is also a sort of red herring.  Success or failure, does the action result in a change in the situation that is closer to or further away from the ultimate climax?  (Keep in mind that the climax may itself be shifting or not even exist yet.)  That's even more up to the GM than success or failure of particular actions are (often because success/failure of 'small' actions are determined by die rolls).  This is the foundation of why participationism and trailblazing work at all, but (I contend) is *also* at the foundation of certain flavors of 'bass playing'.  The GM is making different types of decisions, that result in different play experiences.  The players don't have to do anything different at all.

I think Darren shortcut past an important step (or I'm reading his message wrong).  
If the GM is the one who decides success or failure, over time a pattern will arise as to what the GM 'likes' and allows to succeed, and what the GM 'dislikes' and causes to fail.  The players will notice this trend, and stop doing things the GM doesn't like, because they know it's a dead end.  They'll do what the GM likes, which is what moves the story along - in essence, admitting their choice is an illusion.

James

[Josh Roby]

And while there's a great deal of credit given to the idea that the game system "resolves" or "reconciles" player's ideas about what their character's characterization is, I don't think this is as true as the theory would like it to be. If my character is a doctor with a decent skill roll but the three or four times he tries to save someone the dice come up really unlucky, some people will say he's an incompetent doctor despite what my internal characterization is and (correctly, IMO) continues to be.

If you are talking about the Interaction Model like it looks like you are, first off, that's a different thread and if you want to spawn one off, I'd be happy to be bombastic in it.  That said, the Interaction Model says that roleplaying works towards reconciliation, but does not ever really reach 100%-match.  Some players may think your doctor is inept; others may decide he's had a very bad day.  The point is, they're all drawing conclusions about the same imaginary doctor.

I might very much differ from you about who is a hero and where heroism resides in a given set of character interactions...

My point isn't about character interactions; it's about the distinction between player intent and character initiation.  Someone who tries and fails to do something heroic may very well be heroic; a character who doesn't try at all isn't heroic under any circumstances.  See my IIEE comment in my last post.

[Darren Hill]

I think Darren shortcut past an important step (or I'm reading his message wrong).  

You're right, I did miss that out.

[Josh Roby]

If the GM is the one who decides success or failure, over time a pattern will arise as to what the GM 'likes' and allows to succeed, and what the GM 'dislikes' and causes to fail.  The players will notice this trend, and stop doing things the GM doesn't like, because they know it's a dead end.

...assuming the players are only interested in character success, yes.

[Darren Hill]

If the GM is the one who decides success or failure, over time a pattern will arise as to what the GM 'likes' and allows to succeed, and what the GM 'dislikes' and causes to fail.  The players will notice this trend, and stop doing things the GM doesn't like, because they know it's a dead end.

...assuming the players are only interested in character success, yes.

Remember the context. We are talking about a situation in which the GM controls outcomes, and players control intents. If the GM controls outcomes, then the GM controls what actually happens in the SIS.  It doesn't matter whether the players are interested in character success or not, what matters is whether their additions to the SIS (whether character-focussed or not) are recieving validation or being rejected. If they are being rejected, they will make assumptions about why this is, and change their play accordingly, to find inputs that will recieve validation. Players will manoeuvre to find a smaller subset of intentions that they can find fun, or they'll reject this kind of play completely.
I would expect this to be the case whether the system focusses player interest on their characters, or allows them to influence the SIS in some other way. (Though I find it hard to imagine a system with an emphasis on GM control of outcomes also having support for players controlling elements beyond their character.)