[Pitfighter] SBP: the GM's role in resolution

[David Berg]

"SBP" stands for Story Before Participationism.  It's basically gaming where (a) the GM makes certain things unavoidably happen in play as a part of getting across a "story" (by which I mean a plot, experience, or other sort of vision) they've conceived, and (b) the players are aware of this and are engaged in contributing meaningfully to that story.

The end of my "decoupling rewards from plot" thread serves as a jumping-off point for my current work in developing and cataloging SBP techniques. 

The next step I took was to explore alternatives to rolling for characters' success/failure.  In that thread, it was pointed out that the functionality of such formal mediation depends largely on how it interfaces with the GM's role in resolution.  That's what I want to talk about here.

How do we take the things that good, experienced SBP GMs do, and turn them into a system that is more reliable, clear, easy, and/or fun?

[David Berg]

In college, I ran a Story Before d20-derived game called Pitfighter.  I've described how I wound up running it in my own principled Illusionist style here. 

I think the variety of approaches I used as GM to resolve events in play covers most of the relevant bases for this discussion, so I've bolded them in the play report below. 

I hope this example can serve as a springboard to discuss approaches to SBP GMing and how to systematize them.

The scene:

The player characters Garek (John), Rafaenn (Ben), and Kwailun (Gabe) were sailing to the city of Haven, there to rendezvous with the Silent Wolves, an assassins' guild they were on uncertain terms with.  The Silent Wolves were key to my mega-plot, and I was eagerly anticipating the revelations and questions I'd get to throw at the players once they reached Haven.

One of the Silent Wolves' enemies was a strange behind-the-scenes power-broker organization known as the Diamondbacks.  Before the PCs met the Wolves, I wanted the PCs to establish some sign of their combat-worthiness and allegiance.  Accordingly, I had their boat attacked by some Diamondbacks, intending for the PCs to defeat them and escape, while noting enough distinctive features to surmise the Diamondbacks' identity.

The Diamondbacks' ship appeared out of nowhere, and their warriors quickly attached grappling hooks to the PCs' craft. 

Ben said, "I wait by one rope until the guy's almost within slashing range, and then I cut the rope with my diamond razor." 

I quickly responded, "Nice idea.  He never expected you'd be able to slice it so quickly.  He's in the water.  Well done!"  My rationale was: that's clever, sounds like it ought work, and won't ruin my plans.  So: it works!

Garek and Kwailun didn't have diamond razors, so they dislodged as many hooks as they could and then readied themselves to duel their assailants.  John said, "I'll crouch down, with my sword behind me, so the first guy over the side sees my unprotected back and leaps down at me.  And then I'll spin and skewer him."

This struck me as really cool, but also a bit chaotic and uncertain.  You wouldn't know if this would work until you tried it, right?  There should be some suspense.  "Cool, the pirate sees you, and you see him spring forward!  Roll to attack!"  It was probably around when John was picking up the d20 that I decided that his action was going to work.  (We played with hidden target numbers, so I wasn't forced to communicate this.)  It was pretty open in my head how completely it'd work and what it'd look like; so, though I knew basically what was going to happen, I let the dice guide me on the particulars.  If he'd rolled a 1 on his d20, I would have felt obligated to support the idea that the roll mattered (remember, this was Illusionism, not Participationism!) by announcing failure.  If he'd rolled anything above a 15, I would have announced that he completely skewered the guy and looked slick doing it.  In reality, he rolled a 5.  So I said, "You suckered and surprised him so much that even your crappy attempt to stab him causes him to try an awkward mid-air tumble to avoid your blade.  You don't hit him, but he crashes to the ground at your feet, losing his grip on his sword."  This NPC was still dead meat, but it'd take John another combat turn to make that happen.

(Note: this is close to how Jay's Middle Earth game plays, as described here.  Their system: 1 = horrible disaster, 20 = brilliant success plus XP, 2-19 = whatever the GM wants.)

Kwailun, meanwhile, faced off against a badass Diamondback warrior.  They both had enough hit points that no single blow was going to kill anyone, so I was content to sit back and see what the dice produced.  In general, I didn't involve myself much in combats until permanent death was about to occur.

Rafaenn levitated out of reach and cast some spells, Garek and Kwailun out-dueled their opponents, and I slipped in my distinctive Diamondbacks color, so I was pretty much satisfied with the encounter.  At some point the Diamondbacks' leader, Collan, with all his crew slaughtered, ran to the edge of the boat.  I liked the looks and persona I'd come up with, and wanted the option to use him as a recurring badguy.  John decided to chuck his sword at Collan's back while he ran.  Since this was an action with a weapon, John expected a roll, so I gave him one.  But I didn't think this would work, and didn't want it to.  So, I decided it wouldn't work, and let the die roll guide me on how to implement that result.  John rolled quite well, so I narrated that damage had been dealt but that Collan was just enough of a badass to not slow or stumble in his dash.  He then jumped over the side and was gone.

The players then came up with the brilliant idea of grabbing the Diamondbacks' now empty ship and using it to infiltrate their lair.  It was a cool idea, but would have totally derailed my planned Silent Wolves rendezvous in Haven.  So, I said, "Their boat's on fire" -- basically, no, you can't do that.

The system:

Looking at this 12 years later, my impression of this resolution system is a sort of flow chart:

Step 1:
If the GM wants an action to succeed or fail, it does. 
If the GM has a preference but wishes to include other factors, go to step 2.
If the GM has no preference, go to step 3.

Step 2:
Some additional resolution method is employed to navigate the range of outcomes that are acceptable to the GM.

Step 3:
Some additional resolution method is employed to determine the outcome.

[contracycle]

I think the above is pretty close to the way I do things too.

Just to expand a bit, what I meant by abandoning system also looks as if it shows up here: almost all the stuff OTHER THAN to-hit and damage gets taken ito my hands.  All the stuff that would be governed by maps or intitiative rolls or periodic recharge etc. gets moved to fictional positioning handled by narration. 

Poss. more later, just been interrupted,

[contracycle]

So I'll try to add a bit more.

I think one aspect of the above style is that this fight scene isn;t really there to "see if the players win" or anything like that.  It has quite different goals: it;s primarily a piece of exposition.  The GM wants to communicate with players, telling them things about the world and the NPC's in it.  It's set up as a fight scene partly because that can be a fun activity in its own right, partly becuase it allows the players to show off their characters to each other and the GM, and partly becuase interacting with the world is much more engaging than just sitting through a lecture.

So, the question of who wins the fight is pretty much already answered.  Some would think that because this is not in doubt, is not a Question being asked by the game, it should be skipped over or abstracted so that play can move on to something more interesting.  But this I think is to laregly miss the point, because the right players for this sort of style are going to be interested in "having the experience".

So the question then is how to do this elegantly, without sending false messages to the players, such as implying they can bump off the Big Villain if he or she appears.  Also, as suggested by my comments above, hen I do this sort of thing I don't like "rigid" systems that trap me into tightly governed situations.  Fictional positioning is king, and rules about how far one can move on a map and so on tend to get in the way and detract from the experiential aspects of the effect I'm trying to create.

On that note, Story Games has a current thread on "Are maps bad?"which provides some discussion that is pertinent here.

Now I was quite intrigued by the fairly abstracted system that was used in 3:16.  Working in terms of abstract ranges, abstract threats, rather than concrete descriptions.  I think I could work with something like that, and certainly the AP accounts I've seen suggest a powerful role for fictional positioning, and ionterpreting mechanics into the fiction and out again, all of which looks quite useful.

That I think provides an interesting starting point gfor looking at how you might design a system that was useful for this sort of play, provided some sort of structure to the otherwise pretty vague and possibly misleading impressions that arise from using conventional RPG rules in this manner.

[Anders Gabrielsson]

So is the question now "How to do this well if the players are aware of and embracing the fact that this resolution system is being used"? Or perhaps I'm moving ahead.

[David Berg]

Gareth, I'm with you on all counts.  That scene is interactive exposition with teeth, and that's part of the meat of SBP play, rather than being filler or summary fodder.

Could you elaborate more on what you're envisioning with the below?  What gets abstracted, and how might that help SBP GMing?

Now I was quite intrigued by the fairly abstracted system that was used in 3:16.  Working in terms of abstract ranges, abstract threats, rather than concrete descriptions.  I think I could work with something like that, and certainly the AP accounts I've seen suggest a powerful role for fictional positioning, and ionterpreting mechanics into the fiction and out again, all of which looks quite useful.

That I think provides an interesting starting point for looking at how you might design a system that was useful for this sort of play

[contracycle]

Well the problem arises with the implied objectivism of conventional systems.  If there is a villain declaiming a speech then there is no reason, in principle, that the villain cannot be shot.  But in 3:16, the system operates explicitly against threat tokens, not described individuals, so a villainous named NPC, UNLESS represented by a threat token, is effectively out of the reach of the system.  In addition, there is a clear distinction between a scene where the NPC is not so represented, and one where it is: in that case, the implicit restriction has been exolicitly lifted.

So anyway, that is pretty cliche scenario, but its one that's often cited in this kind of thing.  But I think it is a useful precedent, in that it it labels things in the fiction according to the manner in which the players can systematically interact with them.  In a similar light, in another thread I described some things I do to create the impression, rather than the actuality of danger.  I might narrate gunfire or similar that arrivers near the characters, but which is not a rolled attack.  This is all aimed at creating atmosphere and tone, rather that doing the roll-for-attack-and-then-fudge-it-away thing.  Again, 3:16's abtraction allows this sort of narration as part and parcel of describing a fairly wide, generalised situation, inside of which discrete acts of system occur.  In a conventional reality-modelling system, each of these should implicitly be acts of system, but not so at 3:16's level of abstraction.

On a related note, I offered the HeroWars people an abstract battle system, in which you had abstract range bands and alignment blocks, and units could move between them based on a roll against their attributes.  The idea was to be able to indicate which unit was able to act against which opposing unit without fiddling about with rulers and manoeuvre points and that sort of thing.  Although this was systematically complete, it would, I thought, have left more scope for narration and such rather than breaking out to a top-down, depersonalised battle system, as is usually the case.

In all of these, what I'm describing is a move from modelling to abstraction in order to capture an appropriate subjective experience, and to  make a space in which the GM's authority can be introduced without the need to suspend the action of system.

[David Berg]

In all of these, what I'm describing is a move from modelling to abstraction in order to capture an appropriate subjective experience, and to  make a space in which the GM's authority can be introduced without the need to suspend the action of system.

Gotcha.  Yeah, I think that in-fiction causality-modelling certainly can't be the highest-order logic for resolving things in SBP.  That said, I think we want in-fiction causality to still show up in an important way, right?  Just like we discussed with samurai and Hastings in the other thread.

That 3:16 example with the threat token is a cool option to explore.  An SBP GM could use such an overt mechanism to operate the highest level of resolution, which includes, "Do we use modelling here or not?"  Depending on the options in a given game, it could be as simple as "GM decides Yes or No and signals that" or as complicated as a bidding system that determines the cost of modelling or not modelling what parts of the fiction right now.

Telling the GM, "Depending on the system's output, maybe you can't manage your plot the way you want," defeats Story Before, so that's why I say "cost".  Or, as Frank put it, the system could tell the GM "Yes, but" as opposed to simply "Yes" or "No".

Some games seem to take the approach of, "Yes, but pay the players for any agency you've cost them":

• FATE Compels: GM decides that a character's weakness applies here and now, thus stopping the character from attempting the action the player wanted.  The player is then awarded useful Fate Points as compensation.  (A player can bid against the GM to resist the Compel, forcing the GM to give them more Fate Points or abandon the Compel.  Not sure if that fits with SBP, where we may want the GM to have unlimited points, but just thought I'd mention it.)
• Todd's awards of Style Points for when he takes control of someone's character in Hollow Earth Expedition.

I have mixed feelings on this.  I like the fact that there is GM-player communication.  I like the fact that taking agency away from players is called out as significant.  I vaguely dislike the distance between the GM's action and the player's benefit -- it reminds me somehow of a bribe or a conciliatory payment, and rather loses the sense that even the GM's action which disempowers a player is actually, in itself, constructive and beneficial to the player (if not the character).  I don't have any brilliant ideas for "Hey look, by holding you back I just gave you a cooler experience!" though, so perhaps that's an idle thought.

I do have some ideas for alternative awards to the above "Help You Succeed At Character Actions" points.  Stuff like "your character's sub-plot gets more fully woven into the GM's main plot."  I'll go into that at some point, but I want to see folks' thoughts on the above first.

[David Berg]

Gareth, I just realized I have a major question here:

in 3:16, the system operates explicitly against threat tokens, not described individuals, so a villainous named NPC, UNLESS represented by a threat token, is effectively out of the reach of the system.

Does token presence/absence simply describe scene framing?  ("This NPC's too far away for you to shoot.")

Or does it describe possible outcomes regardless of scene framing?  (This is what I assumed in my previous post.)

If that latter, what happens when the evil NPC walks up to me to taunt me, and I want to shoot them?

- Do I shoot them and then the GM makes up some crazy reason why they don't die?

- Do I make up some crazy reason why none of my guns work right now?

- Do the GM and I brainstorm a mutually acceptable reason why I don't shoot?  I explain my character's thinking and motives, and the GM changes the scene so that now actually isn't a good time to take the shot, and my character would rather wait?

- Something better?  Hopefully?!

[Anders Gabrielsson]

Mutants & Masterminds and Smallville have similar systems, directly (IIRC, in M&M the GM gives the players Hero points when he does things like having the villain escape without the PCs being able stop them) or indirectly (in Smallville the player gets the resources the GM expends in winning against them).

One thing that might be problematic with this type of mechanic is that it could push the players into rebelling to get the reward for having their rebellion defeated: They know they won't succeed in killing the Big Bad before the rooftop showdown, but trying will force the GM to give them goodies in compensation for frustrating them. For SBP I think I'd rather have a system that gives the players goodies when they go along with the story or which makes explicit what things they can and cannot do.

(I think that token thing from 3:16 seems like a great mechanic for this. With something similar, the GM can clearly indicate what goals are reasonable for the players in a specific scene without having to state them explicitly.

Actually, maybe that could work as a kind of bribe mechanic as well. "We can't kill the Big Bad even though he's in the scene because his marker isn't among those available, but there's this one that will give us a bonus in the inevitable showdown.")

[Anders Gabrielsson]

If that latter, what happens when the evil NPC walks up to me to taunt me, and I want to shoot them?

If the GM does this, isn't he just being a dick? I'm not sure this is a system-level problem.

[JoyWriter]

Gareth that sounds to me a rather unusual solution; partially dissociating the mechanics the players have access to from the game world for the sake of making use of that fuzy link for plot protection:

How does that compare to say, creating a game in which HP and character health are tenuously linked, and going to 0 may have little negative effect unless specified by the games master?

It sounds like I'd start to feel like someone in a satelite above the game world, or an experimeter on the other side of glass. I would talk proactively about what I did in the space station, and see how the planet responded.


I made a wushu hack a while ago that played with this kind of distance thing; you roll to see if the GM will sideline your contribution with new information. Sounds harsh, but it was a way to get around the strange abstract resolution wushu uses while retaining clear fictional logic links. Basically you say what your character does to a certain level of detail, within the context of a GM defined conflict, and you do whatever you say you do, but whether it matters depends on how high you roll. It was also a bit of a psychological experiment/training exercise, about dealing with being shot down.

Anyway it occured to me that the resolution system in the second post seems to be doing the same kind of thing; succeed and your action is impactful, even if it is responded to and counteracted by others. The GM will spend some time affirming it, even as he counteracts it. Perhaps at the scene level in that example, you would have one of the pirates intentionally scuttling the ship, rather than random chance burning it down.

But I think in order for that kind of thing to work, you would have to constantly elide descriptions of the distributions of forces in the situation: Knowing who is on top politically, socially or militarily, and people's regions of influence, is always off the table, blurred over and obscured. How many pirates are there on that ship really? If you have that, then you can alternate between using a GM's ability to contextualise player action and the actual forces of the story, depending on dice rolls.

If they win a roll and you don't want them to succeed, escalate the power of their opposition, if they fail and you do want them to succeed, have some other faction/coincidence help them out. What people are rolling for or making decisions toward is how much of their success is of their own making, and how much of their failure is due to things outside of their control.

The nice point of this is that players who know this don't need to worry about assessing the danger levels they are in; if they want to they can enjoy the ride and emerge winners due to luck. There's a few downsides associated with that too though, like accidently slipping into tone damaging over-casual character actions etc.

[contracycle]

Gareth, I just realized I have a major question here:
...
Does token presence/absence simply describe scene framing?  ("This NPC's too far away for you to shoot.")

Or does it describe possible outcomes regardless of scene framing?  (This is what I assumed in my previous post.)

The latter, as I understand it, although I'll confess I'm working more or less at second hand.  But the systematically important things in 3:16 are: 1) whether or not a threat token is removed, and 2) how many enemies your character kills off.  So you roll in an attempt to damage or destroy a threat token, and one of the outputs is a kill count.

From this it follows that killing of this or that specific enemy individual is purely a matter of narration.  As long you are not cheated out of removing a threat token, it's perfectly reasonable for the GM to narrate in such a way that a villainous NPC survives.

If that latter, what happens when the evil NPC walks up to me to taunt me, and I want to shoot them?

Well the point is that system doesn't really work at that level, of resolving attacks on individuals.  Enemies come in swarms, the weapons are powerful, the scale is expressly set at military firepower rather than small unit tactics.  A weapon doesn't do "damage", it kills 1d10 enemies, or whatever.

So one of three things would have to be going on: its not a battle scene, and just down to narration, in which case it can be assumed that the NPC can just be killed; or it is a battle scene, and the NPC either has sufficient, indeed indefinite, minions to absorb the damage, or it is a full scale monster that can absorb damage as if there were minions, like tentacles or hydra heads or something.

In a sense, named individuals don't really matter to the system.  It wasn't designed to do the kind of thing I'm describing, its just coincidentally a way it can be used.  The abstraction that the system governs doesn't bother to say you shot this person, you did so much damage, and that leaves scope by default for narration to step in, without the system having to be fudged, or forced, or suspended.

Which leads me back to the bidding idea.  I don't really like these, becuase either you do allow players to overrule the GM or you don't.  Cost and so on is not an effectove deterrent; you just end up hoping that the players will "realise" that they shouldn't, say, commit all their story influence points to shoot your pet villain in the middle of his villainous monologue, which is pretty much back to square one.

Joywriter wrote:

Gareth that sounds to me a rather unusual solution; partially dissociating the mechanics the players have access to from the game world for the sake of making use of that fuzy link for plot protection:

How does that compare to say, creating a game in which HP and character health are tenuously linked, and going to 0 may have little negative effect unless specified by the games master?

Well the only unusual part is using it for plot protection.  In fact your latter example is a good one, because early D&D did just that: losing hit points did not imply bodily wounds, they could be position or a sort of running out of luck.  It would be perfectly within the rules to narrate every hit that caused HP to be lost as a near miss, and the last one that finally kills as a single, clean, stab to the heart.  I don't think its that weird.

[David Berg]

Ah, okay, I guess 3:16 doesn't answer the question "How do we integrate plot protection smoothly into the fiction?"  Perhaps I was jumping the gun with that line of inquiry.

Returning to my Pitfighter example above, I guess an abstracted system might concern some sort of direction or progress?  "No, don't hijack the pirate boat!" could be signaled by either a "No Plot This Way" token or a "Next Destination: the City of Haven" token.  The former would have to be played responsively, while the latter could be played during scene framing.  Or perhaps some text-based computer adventure game logic could be used in scene framing to express possible directions ("forward"/"back"/"anywhere"/"subplot A"/etc.)? 

Or perhaps destinations could be like 3:16 Threats, with tokens put down to indicate they are accessible (and perhaps how challenging they are to reach).  So any destination the players propose that doesn't elicit a token from the GM is a no-go.

Was there some other application you had in mind?

Separately, are we agreed that we might well want modelling pending the determinations of the abstract system?  E.g., GM drops a Kill Me token on the Saxon army so now we go to the Formation Fighting rules.

[contracycle]

I don;t have any other specific applications, the idea with raising 3:16 as a jumping-off point was to suggest that a useful approach is in framing the "moving parts" of the game, distinct from what the imaginary space implies.

The distinction in your Pitfighter case is essentialy between the implicit nature of the boat as a vehicle, and its functional relevance to the game as terrain. Some computer games, even like Streetfighter etc, have used that sort of thing; you jump back on forth on a sinking ship or what have you.

I'd be inclined to suggest something like "publishing" a list of chapter headings and objectives.  So this scene would have been explicitly labelled as a ship fight, and the next scene as, say a caravan journey.  Or maybe, in each discrete scene, the GM provides a list of "exit conditions", which in this case says stuff like defeat the minions and head west to the caravanserai.  Both of these make it quite clear the taking off with the ship is not a feasible option.

As for the modelling thing, I'm not sure that applies to all cases.  I just want to be do-able in the cases which need it, ratrher than moving on to assuming that an inherent step for SBP is to eliminate modelling of reality in exchange for modelling psychological states or whatever.  You could use chapter headings alongside physics modelling easily enough, for example.