What is *worldbuilding* for?

[AbdulAlhazred]

The fiction may be awesome or not - that seems mostly a matter of taste. If players enjoy the GM presenting them with the products of his/her imagination, no doubt that's a reason for the GM to create and present such products.

My claim is simly about agency. In that situation, the players do not seem to have a great deal of agency over the shared fiction. This was a point that was made upthread and treated as controversial. But now it seems that it is a point that attracts widespread agreement.

Where do the constraints come from, then. Eg how is it determined where the map is located?

I suspect, since most of the DMs of 2e/3e/5e games I've played in work something like this, that in [MENTION=34194]byron[/MENTION]'s play process the players are making pretty high-level decisions about where to go and what to do, probably often informed by intelligence they have acquired about different areas of the GM's world. So in effect you could model it like a 'dungeon maze' where the players have maps and descriptions of some areas and a fair idea that going left or right at the next passage will lead to areas with some specific character to them, or something that they want. Also, being an RPG, there's always the alternative of subverting some of the GM's arranged material. In 'dungeon maze' terminology someone prepared a scroll with 'Passwall' on it and can use that to partially rewrite the map at some point if they don't like the direction things are going.

Players are also likely to find ways to have their characters exercise significant agency in terms of restructuring the world at a higher level, building a castle, conquering a city, assassinating a king, whatever. As long as the GM in question doesn't fall to exercising force and illusion to artificially limit this, then its cool. This is a treacherous problem though because there's always SOME logic inherent in the situation for resisting player-established goals. I mean, it wouldn't even be exciting in a challenge sense (which is part of this modus of play to one extent or another) to simply allow the PCs to roll over things and get what they want. Where is the line between challenging and railroading/controlling? Its a big grey area. Clearly fudging die rolls in order to limit (or enhance) character success crosses into force/illusionism, but there's a LOT of territory short of that.

Lets suppose the PCs decided to assassinate the king. Just how strong are his bodyguards and what precautions does he take? That probably isn't established in advance, at least in detail. The GM now is in the position of effectively deciding if the task is within the resource limits that the PCs can deploy (and those are purely in-game resources, the players have no meta-game power outside of 'convince the GM to let us do X').

Now we begin to see clearly WHY the alternative forms of play evolved. In MY game, this would simply be cool beans. The PCs wouldn't know if they were going to succeed or fail ahead of time, but the players and GM could establish that the risks were "either you succeed and the kingdom falls into chaos as you wish, or you fail and the King's army will lay waste to your lands and besiege your castle!" Maybe there's a 'lesser risk' version where the PCs hire some guy to try to poison the King, little chance of success, but maybe even the attempt is a useful ploy, and no real chance it gets back to them either. This can all be established using the sorts of techniques I've noted here and in other threads, and/or with those of [MENTION=42582]pemerton[/MENTION] or [MENTION=6696971]Manbearcat[/MENTION] (which are substantially similar though maybe not identical).

IMHO the problem with the 'classic' approach is it ALWAYS ends up falling to the DM's shoulders. Even if the players are substantially driving the story with in-game decision making on a big stage there are limits. The GM is always free to derail them at any time, there is always the great likelihood of unknown, maybe unknowable, backstory entering into play to thwart them, etc. AT BEST in the final analysis it is a GM-gated story.

The 'story driven' approaches DO go outside this limitation. By establishing some level of focus on character needs and player interests, and always allowing those to be entirely the scope of the players to establish, the role of the GM shifts to 'provider of narrative framing'. Instead of presenting fictional elements as pre-established world canon the GM is throwing up story elements to act as food to feed the agenda established by the players. It could go ANYWHERE. In the assassination example above its still within the GM's scope to declare the consequences of actions, but that will be within the framework of 'say yes' or 'roll the dice'. The ultimate result can easily be that the character's plot is foiled, they are unmasked, and their castle is laid in ruins and they flee into exile. That's a fun outcome! It is of course likely that the players are partisan to some degree and WANT to succeed in the character's plan, but they still have to wager, and part of the fun in that is the possibility of failure. If they fail then they'll have plenty of material to fuel further conflict narrative and perhaps in the end they will triumph in some other fashion, or simply go on to other concerns and leave that incident as a backstory element.

I think you could do all of this using 2e if you were really focused on it and know the techniques well. It just doesn't HELP you.

 

[Lanefan]

I suspect, since most of the DMs of 2e/3e/5e games I've played in work something like this, that in [MENTION=34194]byron[/MENTION]'s play process the players are making pretty high-level decisions about where to go and what to do, probably often informed by intelligence they have acquired about different areas of the GM's world. So in effect you could model it like a 'dungeon maze' where the players have maps and descriptions of some areas and a fair idea that going left or right at the next passage will lead to areas with some specific character to them, or something that they want. Also, being an RPG, there's always the alternative of subverting some of the GM's arranged material. In 'dungeon maze' terminology someone prepared a scroll with 'Passwall' on it and can use that to partially rewrite the map at some point if they don't like the direction things are going.

Players are also likely to find ways to have their characters exercise significant agency in terms of restructuring the world at a higher level, building a castle, conquering a city, assassinating a king, whatever. As long as the GM in question doesn't fall to exercising force and illusion to artificially limit this, then its cool. This is a treacherous problem though because there's always SOME logic inherent in the situation for resisting player-established goals. I mean, it wouldn't even be exciting in a challenge sense (which is part of this modus of play to one extent or another) to simply allow the PCs to roll over things and get what they want. Where is the line between challenging and railroading/controlling? Its a big grey area.

Corollary question: where's the line between a DM introducing elements in order to frame a scene and a DM building the world? This also seems to be a rather fuzzy division.

Clearly fudging die rolls in order to limit (or enhance) character success crosses into force/illusionism,

Or cheating, depending on how bluntly one wants to put it.

Lets suppose the PCs decided to assassinate the king. Just how strong are his bodyguards and what precautions does he take? That probably isn't established in advance, at least in detail. The GM now is in the position of effectively deciding if the task is within the resource limits that the PCs can deploy (and those are purely in-game resources, the players have no meta-game power outside of 'convince the GM to let us do X').

Now we begin to see clearly WHY the alternative forms of play evolved. In MY game, this would simply be cool beans. The PCs wouldn't know if they were going to succeed or fail ahead of time, but the players and GM could establish that the risks were "either you succeed and the kingdom falls into chaos as you wish, or you fail and the King's army will lay waste to your lands and besiege your castle!" Maybe there's a 'lesser risk' version where the PCs hire some guy to try to poison the King, little chance of success, but maybe even the attempt is a useful ploy, and no real chance it gets back to them either. This can all be established using the sorts of techniques I've noted here and in other threads, and/or with those of [MENTION=42582]pemerton[/MENTION] or [MENTION=6696971]Manbearcat[/MENTION] (which are substantially similar though maybe not identical).

IMHO the problem with the 'classic' approach is it ALWAYS ends up falling to the DM's shoulders. Even if the players are substantially driving the story with in-game decision making on a big stage there are limits. The GM is always free to derail them at any time, there is always the great likelihood of unknown, maybe unknowable, backstory entering into play to thwart them, etc. AT BEST in the final analysis it is a GM-gated story.

The 'story driven' approaches DO go outside this limitation. By establishing some level of focus on character needs and player interests, and always allowing those to be entirely the scope of the players to establish, the role of the GM shifts to 'provider of narrative framing'. Instead of presenting fictional elements as pre-established world canon the GM is throwing up story elements to act as food to feed the agenda established by the players. It could go ANYWHERE. In the assassination example above its still within the GM's scope to declare the consequences of actions, but that will be within the framework of 'say yes' or 'roll the dice'. The ultimate result can easily be that the character's plot is foiled, they are unmasked, and their castle is laid in ruins and they flee into exile. That's a fun outcome! It is of course likely that the players are partisan to some degree and WANT to succeed in the character's plan, but they still have to wager, and part of the fun in that is the possibility of failure. If they fail then they'll have plenty of material to fuel further conflict narrative and perhaps in the end they will triumph in some other fashion, or simply go on to other concerns and leave that incident as a backstory element.

"The PCs hire an assassin to poison the king" is an excellent example of my idea from way upthread about "pushing over a domino" and subsequent off-stage results. The difference between this and my earlier not-the-best play example involving the spy-harlot is that here the PCs are very intentionally pushing over the domino to set some events in motion; events which in this case the PCs quite understandably don't want to be involved in or assiciated with if possible.

But then what? No matter what system you use, in a case like this it's squarely on the DM to determine how the dominoes fall and-or whether there's any subsequent narratable effects either indirectly (next morning, news of the king's sudden death sweeps the streets of Kerratrim) or directly (next morning a squadron of soldiers shows up at the PCs' inn) affecting the PCs.

"We hire an assassin to poison the king" is a large-scale action declaration (most DMs would quite reasonably want to break this down into some smaller-scale details e.g. where are you finding an assassin, what terms are you offering, how does the assassin contact the PCs before-after the job, etc.) that, on a success, means an assassin has been hired to poison the king; and there the involvement of the PCs ends for a while, if not forever (e.g. the PCs set this in motion and then immediately skip town).

It doesn't mean the assassin has successfully poisoned the king yet. The DM has to by whatever means play this out either with herself or (and maybe the best if not always most practical option) by having someone else not playing in that game do the rolling etc. for the assassin when the usual players aren't around.

In a normal DM-driven game this isn't a problem at all - in fact, it's pretty much business as usual. The question is more one of how would a player-driven system or game handle something like this, where the players/PCs are acting as backroom directors rather than front-line agents?

Lan-"sometimes it's nice as a PC to sit back and let someone else do the dying"-efan

 

[Aenghus]

"We hire an assassin to poison the king" is a large-scale action declaration (most DMs would quite reasonably want to break this down into some smaller-scale details e.g. where are you finding an assassin, what terms are you offering, how does the assassin contact the PCs before-after the job, etc.) that, on a success, means an assassin has been hired to poison the king; and there the involvement of the PCs ends for a while, if not forever (e.g. the PCs set this in motion and then immediately skip town).

IMO this is the sort of large scale action which conventional procedural games can struggle with on both micro and macro levels.

On the micro level the probabilities of repeated dice rolling can all but guarantee eventual failure if the GM is inclined to keep asking for rolls till a critical one fails. I have seen this happen more than once in real games, typically when the GM doesn't think the players can succeed or doesn't want them to succeed, or just doesn't understand probability. Very rarely I've seen players succeed by sheer awesome dicerolling despite such GM tactics, which in some cases is accepted bu the GM in good faith, but in others explodes the game in flames as the GM can't handle large PC-initiated changes in the setting.

At the macro level I've seen GMs refuse point blank to adjudicate such actions. Sometimes the players can still sneak in the action by assembing all the parts without telling the GM what the overall plan is, or perhaps having a cover plan more acceptable to the GM. The results are similar to above with a bigger chance of GM bad will.

In some cases the outcomes of the attempt prove to be unacceptable to the players, given the lack of a stakes negotiation phase. Depending on the level of unhappiness, there can be arguments, players leaving, the group breaking up, or player revolution with a new GM installed.

It doesn't mean the assassin has successfully poisoned the king yet. The DM has to by whatever means play this out either with herself or (and maybe the best if not always most practical option) by having someone else not playing in that game do the rolling etc. for the assassin when the usual players aren't around.

In a normal DM-driven game this isn't a problem at all - in fact, it's pretty much business as usual.

Handled competently, yes. There plenty of examples in original games or even in printed works of similar situations handled incorrectly as outlined above, thus the number of gun-shy players who feel obliged to check for probability traps.

The question is more one of how would a player-driven system or game handle something like this, where the players/PCs are acting as backroom directors rather than front-line agents?

Lan-"sometimes it's nice as a PC to sit back and let someone else do the dying"-efan

I suspect player-based games with stake setting could find such problems much easier, with a few dice rolls or even just one if the king and government isn't important to the main player game goals. Stake setting allows the players and GM to agree as to outcomes, win or lose, and avoid most of the "unhappy player" issues above. With fewer dice rolls it's much easier to get the probabilities correct and give the players a transparent idea of their odds of success or failure, and what's likely to fall out either way.

 

[Lanefan]

IMO this is the sort of large scale action which conventional procedural games can struggle with on both micro and macro levels.

On the micro level the probabilities of repeated dice rolling can all but guarantee eventual failure if the GM is inclined to keep asking for rolls till a critical one fails.

True, though the odds of success on any given roll (including automatic) would realistically be swayed by situation. For example if the PCs are already familiar with the town and have contacts etc. then finding and hiring someone would be much easier than if they were a bunch of total strangers here for the first time.

I have seen this happen more than once in real games, typically when the GM doesn't think the players can succeed or doesn't want them to succeed, or just doesn't understand probability. Very rarely I've seen players succeed by sheer awesome dicerolling despite such GM tactics, which in some cases is accepted bu the GM in good faith, but in others explodes the game in flames as the GM can't handle large PC-initiated changes in the setting.

The not-understanding-probability is fair, but otherwise it comes back to whether the DM is both willing and able to hit the curveball thrown by the PCs/players.

At the macro level I've seen GMs refuse point blank to adjudicate such actions.

I can certainly understand a "wtf?" response if a bomb like this gets dropped on a DM without warning - it's happened to me numerous times - but even if the DM thinks the result will 99%-likely be a TPK (or at worst TPLifeInPrison) she should at least go through the motions.

That said, if a DM sees disaster coming there's nothing wrong with asking the age-old question: "Are you sure about this?". If the players carry on after that, so be it...

In some cases the outcomes of the attempt prove to be unacceptable to the players, given the lack of a stakes negotiation phase. Depending on the level of unhappiness, there can be arguments, players leaving, the group breaking up, or player revolution with a new GM installed.

I think here you're being a bit extremist. The players/PCs have in effect given up any say over the final outcome via their putting someone else (in this case the hired assassin) on the front line, and in so doing have implicitly agreed to a) trust the DM to fairly and reasonably adjudicate what happens behind the scenes and b) accept whatever outcome comes out from however the dominoes fall - whether it's what they originally wanted/intended or not.

Attempting to kill a king whether directly or indirectly is obviously not just a very high stakes move, it's almost certainly an all-in move. No negotiations required.

And a player who would leave a game over this is probably a player I didn't want in the first place as he or she is clearly not willing to accept losing a gamble: in this case the 'loss' being a negative outcome of actions s/he (probably along with the other players) intentionally and willingly set in motion in the first place.

I suspect player-based games with stake setting could find such problems much easier, with a few dice rolls or even just one if the king and government isn't important to the main player game goals. Stake setting allows the players and GM to agree as to outcomes, win or lose, and avoid most of the "unhappy player" issues above. With fewer dice rolls it's much easier to get the probabilities correct and give the players a transparent idea of their odds of success or failure, and what's likely to fall out either way.

Problem here is this immediately crashes into the wall of player knowledge vs. PC knowledge.

The PCs (and thus players) have no real way of knowing the assassin's chance of success. At best they can use their own observations, intuitions and information gathering to give it an educated guess. But once they send the assassin on his way the - pardon the pun - die is cast, and the players/PCs at this point should and IMO must have no way* of knowing what's happening except via events/actions/outcomes in the game world that their PCs can actually observe, as - hopefully competently - narrated by the DM.

"You wish your operative well and send him on his way. For a few hours the night is quiet, then suddenly shouting erupts from the general direction of the palace; and soon it's clear the shouting is getting closer."

I posit the players'/PCs' reactions to this will be a lot different if the players already meta-know the assassin has 1) succeeded or 2) failed than 3) if they don't know. Option 3 here is the only one that can give a player/PC reaction untainted by meta-knowledge.

* - for these purposes let's assume the PCs and the assassin do not have long-range communication with each other. If they did, this whole set-up would change dramatically and the players in fact could play through the actual assassination attempt via giving direction and suggestions to the assassin and receiving reports and updates in return.

Lan-"assassination before the age of radio"-efan

 

[pemerton]

So long as they are satisfied with the amount of agency they do end up with, it doesn't matter. More agency isn't objectively better

Sure. That's not in dispute.

My goal in starting the thread was analysis, not aesthetic evaluation.

pre-authored backstory can be an advantage, as such worlds can have enough detail to allow them to keep making a bunch of small incremental decisions, rather than much smaller number of big dramatic ones.

I want to write something about boxed text

<snip>

Done right, boxed text allows newbie GMs to run a tolerable game, and newbie players to figure out how to play. When the players make some progress in the adventure, it's often signified by a new passage of boxed text. When run in this style pretty quickly players associate boxed text with adventure progress.

I think it's fair to describe this sort of approach to RPGing as one in which the players declare moves, the purpose of which is to trigger the GM reading material from his/her notes. Progress through the adventure consists in making some of those moves (ie the ones that trigger boxed text) rather than others (ie the ones that trigger the GM elaborating on established details, or having to ad lib such that the "notes" become notional rather than literal).

EDIT:

I suspect, since most of the DMs of 2e/3e/5e games I've played in work something like this, that in [MENTION=34194]byron[/MENTION]'s play process the players are making pretty high-level decisions about where to go and what to do, probably often informed by intelligence they have acquired about different areas of the GM's world. So in effect you could model it like a 'dungeon maze' where the players have maps and descriptions of some areas and a fair idea that going left or right at the next passage will lead to areas with some specific character to them, or something that they want.

This would be an instance of a significant element of play being to make moves that trigger the GM narrating material from his/her notes. intelligence acquired about the GM's world, after all, just means having learned the content of the GM's notes.

Players are also likely to find ways to have their characters exercise significant agency in terms of restructuring the world at a higher level, building a castle, conquering a city, assassinating a king, whatever. As long as the GM in question doesn't fall to exercising force and illusion to artificially limit this, then its cool. This is a treacherous problem though because there's always SOME logic inherent in the situation for resisting player-established goals.

<snip>

Lets suppose the PCs decided to assassinate the king. Just how strong are his bodyguards and what precautions does he take? That probably isn't established in advance, at least in detail. The GM now is in the position of effectively deciding if the task is within the resource limits that the PCs can deploy (and those are purely in-game resources, the players have no meta-game power outside of 'convince the GM to let us do X').

Now we begin to see clearly WHY the alternative forms of play evolved.

<snip>

IMHO the problem with the 'classic' approach is it ALWAYS ends up falling to the DM's shoulders. Even if the players are substantially driving the story with in-game decision making on a big stage there are limits. The GM is always free to derail them at any time, there is always the great likelihood of unknown, maybe unknowable, backstory entering into play to thwart them, etc. AT BEST in the final analysis it is a GM-gated story.

There's a lot going on here!

I agree about the GM-gated story, where a significant structural element of the gate is the GM's pre-authored setting material ("worldbuilding"). The players have the capacity to declare actions - but the GM exercises a very high degree of control over how these actions resolve.

Want to build a castle? The GM has to decide how much stone is available from local quarries.

Assinate the king? As you say, the GM decides on the bodyguards etc.

There are potential issues of "fairness" here, but I don't think that's the main thing. The GM can be as fair as you like - that doesn't change the fact that assassinating the king is likely to involve a lot of play time spent with the players triggering the GM to tell them stuff (in the fiction, this will be spying etc). Or in the building example, the players spend a lot of time learning information about availability of materials, employees, etc.

I've known of GMs who string this stuff out endlessly to avoid having to deal with the actual goal of the players' action declarations (to perform the assassination, build the castle or whatever).

 

[pemerton]

So you came up with the idea of te NPC beforehand, but because you used two ideas introduced during play to introduce him, you claim that the NPC is an idea and not a pre-authored element?

Yes. I've been posting about the contrast between preparation and establishing setting for much of the thread.

To give more examples: the murderous mage could have been introduced as a potential ally of the PCs. Or: when the player failed the Circles check to make contact with Jabal, instead of taking the approach that I did (Jabal sends Athog to tell the PCs to leave town) I could have had the murderous mage approach the PC ("Jabal's after me - I hear you can help me get away from him!").

Another way to come at the distinction is this: who sold the angel feather to the peddler? In the session I ran, it turned out that the murderous mage did so. But what if the players had decided to investigate this directly - so that it becomes like the map? There is no established setting element that it was the murderous mage who did so, which would then become part of the (unrevealed) ficitonal positioning that would inform adjudication of actions delcared in that investigation attempt.

The GM may introduce elements to the story because he thinks they are interesting.

Yes.

But identifying a trope, or an idea, as being potentially interesting is different from establishing the setting in advance of play.

There's always the danger here that such prep will graduate onto the level of being canonicalized by the GM and thus become a driver of play for the GM.

Agreed. Luke Crane talks about some aspects of this in the Adventure Burner for BW.

What if there was someone who introduced material in response to PC dramatic needs, and someone who established color, and someone else who dealt with pacing? I mean I don't know how to parse it actually, but who's done this analysis?

I think separating your second - introducing material - and fourth - pacing - could be tricky. But that's conjecture, not based on actual experience.
[MENTION=99817]chaochou[/MENTION] has much better knowledge than me of systems and approaches that depart from the "mainstream" GM/player division of roles.

 

[AbdulAlhazred]

Corollary question: where's the line between a DM introducing elements in order to frame a scene and a DM building the world? This also seems to be a rather fuzzy division.
Or cheating, depending on how bluntly one wants to put it.

Its also a quite common technique in actual use, and endorsed by no less than EGG, amongst many others. In later times it has come to be seen as less desirable, but still quite heavily utilized in many games. To be quite honest vanilla 5e doesn't really give you much support for alternatives, though there are some optional rules that kind of help a bit, and obviously GMs/tables can incorporate any technique they want.

As for the line between introducing elements and world building, yes, it isn't really a line, its a continuum and thus I don't really think it makes too much sense to dwell heavily on the middle of that range. The extremes are interesting though.

"The PCs hire an assassin to poison the king" is an excellent example of my idea from way upthread about "pushing over a domino" and subsequent off-stage results. The difference between this and my earlier not-the-best play example involving the spy-harlot is that here the PCs are very intentionally pushing over the domino to set some events in motion; events which in this case the PCs quite understandably don't want to be involved in or assiciated with if possible.

Right, I wanted to construct it as the kind of example that would work well in a variety of games. If I framed it as "some NPC that has a grudge against the party stemming from stuff that happened years ago frames them for assassinating/attempting to assassinate the king and..." that would fit fine within the 'domino theory', but it would obviously involve 'hidden backstory' (assuming the players had no way of discovering the plot and no inkling that they might even need to be wary of such). So we have again probably a continuum, and that illustrates how there aren't really totally distinct practices of play (as opposed to theories).

But then what? No matter what system you use, in a case like this it's squarely on the DM to determine how the dominoes fall and-or whether there's any subsequent narratable effects either indirectly (next morning, news of the king's sudden death sweeps the streets of Kerratrim) or directly (next morning a squadron of soldiers shows up at the PCs' inn) affecting the PCs.

Agreed. In my formulation, the players chose the wager. The GM framed the situation in some degree "As you contemplate the various ways to thwart the King's evil plans you recall your contacts with the Assassin's Guild." (maybe this is narrated as the result of a check or in HoML it would simply be part of an interlude, since nothing is yet at stake). Once they've committed and the challenge is afoot then any consequences are theirs to bear. I have no problem with this. They fail the challenge and soldiers show up at their door! 4e/HoML handle this quite easily.

"We hire an assassin to poison the king" is a large-scale action declaration (most DMs would quite reasonably want to break this down into some smaller-scale details e.g. where are you finding an assassin, what terms are you offering, how does the assassin contact the PCs before-after the job, etc.) that, on a success, means an assassin has been hired to poison the king; and there the involvement of the PCs ends for a while, if not forever (e.g. the PCs set this in motion and then immediately skip town).

It doesn't mean the assassin has successfully poisoned the king yet. The DM has to by whatever means play this out either with herself or (and maybe the best if not always most practical option) by having someone else not playing in that game do the rolling etc. for the assassin when the usual players aren't around.

In a normal DM-driven game this isn't a problem at all - in fact, it's pretty much business as usual. The question is more one of how would a player-driven system or game handle something like this, where the players/PCs are acting as backroom directors rather than front-line agents?

Yeah, I think the SC (HoML General Challenge) system of 4e is fine with this. The fictional position is one of pulling the strings from the shadows. Part of the consequences of a significant enough failure is that the shadows are drawn away and the puppet master is revealed. It could be that there are lesser possibilities in a range. Failing after several successes might result in being forced to leave town to avoid being unmasked, or being forced to frame one of your allies to take the fall in your place. Limited success might produce similar results, the King dies but you or someone/something you value is lost, or even that you are fully revealed and your 'victory' becomes hollow or much more equivocal.

If the game is one in which the players are heavily invested in the fiction and their characters then these sorts of indirect consequences work fine. In a sort of simpler setup where the PCs just cart around the setting and don't really have much concern for allies, property, allegiances, etc. then maybe a setup like this doesn't work too well. At worst there's a fight scene and otherwise perhaps the party simply moves on to some other location and doesn't look back.

I'd note that I wouldn't have other people or the GM by himself 'play out' anything like an assassination. I'd simply have it either succeed or fail based on whatever the PCs did and the check results they got. If they plan it really well, hire the best guy, equip him with everything he needs, etc then things go off as planned (IE 12 successes and no failures). If there's one failure, then maybe there's clues left behind that point to the PCs, perhaps requiring them to undertake further actions to avoid being revealed, 2 failures maybe means the assassin is captured afterwards and fingers the PCs (or their agents). I never roll dice between one NPC and another, the relevant part of the game is what the players do. I guess it would be OK to have an opposing NPC roll to accomplish something that thwarts or complicates a PC action, sort of like in combat, but I don't really use that technique.

 

[pemerton]

An example that has come up in other discussions is searching for a secret door. A player declares her PC is searching for a secret door at the end of a dead-end passage, after which the DM can respond in one of two ways depending on the situation:

1. Say yes. On this the player - not the DM - has just introduced a secret door into the fiction.
2. Roll the dice. On success, the player - not the DM - has just introduced a secret door into the fiction. On failure quite likely she has not, but she still may have.

While the player can't usually bypass the check if the DM calls for one, the success (automatic on say-yes) or failure of the check itself establishes whether the player's attempt to stipulate a new fiction element works or not. But the result remains: sometimes players can stipulate fiction elements, or try to.

In your example, the player isn't stipulating anything but his/her PC's action: the PC searches for a secret door.

That's the whole point: the player doesn't have to think about the gameworld expect as a place his/her PC is engaging with: "I look around for a vessel to catch the blood!" "I search for a secret door so we can get out of here!" "I search the study for the map!"

The action resolution tells us whether the attempt succeeds or fails: does the PC find a vessel or not? a secret door, or not? the map, or not?

Only a rotten metagamer would be spending time outside the PC perspective to worry about the process whereby the shared fiction is established!

 

[Aenghus]

I think here you're being a bit extremist. The players/PCs have in effect given up any say over the final outcome via their putting someone else (in this case the hired assassin) on the front line, and in so doing have implicitly agreed to a) trust the DM to fairly and reasonably adjudicate what happens behind the scenes and b) accept whatever outcome comes out from however the dominoes fall - whether it's what they originally wanted/intended or not.

I did say "sometimes", there are lots of acceptable resolutions I didn't discuss, precisely because they are acceptable. I'm interested in failure modes and how to avoid them, so I tend to focus on stuff going wrong.

The stuff going wrong in this case is generally when the players and GM's view of the gameworld differ significantly. In a conventional game, the GM has the overwhelming advantage, but if the players actively dislike the resulting game it's a lose-lose scenario for everyone concerned. As a player, nowadays I'm much less likely to play through extreme frustration, boredom or misery.

Attempting to kill a king whether directly or indirectly is obviously not just a very high stakes move, it's almost certainly an all-in move. No negotiations required.

And a player who would leave a game over this is probably a player I didn't want in the first place as he or she is clearly not willing to accept losing a gamble: in this case the 'loss' being a negative outcome of actions s/he (probably along with the other players) intentionally and willingly set in motion in the first place.

I was actually imagining disappointment over a "success" that was fiated into a state the players saw as an unfair failure.

Problem here is this immediately crashes into the wall of player knowledge vs. PC knowledge.

This wall is potentially a lot shorter or even non-existent in player-driven games.

The PCs (and thus players) have no real way of knowing the assassin's chance of success. At best they can use their own observations, intuitions and information gathering to give it an educated guess. But once they send the assassin on his way the - pardon the pun - die is cast, and the players/PCs at this point should and IMO must have no way* of knowing what's happening except via events/actions/outcomes in the game world that their PCs can actually observe, as - hopefully competently - narrated by the DM.

In some systems the players are allowed access to these probabilities either as players, PCs or both. Your preferred style isn't an universal rule and may not apply to other people.

In some systems the assassin could be a resource on the character sheet of one of the players

I find strangely attractive the idea of resolving the whole task with one dice roll, again in the case where this was a side show and the main business of the game rested elsewhere. Amateur PCs would probably fail badly, experienced ones would likely have a better chance.

"You wish your operative well and send him on his way. For a few hours the night is quiet, then suddenly shouting erupts from the general direction of the palace; and soon it's clear the shouting is getting closer."

One way of doing it. Depending on how ruthless and professional the PCs are, they could set up a false flag operation and kill the assassin afterwards if he survives.

I posit the players'/PCs' reactions to this will be a lot different if the players already meta-know the assassin has 1) succeeded or 2) failed than 3) if they don't know. Option 3 here is the only one that can give a player/PC reaction untainted by meta-knowledge.

I don't see meta-knowledge as necessarily forbidden or unthinkable. As a GM I have to constantly handle meta-knowledge, some players can do so as well if they are willing and able. And as I wrote above, the failure modes I was most concerned about is apparent success turned by the GM into what the players saw as failure, or changing the gameworld to something unfun for the players.

e.g. state of emergency and martial law closes down much of the nation and prevents the players from pursuing any of their goals. They could hide in a hole for an extended period, but if the GM forces them to play through their confinement one day at a time I can see good reason for player unhappiness.

Extended periods of meaningless boredom is possibly the worst type of game, at least PC death allows new PCs.

* - for these purposes let's assume the PCs and the assassin do not have long-range communication with each other. If they did, this whole set-up would change dramatically and the players in fact could play through the actual assassination attempt via giving direction and suggestions to the assassin and receiving reports and updates in return.

Lan-"assassination before the age of radio"-efan

In D&D for instance there are a bunch of spells and items for long range sensing and communication.

But the competence and loyalty of the assassin could inform a single die roll that indicated the result in non-procedural systems. In some games this might be a secret GM roll, in others it could be open with pre-negotiated stakes determining the outcomes.

 

[pemerton]

IMO this is the sort of large scale action which conventional procedural games can struggle with on both micro and macro levels.

<snip>

Very rarely I've seen players succeed by sheer awesome dicerolling despite such GM tactics, which in some cases is accepted bu the GM in good faith, but in others explodes the game in flames as the GM can't handle large PC-initiated changes in the setting.

<snip>

Sometimes the players can still sneak in the action by assembing all the parts without telling the GM what the overall plan is, or perhaps having a cover plan more acceptable to the GM. The results are similar to above with a bigger chance of GM bad will.

In some cases the outcomes of the attempt prove to be unacceptable to the players, given the lack of a stakes negotiation phase.

Personally, I would regard all this as a sign of something having gone pretty badly wrong at the table.

EDIT: I just read your follow-up post which talks about the GM narrating success as failure (eg the assassination succeeds, but the outcome still thwarts the players eg because of the outbreak of civil war or similar). That, to me, is another sign of things having gone badly wrong.

The idea that success is success, not just another mode of failure established by the GM deploying hidden backstory material, is the first step towards establishing genuine player agency over the shared fiction.

(Obviously this also connects directly to what [MENTION=82106]AbdulAlhazred[/MENTION] has talked about as GM-gated stories.)

FURTHER EDIT:

A micro version of this is when the GM doesn't treat any non-combat resolution as final - so eg the NPCs surrender and promise compliance, but then betray the PCs as soon as possible. Or the PCs find the map, but its withered parchment turns to dust in their hands. Etc.