What is *worldbuilding* for?

[Jester David]

Just ran a session today. 5e. Generally a more sandbox game. I try not to create heavy plots but I do throw out seeds that the players can latch onto and decide to turn into plots. But my players do like a little guidance now and then, so I'm slipping in Tomb of Annihilation as a big "heroic deed of legend".

The players spent the first third of the session slowly making their way overland to the location where the tomb is located, a distant lost city. But I'm placing the tomb in my homebrew campaign setting, which is loosely a kitchen sink world and heavily detailed. But the players' chosen route just happened to pass alongside a gnomish vault. (Think a Fallout franchise vault.) I hadn't planned that and had very few ideas for the vault. So I planned to just let the players walk passed and use it as a dash of world flavour. Like passing giant statues in Lord of the Rings, having it too sealed up to enter.
But my players fixated on it and really wanted in. Because treasure. And they managed to magic a solution.

So they entered.
I pulled some details of the location out of my ass. Gave it some flavour and tied it into the "plot" of Tomb of Annihilation by having one of the leaders suffering the Death Curse. There was a fight with some kobolds, some exploration, and a huge moment of character growth as a PC had erased memories restored. Which allowed me to point them in the direction the plot needed them to go, preventing random wandering and wasting of time in exploration.
This random side-quest incidental encounter, which only occurred because I threw a random point on a map and never detailed it, suddenly fixed a gap in the plot.
Worldbuilding!


So, what is worldbuilding? Really, it's creating the setting. Which can be as big as making a huge sprawling campaign setting for the players to place dozens of campaigns in. Or it can be as small as just keeping the continuity of people and locations in a campaign: selling treasure to the same merchant, sleeping in the same inn between trips in the dungeon, befriending a local drunk at the tavern.
They can even be entirely confined to the dungeon. Say the party has a random encounter with an evil dwarf who manages to survive and escape the PCs. If that dwarf shows up later, that generates continuity. That's worldbuilding right there.

I don't see worldbuilding as unrelated to "classical" D&D play or design. Lots of early modules had monsters that moved from place to place, having a certain percentage of a chance of being found in one location. And there was discussion in a few of guards being replaced if the PCs killed the previous guards and left. Or monsters that call for reinforcements or run away. All that creates the illusion of a living world. It's as much worldbuilding as creating a backstory, kingdom, and detailed world to place the modular dungeon in.

 

[pemerton]

So the point of world-building, in this context, is similar to the point of creating a setting for a novel: it provides a context for story, and a space for the reader (or player) to explore and enjoy.

"Providing context for story and a space for the (players) to explore and enjoy" is not the same as the GM "present(ing) the players with the product of his/her imagination."

The main difference is the purpose (or point).

<snip>

In my phrasing, the purpose is centered on the play itself--both GM and players--for the point of worldbuilding is to provide context for game play (or story)

But am I correct in understanding that this is to be achieved by the GM presenting to the players certain products of his/her imagination?

I offered the paraphrase to which you are responding in response to your explanation that it matters, in an important fashion, that it is the GM and not the players who do the worldbuidling. If we just centre on the play itself - both GM and players - why would it matter that it is the GM rather than the players who do it?

this is not at all what I'm saying and I'm a bit baffled by why you'd think this. It is not about "ensuring passivity." It is about enabling a certain kind of immersion into otherworldliness, mystery, and uncertainty that I find is better facilitated by the GM being the primary creator and authority on the world.

I guess I'm trying to understand why having someone else do it facilitates the immersion.

To pick a concrete example: the player can decide that his PC has a brother who was possessed by a balrog; or the GM can decide that. The player can decide that his PC's immediate goal is to obtain an item that might help free someone (namely, his brother) from balrog possession; or the GM can decide that (eg by having a patron approach the PCs in a tavern and ask them to find such an item).

I'm not sure why this is all more immersive when it comes from the GM rather than the player - unless the idea is that one can't be immersed in the creations of one's own imagination, which is why I suggested a contrast between activity and passivity. (Not unlike the audience notion that some other posters have used upthread.)

Let's say my PC finds a chest, opens it, and then the DM says "pick any magic item from the DMG that is worth 50,000 GP or less." That's pretty fun but...something is lost. A sense of mystery, uncertainty, and I would say immersion.

Well, I've never run a game like that, or played in one, so I couldn't comment.

The closest I can get is the following: the PCs had trekked across the Bright Desert to a ruined tower. 14 years ago (ie well into backstory territory) that tower had been the home of one PC and his brother, before they had to flee in the face of an orc attack. It was in trying to stop that orc attack that the brother had tried to conjure up a storm of lightning, failed, and thus become possessed by a balrog.

The PC now returned to his tower after 14 years away, and the first thing he did - after the group had some water to drink, and a bit of a rest - was to look to see if a silver-nickel mace that he had been working on 14 years ago was still there. (That particular bit of backstory was authored not long before this episode of play.) So we framed a check - in D&D this might be Perception or Search or Investigation or an INT or WIS check (depending on details of the edition); in BW it is a Scavenging check.

An appropriate difficulty was set, extrapolated from the examples given under the Scavenging skill description. The dice were rolled. The check failed.

So the mace wasn't there (and the player straight away intuied where it would be - "Of course that b*stard GM will have it in the hands of that dark elf whose been harassing us ever since we entered the Abor-Alz"). But they did find something - in the ruins of what had been the brother's private workroom, they found black arrows, cursed to make it harder to heal the wounds they inflict. One of the PCs was very familiar with these arrows, because he carried one around his neck - a token of the death of his captain, which had led him to become a "ronin", leaving the elven lands to travel among the humans and try to find some way to lift his shame and grief at having failed his captain. And so all hell broke lose among the party - the PC has been insisting that his brother became evil because possessed by a balrog, but now it seemed that his brother was a suitable vessel for the balrog because already evil - a creator of black arrows for the orcs to use in their wars against the elves.

That was pretty immersive. I don't think it would have been more immersive if I had been the one to make up all the relevant PC backstory - about the brother, the tower, the orcs, the balrog possession, the mace, the ronin carrying an orcish arrow as a token of the death of his captain. I think it would have been less immersive.

I am reminded of how in video games, if you don't like the result you can always try again, or save the game at a certain point and keep going until you make it through.

I'm not sure what you are saying is reminding you of this. I can't remember if I've posted about retries or not in this thread, but different games have different approaches to retries. As best I understand, for instance, 3E and 5e D&D are very liberal in respect of them. AD&D tends to be rather limited in retries (most thief abilities are limited- the only exception I can think of is climbing walls, though hit point damge from falling might be seen as imposing a de facto limit; listening at doors is limited - after 3 goes the character must wait a turn; and opening magically held doors and bending bars are both limited). Burning Wheel has a "no retries" rule (called "Let it Ride"). During the 4e era, Stephen Radley McFarland posted a blog on the WotC site advocating a similar rule for 4e - my table was already using it.

Anyhow, you are positing a false entity that you entitle the "ENWorld Collective."

Well, all critical analysis (be it of visual arts, film, literature, or even RPGing) requires the making of some generalisations. We group creators and performers into schools and movements, and talk about trajectories of influence and development, even though we know that each creator and performer is an individual, different in some ways from any other.

You, for instance, pick up on [MENTION=6696971]Manbearcat[/MENTION]'s point (2) and dispute it - many ENworlders, you say, don't worldbuild as an art/creative outlet.

Well, be that as it may (and there are certainly some in this thread who have embraced it, and no one had repudiated until [MENTION=6696971]Manbearcat[/MENTION] put it forward as a conjecture), does that really distort the overall thrust of Manbearcat's analsysis? Suppose that we substitute in 2' - the GM worldbuilds because someone has to do it? That leaves (3) through (8) largely undisturbed - the GM is still using his/her world to establish framking and to manage the adjudication of action resolution - and (9) changes only a little bit - instead of the players being expected to appreciate the GM's art, they're expected to appreciate - in a utilitarian sense - the GM's effort, and if they don't think it's very good then they can "put up or shut up".

In other words, if (2) needs modification/correction we can still conjecture that there might be a broad consensus that reflects (3) through (8) plus an appropriately modified (9). Now, if you think that there is no such consensus, what are you pointing to (in this thread, or more generally)?

Now maybe there is a status quote, and shared assumptions that are vaguely aligned with your nine points, but what I see you doing is taking a relatively extreme end of it and saying the entire collective operates within that narrow range. I simply don't think this is the case.

What do you think is extreme about Manbearcat's conjecture? I'm not saying that you're right or wrong about that, as you haven't yet identified what you have in mind.

#8. Just because a GM can use fiat doesn't mean he or she does with any frequency. It is a kind of wild card that can be used in case of "catastrophic failure," or even simply when the GM feels like the overall game experience would be enhanced by a little nudge here or there. Each GM has a different take on what that means, and thus when to use it.

I don't see how this in any way conradicts what [MENTION=6696971]Manbearcat[/MENTION] has said. It seems to be a straightforward reiteration of his point!

Maybe you think enhancing the overall game experience (your phrase) is different from making for a better/more interesting story outcome (Manbearcat's phrase) - but you haven't explained what that difference is. And at this point I'm not seeing what it is.

for many (most?), the existence of GM fiat has no negative impact on the game experience, and is actually used as a way to enhance it.

That also doesn't seem to contradict anything that [MENTION=6696971]Manbearcat[/MENTION] has said, unless you're suggesting that it contradicts his (9) ie his claim that (4), (7) and (8) will cause tensions with (ie "inevitably bump up against") ideas of player choice or player discretion.

Let's treat "inevitably" as a rhetorical flourish - Manbearcat hasn't posited a time frame for this inevitability, after all - and look at the real claim: namely, that there is a tension between (4), (7) and (8) (ie the GM has various veto/manipulation powers as part of the process of action declaration, used to ensure fidelity to the setting and/or "metaplot" and/or "to enhance the overall experience) and the idea that the players get to choose what their PCs do.

The existence of that tension doesn't mean that people don't enjoy their RPGing. It does mean that RPGing done in this style might exhibit a recurring pattern of issues - eg "problem players" who buck the GM's authority over the fiction; debates about whether or not fudging is permissible; claims that the GM can do whatever s/he wants in relation to action resolution, as long as the players don't know; etc.

If everyone was running games in the way that Luke Crane describes in his BW books there might be recurring topics of discussion and debate, but the ones I've just mentioned would probably not be among them!

I think we're talking about two different approaches based upon different underlying assumptions about the roles and power of the GM and players. One approach assumes that the GM is omnipotent, and the player's relationship to the world is akin to our own relationship to our world; the players--through their avatars, the characters, have agency but not the capacity to alter reality (at least as far as we know!).

I find these metaphors relatively unhelpful for analysis.

My relationship to the real world is a product of various physical and biochemical processes. But no player's action declaration for his/her PC unfolds in virtue of such processes (except the ones going on in his/her head, and the head of the GM; plus the mechanical forces that govern the roll of the dice).

I also have the capacity to alter reality - just now, for instance, I'm making it true that certain words are "typed" on a keyboard and hence appear on a monitor. But when it comes to RPGing, we're not talking about any sort of reality - we're talking about a shared fiction, and the process whereby a group of people agree on what it includes.

This is why I ask if, on your picture of GM-worldbuilding, action declarations should be seen as suggestions to the GM to change or develop the fiction in a certain way? Because if that's not what they are - for instance, if the player has the power to change the shared fiction directly (eg by declaring an action, and then rolling some dice which result in a success) - then it ceases to be true that the GM is omnipotent in respect of the shared fiction.

 

[pemerton]

So, what is worldbuilding? Really, it's creating the setting.

That's largely what I had in mind in my OP, with one caveat - I was focusing on the GM creating the setting in advance of play.

I don't see worldbuilding as unrelated to "classical" D&D play or design.

I agree. The dungeon is a paradigm of setting.

there was discussion in a few of guards being replaced if the PCs killed the previous guards and left. Or monsters that call for reinforcements or run away. All that creates the illusion of a living world.

Gygax talks about this stuff in his DMG, yes. I can't remember what B2 says about it.

Referring back to some of the points I tried to make in the OP, I think this starts to bump into the limits of effective dungeon play as Gygax himself describes and advocates for in the closing pages (prior to the appendices) of his PHB.

For instance, the advice in those closing pages takes for granted that the PCs can enter the dungeon, engage in scouting, learn stuff about it (including the locations of "placed" or "permanent" dungeon denizens, their treaures, etc) and then leave and come back. So on the scouting foray, the spell load-out is lots of divination magic, the approach to encounters is to evade or parley, etc; whereas when the PCs come back to carry out the assault, the spell load-out is more offensive in nature, and the PCs are ready to beat up on the monster if it doesn't hand over its treasure!

The more the GM runs the dungeon as a "living, breathing world" the less feasible the approach to play just described becomes, because - as a practical matter - it becomes impossible for the players to make rational plans, choose rational spell load outs, realise the fruits of their scouting, etc.

Of course, the relatively static dungeon, whose ecology makes little sense and where dungeon denizens hang out waiting to be raided, is artifical at best and absurd at worst. These pressures of "verisimilitude" seem to have grown over time, and clearly inform Gygax's remarks on how dungeon denizens will respond to raiders; he also (as best I can judge) wants to make planning for, and responding to, denizens' responses itself an element of skilled play. But taken in its natural directions, this move towards verisimilitude creates puzzles/challenges that are, in practical terms, not reliably beatable through skill. Eg if the PCs raid the orcs, and beat up on the orc's ogre friend; and the GM decides that the ogre has a family who will mourn him/her and come to join the orcs to help beat up on these hubristic raiders; well, how can the players reasonably plan for that eventuality?

This is why I think that, once we look at a game like Runequest, or most contemporary D&D games, where the scope of play (in terms of NPCs, geography, locations, etc) is more-or-less boundless, the setting can't be doing the same sort of work it was doing back then. It's not a puzzle/maze for the players to unravel and beat through skillled play. It's something else - to the extent that some consensus is emerging in this thread, it's a source of context/meaning for play; and also a source of tools/resources/"levers" for the players to deploy in their play of their PCs.

Do you imagine the maze is populated or is just filled with doors to open, chairs to sit on and chest lids to open? So a dungeon is just a dungeon is just a dungeon, the non-player characters are about as animate as stone tiles and do not offer any story, rumours or intrigue in your adventures?

Pemerton IS talking about comparing the worldbuilding of classical dungeons with the worldbuilding of the wilderness. Once again I ask, I assume the classical dungeon is populated with sentient creatures and not only doors to open, chair to sit on and chest lids to open?

Would you say ToEE is a classical dungeon? Because within that module there are plenty of NPCs which with DM fiat will allow for the story to unfold in as many ways as an urban mystery or a wilderness exploration adventure.

EDIT: Because it does sound like what Pemerton should be asking is, Is worldbuilding only useful for non-animate objects? or something to that effect.

Hopefully what I have said just above in response to Jester David provides some further clarity here.

Early dungeons had NPCs, but their presentation, motivations etc were extremely narrow compared to even the crasser pulp literature. And, as [MENTION=6696971]Manbearcat[/MENTION] said, often their reactions were randomly determined (or there might be a note: "attacks anyone wearing the helmet taken from the altar in room 10", but otherwise reactions are random). Thus they are part of the "puzzle" to be dealt with. (I can't comment on ToEE, but what I've said seems true of Lareth the Beautiful in T1, and is true of the NPCs who populate the modules in early White Dwarf magazines, in X2 Castle Amber, in B2 Keep on the Borderland, and other early modules I'm familiar with.)

But as soon as we suppose that the thief who lives in room 6 on the 2nd level can, if the GM thinks it makes sense, call upon his brethren in the Thieves' Guild, the situation becomes on that is so open-ended in its parameters that, as a practical matter, I don't see how players are supposed to map it out (literally and metaphorically) and then beat it in the way that Gygax talks about in his PHB.

techniques that worked in the dungeon context - obtaining information by way of sheer fictional positioning and free roleplay ("We open the door and look in" "We lift the lid of the chest" "How many goblins can we see through the peephole?") - become far less feasible. The players become far more dependent on the GM to dispense information (eg in the form of rumours; encounters and interactions with various city inhabitants; etc).[/qjuote]What a strange thing to say (bolded part). Who dispenses information when a character opens a door, lifts the lid of a chest or peeps through a peephole?

When the environment is confined in the way a dungeon is, it is relatively easy for the players to declare actions which oblige the GM to give them the information they need.Eg they can declare that their PCs listen at a door, which requires the GM to make a roll, and then the GM has to tell them what, if anything, they hear - of course hearing nothing is no guarantee of anything, as maybe the roll failed or maybe the room has undead or bugbears in it, but make enough checks and you can start to be a bit confident that probably not all of them failed, and the information you do get - ie when the check succeeds, the monsters aren't bugbears, and so the GM tells yo what you hear - is very useful for your dungeon raiding.

Likewise there is divination magic, from spells, wands, potions, medallions (of ESP), intelligent swords, etc. All of which gives useful information for dungeon raiding.

Then there are treasure maps which - if the GM follows the standard advice - at least sometimes will give you useful information about the dungeon you're exploring.

But once the imaginary vista is a whole wilderness, or a city, the idea that the players can get their PCs into the right fictional positioning to declare actions to obtain all the information that is broadly salient to their endeavours becomes unrealisitc. The field of action, the scope of imagined events, just becomes too broad. When I say that the players become dependent upon the GM to dispense information, I'm not talking about the GM's roll in narrating the outcome of action declarations - I'm saying that they depend upon the GM for all the framing in the first place. Eg if they want to inspect a library, first the GM has to decide whether or not there is any library to be inspected, and then has to decide whether or not the library, if there is one, contains any relevant books/scrolls, and [i[then[/I] has to decide what language they are in, and then has to paraphrase what might be hours of reading thousands of words into a few pithy bits of information to narrate to the players. The play dynamics of this are completely different from declaring "I look through the archway - how many orcs do I see?" - where the GM already has a note about how many orcs there are, and just tells the player the number read from those notes. The attempt to obtain relevant information is not mediated through all these GM decisions about the content of the fiction, and how to paraphrase it.

Perhaps the player characters have a choice of travelling as passengers on a barge, alone on horseback or guards as part of a travelling caravan as they have to be discreet for story reasons. Perhaps they have to take the fastest possible route due to x or they desire to take the safest route. The goblins tracks split, some continue along the river, but a small contingent is now heading towards the mountains... These examples are similar in fashion to your typical T-junction decisions in a dungeon.

"The mountains" is very different from the left rather than the right corridor in a classic dungeon. The latter is mapped. The players can learn it's layout by declaring actions for their PCs (of moving through it), having the GM describe the layout, and drawing a map.

"The mountains" aren't mapped in anything like the same detail. The players' exploration of the mountains is going to be highly mediated through general GM descriptions, GM choices as to what to make salient and what not, etc. It's a very different thing.

 

[pemerton]

I see, so in your personalised definition of worldbuilding introducing a 'map with names' is not worldbuilding. So worldbuilding depends on the extent of the worldbuilding. So 1 map is fine, 3 is borderline but 5+ is definitely worldbuilding?

The number of maps has nothing to do with it. This thread asks about the point of GM pre-authored setting. That contrasts with other ways of establishing setting, eg by showing everyone a map at the start of the campaign and saying, "Hey, let's use this."

In practice, there's going to be a limit on how much stuff can be shared. There's also another limit that is relevant if the idea is for setting to emerge through play, rather than be established by someone in advance, namely - you have to leave room for that stuff to emerge!

(A loose analgoue of the issue for classic D&D: the traditional use of wandering monsters isn't really compatible with assuming that every being in the dungeon is fully detailed in advance by the GM. Because if the latter was the case, then where are all those wanderers coming from?)

Let's play in Mystara isn't worldbuilding it's pitching a game despite the fact that this setting has a unique cosmos with its own planes and includes specific races and concepts and excludes others. Is that how you see it?

It depends on the details. If "Let's play in Mystara" means "Here's this cool map from the back of my Expert book, and we're only going to have elves, dwarves, halflings and humans" then absolutely - that's pitching a tropes for a game.

If it means "I've got this pile of gazzeteers and stuff full of backstory that you don't know about, and that I will rely on to adjudicate your action resolution, but without telling you" then that's not just pitching tropes - it's pitching a whole playstyle, of GM preauthored worldbuilding that will be used to determine the outcomes of action resolution in the style [MENTION=6696971]Manbearcat[/MENTION] described a little bit upthread; and in the style that [MENTION=59082]Mercurius[/MENTION] has described as "omnipotent GM".

Does Marvel Heroic include any DC character? (I've never played the game, so I'm asking)

It was published with a licence from Marvel, not DC. So it contains Marvel characters, not DC ones. DC characers would be easy enought to build for it, though I've not tried as I'm not so much into DC as Marvel. (Eg you'd use Hawkeye as your starting point for Green Arrow; Quasar as your starting point for Green Lantern; etc.)

 

[pemerton]

A slight tangent:

A hammer has strengths and weaknesses that depend upon context - that is, what you want to use it for.

Not really. It's a weakness (or flaw) in a hammer that it's handle is brittle, or that it's head is not flat - because that is an obstacle to the hammer serving its core function.

I'm not much of a carpenter, but I can imagine it counting as a weakness in a hammer intended mostly for driving nails into timber that it lacks a claw on the back for pulling out nails. But such a lack is not a weakness in eg a sledge hammer, which is not designed for such use.

It's not any sort of weakness at all in a hammer that it not very good for eating eggs off - and if you ever found yourself in a situation where it was crucial to eat your eggs of something and a hammer was all you had, well you may well be grateful for the hammer as being better than nothing at all!

In the context of which we speak different approaches to world-building could have strengths and weaknesses depending upon the effect you want to manifest and the agreement of the gaming group. If everyone except the GM wants a more co-creative experience, but the GM mostly wants to present their brilliant creation, then there are weaknesses to his approach given the context.

I just find this quite unhelpful and even obscurantist.

If I want to play a no myth game, it makes sense to ask whether the formal approach of FATE-based games, or the more informal approach set out in Burning Wheel, and that I also have used in my 4e game, will work. And you might say of these that each has strengths and weaknesses - the existence of a structure in FATE is probably a strength as such, as in game play structures help mediate expectations and manage or eliminate conflicts - but the output of the structure in FATE might count as a weakness if it leads to "Aspect bingo", whereas the more informal and somewhat organic approach in BW might be seen to produce a more intimate and visceral engagement with the fictional positioning.

Now maybe all the above is wrong about strengths and weaknesses - I've never played FATE, and what I've said is the barest of conjectures - but it makes sense.

But to say that a weakness of the Burning Wheel approach is that it doesn't allow for a GM driven metaplot is just nonsense - like saying that a weakness of a hammer is that it's hard to eat with. No one who wants to run a GM driven metaplot-type game would pick up Burning Wheel and try to use it for that purpose.

Of course maybe we don't need to frame this as strengths/weaknesses, and more as contextual appropriateness.

Or just purpose or utility. Different tools and techniques are useful for different things.

 

[Caliban]

What a strange game. The only winning move is not to play.

 

[billd91]

But once the imaginary vista is a whole wilderness, or a city, the idea that the players can get their PCs into the right fictional positioning to declare actions to obtain all the information that is broadly salient to their endeavours becomes unrealisitc. The field of action, the scope of imagined events, just becomes too broad. When I say that the players become dependent upon the GM to dispense information, I'm not talking about the GM's roll in narrating the outcome of action declarations - I'm saying that they depend upon the GM for all the framing in the first place. Eg if they want to inspect a library, first the GM has to decide whether or not there is any library to be inspected, and then has to decide whether or not the library, if there is one, contains any relevant books/scrolls, and [i[then[/I] has to decide what language they are in, and then has to paraphrase what might be hours of reading thousands of words into a few pithy bits of information to narrate to the players. The play dynamics of this are completely different from declaring "I look through the archway - how many orcs do I see?" - where the GM already has a note about how many orcs there are, and just tells the player the number read from those notes. The attempt to obtain relevant information is not mediated through all these GM decisions about the content of the fiction, and how to paraphrase it.

And how is this really different from the DM placing a library within a dungeon (perhaps a wizard's library or the library of a hidden temple) and defining what it roughly encompasses? How does taking it out of the dungeon really change things? Because it's in reaction to the players saying they're looking for a library and the DM having to define more of it on the fly rather than defining it beforehand? Ultimately, that's not all that different. What is it about taking a situation like that out of the dungeon makes it different? Because there are more variables involved?

 

[pemerton]

And how is this really different from the DM placing a library within a dungeon (perhaps a wizard's library or the library of a hidden temple) and defining what it roughly encompasses?

It's not. But that's not the canonical source of information for staging a dungeon raid. In all the discusssions of skilled play published back in the day, libraries figured barely at all - as opposed to divinatinion magic and actual scouting.

Contrast Call of Cthulhu, where libraries are the canonical source of information.

My contention (or, if you prefer, confidently advanced conjecture) is that contemporary settings in RPGs resemble CoC more than the artificial dungeon environment of classic D&D.

 

[Aenghus]

The alternative to heavy worldbuilding is generic fantasy, which is what a lot of people who aren't interesting in worldbuilding, or lack the time for it, resort to. Generic fantasy typically has huge issues with consistency and logic (castles and flying foes, a crazy, anachronistic mix of technologies and politics, terrible geography etc etc), but these issues only bother a minority of the audience. For one off games and short campaigns there isn't time to impart a complex setting the players aren't already familiar with.

For me the differences between a plot-important library in a dungeon and one a city concern action economy and suspension of disbelief. Dungeons constrain player decision making, and there is a more or less finite set of meaningful actions in the dungeon. Even when an action in the dungeon seems not to accomplish anything, it marks another element off the checklist.

Whereas in a city there are an infinity of possible actions the players could take, most of which accomplish nothing (arguably). The risk that checking out the library is a waste of playtime is to me higher, because there are a million stories in the naked fantasy city, and most of them are not relevant to the pcs/ plots (in a conventional prepared game where the plot doesn't follow the PCs).

The more freedom the players have, the harder it is to hook them in, so that everybody including the referee is having fun. I've seen it happen before that a game that worked fine within the confines of the dungeon, (constrained adventure space) falls apart in the wilderness/ city as the players don't notice or ignore the referees hooks and hare after red herrings and the stuff that interests them in the setting, which may not be what interests the referee.

 

[Jester David]

That's largely what I had in mind in my OP, with one caveat - I was focusing on the GM creating the setting in advance of play.

It's one way, but it's not the only way.
The GM can create the setting wholly before play. They can create chunks of the setting before each session, expanding as needed. They can create the broad strokes and flesh it out during play. They can entirely improvise during play. They can tap the players to add locations and places. Or the whole setting can be collaboratively created at the table by the players.

I agree. The dungeon is a paradigm of setting.

I'd use microcosm.
It's an environment, which can be part of a larger ecosystem or on it's own. It can be it's own sub-setting or just a quick location.

Gygax talks about this stuff in his DMG, yes. I can't remember what B2 says about it.

Referring back to some of the points I tried to make in the OP, I think this starts to bump into the limits of effective dungeon play as Gygax himself describes and advocates for in the closing pages (prior to the appendices) of his PHB.

For instance, the advice in those closing pages takes for granted that the PCs can enter the dungeon, engage in scouting, learn stuff about it (including the locations of "placed" or "permanent" dungeon denizens, their treaures, etc) and then leave and come back. So on the scouting foray, the spell load-out is lots of divination magic, the approach to encounters is to evade or parley, etc; whereas when the PCs come back to carry out the assault, the spell load-out is more offensive in nature, and the PCs are ready to beat up on the monster if it doesn't hand over its treasure!

The more the GM runs the dungeon as a "living, breathing world" the less feasible the approach to play just described becomes, because - as a practical matter - it becomes impossible for the players to make rational plans, choose rational spell load outs, realise the fruits of their scouting, etc.

I was actually thinking of Against the Giants where there's a random chance certain changes happen. (Plus the added worldbuilding of creating a new threat—the drow—and expanding the backstory of the world with the inclusion of fallen elves and Lolth, which is also textbook worldbuilding.)

Planning is impossible if things are shifting continually. But there's already a chance of uncertainty in random monster encounters, which make planning tricky. By the rules, there's meant to be a chance that plans get derailed due to bad luck. A skilled DM will find a middle ground, having most things feel similar but changing a few details to reflect the actions of the players or passage of time. Some changes due to having a "living world" might be beneficial to the players.
For example, if they're watching the Steading of the Hill Giant Chief for 3-4 days from a long distance to observe their patters, they might find raiding parties going out for food every third day, after which there's a larger feasts. The PCs might react to that by attacking after the raids when there are fewer hill giants present or following the feats when the residents are bloated and sleepy. There's no advantage in this game to have everyone in the Steading maintain their book position for several days in a row.

Of course, the relatively static dungeon, whose ecology makes little sense and where dungeon denizens hang out waiting to be raided, is artifical at best and absurd at worst. These pressures of "verisimilitude" seem to have grown over time, and clearly inform Gygax's remarks on how dungeon denizens will respond to raiders; he also (as best I can judge) wants to make planning for, and responding to, denizens' responses itself an element of skilled play. But taken in its natural directions, this move towards verisimilitude creates puzzles/challenges that are, in practical terms, not reliably beatable through skill.

I think that increase of verisimilitude is just the effect of people having played more. If I take a new player an introduce them to the game, they're not going to think much of verisimilitude, but they're just focused on the new things they can do in the game, playing the game itself. You need to know the confines of the box before you think outside it...
Like monsters sleeping. The modules never account for big NPC monsters sleeping or going to the washroom or the like. They're always found in their throne-room or dining halls. There's no advantage for attacking at noon or midnight. Which is something you'll quickly consider as an experienced player, but might not as a newbie. But the module can't give two complete descriptions of the dungeon for a diurnal and nocturnal attack.

When Gygax was writing the 1e books in 1976 to 1977, he'd been DMing for 4 to 6 years. I started playing in 1992 when I was 13. I had 6 years of DMing experience in University. Now I have over two decades of experience running and playing D&D. Gygax was a creative guy, but me right now probably knows more

Eg if the PCs raid the orcs, and beat up on the orc's ogre friend; and the GM decides that the ogre has a family who will mourn him/her and come to join the orcs to help beat up on these hubristic raiders; well, how can the players reasonably plan for that eventuality?

That's a pretty absurd example, but it is a fair point.
A better one would be responding to raids by increasing the guards and setting up traps. Which might be unexpected, but isn't unrealistic. And reacting to the unexpected with aplomb is a great demonstration of player skill.

This is why I think that, once we look at a game like Runequest, or most contemporary D&D games, where the scope of play (in terms of NPCs, geography, locations, etc) is more-or-less boundless, the setting can't be doing the same sort of work it was doing back then. It's not a puzzle/maze for the players to unravel and beat through skillled play. It's something else - to the extent that some consensus is emerging in this thread, it's a source of context/meaning for play; and also a source of tools/resources/"levers" for the players to deploy in their play of their PCs.

I'm amused by how you keep saying "RuneQuest and modern games" when RuneQuest was first published in 1978 and predates the 1e DMG...

Which also confuses me as I don't see the game system at play here at all. You can play the DM vs Player dungeon crawl skill game with most modern systems. Conversely, many DMs playing OD&D and 1e D&D created expansive setting. Or, like Ed Greenwood, used D&D to play in an already created fantasy setting. The modules of D&D were added into their worlds, existing as locations as part of the larger setting.
Really, I would argue that was the place of worldbuilding in classic D&D: worldbuilding was creating the setting where you could put the published adventure modules and then build a story around them. That's why sites like Grognardia were so hard on Dragonlance: they had their own story, making it harder for DMs to add those modules to their own world or create their own plots around the dungeon.