The Horriffic Emptiness of Dungeons

[sparkletwist]

For Diderot the character is a model that the actor imitates, and in order to make the right sorts of calculations to portray the character as effectively as possible, the actor has to maintain a distance between themselves and the character, to ask themselves first "how would this character act?" and then "how can I achieve a representation of that action?"

There are some pretty big differences between acting and roleplay that make the analogy kind of fall apart if you look at it too closely, but I see what you're saying and I think this is a pretty good take on our different approaches to playing a character.

The conversation is about using in-universe resources and the environment and skills the characters possess, not numbers.

As you say: maybe. Maybe not.

I mean, first of all, when they get to making plans, the wizard is going to want to roll his knowledge skills and get some information on the thing that they just faced. The DM can try to dress this up and say artful things like "the creature's scales are as hard as full plate" but the players are going to know that means (and, in OOC discussions, probably just say) that its AC is at least 19. They might have said in character that the fighter should face the monster because he is the "most hearty of constitution," but what they really mean is he has the best Fort save. They might be talking about things in character, but I feel like the real plans are still being made using the numbers and the system, because ultimately that's how the players interact with the world. I've honestly never played a game of D&D or Pathfinder where any IC planning isn't supplemented (or just replaced) with OOC planning using game terminology, because those are ultimately the terms that players think in when their characters' survival is at stake.

There's a further issue as well. As I observed above, one thing about the conversation from Alien that makes it such good drama is that they're obviously scared and confused, not operating with perfect tactical efficiency. D&D parties are rarely so hampered, even if they are trying to think in character. What happens when the rogue with a fear of snakes is the one chosen to lead the creature into the trap, because she is the quickest and sneakiest? Realistically, this could impose a severe downside on what is otherwise a very good plan; however, lacking any sort of mechanics for this, most players would simply rationalize that she sucked it up for this one time so as not to add another hurdle to what is otherwise already a very difficult encounter. In fact, I suspect that in many D&D groups the player that decided to play this out and refused to participate on roleplay grounds would probably be considered to be being unreasonable. On the other hand, in a system like Fate, this would be far more acceptable, as the character has the aspect written right on her sheet, and the player pockets a fate point as compensation for the party having to work around this character flaw.

On the other hand if you have 100% assurances from the DM that death is completely off the table for your character unless you give prior permission - which is a way some people for sure actually play - isn't that effectively a kind of low-key god-mode?

I... guess? I feel like there is (or at least should be) an understanding in there that this is for the sake of making sure that everyone enjoys the story to their satisfaction, and to abuse it is basically the same as cheating. Characters in dangerous situations should still react appropriately to the situation.

[Steerpike]

I mean, first of all, when they get to making plans, the wizard is going to want to roll his knowledge skills and get some information on the thing that they just faced. The DM can try to dress this up and say artful things like "the creature's scales are as hard as full plate" but the players are going to know that means (and, in OOC discussions, probably just say) that its AC is at least 19. They might have said in character that the fighter should face the monster because he is the "most hearty of constitution," but what they really mean is he has the best Fort save. They might be talking about things in character, but I feel like the real plans are still being made using the numbers and the system, because ultimately that's how the players interact with the world. I've honestly never played a game of D&D or Pathfinder where any IC planning isn't supplemented (or just replaced) with OOC planning using game terminology, because those are ultimately the terms that players think in when their characters' survival is at stake.

Yeah, I mean I'm not actually so hard-line on "no rules talk!" as this, I just think the more rules-talk there is, the more meta the discussion becomes, and the more genuine dread gets minimized. But I don't think that any rules talk automatically destroys all fear of the monster.

In any case, I sort of see statistics in an inverse way. When we say the Fighter has the best Fort save, we really mean he has the most hearty Constitution, the Fort save is just an abstracted representation. So it makes sense in universe to have the fighter face the monster because he actually is the toughest and most likely to beat the venom, within the world of the game. At a certain point, talking in broad strokes about strengths and weaknesses of the monster and of characters in terms that map to mechanics is going to happen, and isn't necessarily totally corrosive to immersion; the question is how to prevent that from spiraling into full-on meta-discussion that totally disenchants the monster. I think system matters here, but I also think keeping the monster's abilities as un-quantified as possible and conversations as in-character as possible are mitigating tactics.

In fact, I suspect that in many D&D groups the player that decided to play this out and refused to participate on roleplay grounds would probably be considered to be being unreasonable. On the other hand, in a system like Fate, this would be far more acceptable, as the character has the aspect written right on her sheet, and the player pockets a fate point as compensation for the party having to work around this character flaw.

The snake example is interesting. In 5th I'd be probably give the character disadvantage on certain rolls versus snakes assuming this was a formal Flaw.

I'm worried this could devolve into another round of Steerpike & sparkletwist debate the merits of Fate, which maybe we should avoid. I did mention it first, and should have known that doing so would conjure you  :grin:

I will say that I think panic/confusion is quite hard to produce in players, but that I'm not sure either D&D or Fate has a good way of cultivating it, mechanically. I'll grant Fate has a clearer way of encouraging players to represent their characters as panicked, but does that translate into actual player panic? Because it's ultimately player emotions that I'm really after here - producing actual emotional reactions from players that mirror those of characters.

This is, I think, a good case in point for our different approaches to player-character interaction. For you - and I hope I'm being fair here, so correct me if I'm wrong - the point is getting the characters in a scene to be represented as scared and confused in order to craft a dramatic section of the story and add complications to the narrative. The emotional state of the players themselves is not really important here; the players can be totally calm or laughing at their characters' predicament, but they're feigning panic when the play their characters in the scene, and they can appreciate the drama of the scene as if they were an audience watching the characters react. In my view, the point is to try and elicit actual emotional reactions from players who are immersed in the setting and "fused" with their characters, inhabiting them, and then watch them react organically to the threat without thinking about the shape of the narrative or consciously trying to create a more dramatic situation.

Question for you - when you're DMing IRL, do you address people by their character names, or by their real names, and then ask them what they want their character to do?

[sparkletwist]

I think system matters here, but I also think keeping the monster's abilities as un-quantified as possible and conversations as in-character as possible are mitigating tactics.

Sure, but my point is that when dealing with a situation where there is no assurance of the encounter being "fair" and permadeath being on the table, a large majority of players (or, at least, a very large majority of players I've ever played with) are going to prioritize survival over immersion and plan so that their characters have the best chance of mechanical effectiveness-- that is to say, not rely on vague in-character abstractions but rather talk about things in the terms that are actually used when the dice fall and the determination is made whether the character is alive or dead. So, in my experience, one good way to avoid this is to not put players in situations where their characters' survival depends on using the mechanics effectively.

This is, I think, a good case in point for our different approaches to player-character interaction. For you - and I hope I'm being fair here, so correct me if I'm wrong - the point is getting the characters in a scene to be represented as scared and confused in order to craft a dramatic section of the story and add complications to the narrative. The emotional state of the players themselves is not really important here; the players can be totally calm or laughing at their characters' predicament, but they're feigning panic when the play their characters in the scene, and they can appreciate the drama of the scene as if they were an audience watching the characters react. In my view, the point is to try and elicit actual emotional reactions from players who are immersed in the setting and "fused" with their characters, inhabiting them, and then watch them react organically to the threat without thinking about the shape of the narrative or consciously trying to create a more dramatic situation.

This is a fair summary of it. I agree with you that it can be quite difficult to produce genuine emotions in players, so I mostly don't even bother, and instead try to focus on telling a good story and conveying the characters' reactions well. I also feel like, often enough, any feeling the player feels may not even be what that character realistically would feel in that situation, so that sort of "immersion," even when it works on some level, might still just be getting in the way.

Question for you - when you're DMing IRL, do you address people by their character names, or by their real names, and then ask them what they want their character to do?

I haven't had an RL group in a long time, to be honest. When I did, I used their actual names.

[Steerpike]

Sure, but my point is that when dealing with a situation where there is no assurance of the encounter being "fair" and permadeath being on the table, a large majority of players (or, at least, a very large majority of players I've ever played with) are going to prioritize survival over immersion and plan so that their characters have the best chance of mechanical effectiveness-- that is to say, not rely on vague in-character abstractions but rather talk about things in the terms that are actually used when the dice fall and the determination is made whether the character is alive or dead.

Maybe let's bring this back around to the concept of mostly-empty dungeons here. I agree that in certain types of situations, like in the very midst of an actual fight with a monster, you've got to deal with mechanics to a degree. As a DM you hope that the mechanics map reasonably well to how the situation would play out if it were real, but sure, when you're 3 rounds deep with the monster and it's time to kill this thing, you're going to have to break out a certain degree of mechanical terminology. Attack bonuses are going to be calculated. I try to describe things as vividly as possible to make these seem visceral and real, but there's dice being rolled. I do tend to hope that "survival" would be the priority of characters as well, but I get that what players would do to survive might not be identical with what their characters would do to survive, and that that dis-junction is what threatens immersion.

But this is precisely the point of the most-empty dungeon with the vague, mostly-unseen, and mechanically inscrutable monster, a monster you know could potentially totally outmatch your characters but which for almost all of the dungeon you are not actually facing. 90% or more of the time you're not going to be in direct conflict with the thing, you're going to be soaking up atmosphere and getting creeped out and learning about this thing in primarily non-mechanical ways and experiencing other aspects of the dungeon that don't pertain to hit points or armour class. And you're also going to be predisposed to avoid conflict with the thing (as real adventurers probably would, unless they're in the dungeon with the express purpose of killing it) because you strongly suspect, without necessarily any numeric particulars, that an encounter with the monster will go south fast, which tends to mean a lot more fleeing, sneaking, and other stuff that tends to be more minimally crunchy - not mechanics-free, necessarily, but certainly less "gamey" than the thick of combat where almost every decision is mechanized.

The actual fight with the monster is never the scariest part of horror, really. The good part is the build-up, the bits where you're trying to avoid the monster or where it hasn't shown up yet, but you think it might. If you know for a fact that the monster is going to be reasonably easily defeatable at the first encounter, then there's no real reason to avoid it from the player's perspective, right? You can use compels and other narrative mechanics to reward players for roleplaying their characters as if they were afraid of the thing, but the players will still be aware that if push came to shove they could just dispense with the thing without serious risk of being maimed/killed etc. I can see why this is perfectly fine from your perspective, because the point is getting the characters to act in a certain way to make a narrative that resembles a horror story, in this case, but from my perspective, where as far as possible I want to collapse player-character distance for these sorts of moments, it doesn't really solve the core problem, i.e. making the player identify so strongly with a character that they feel a genuine sense of dread, anxiety, horror, etc.

I agree with you that it can be quite difficult to produce genuine emotions in players

I don't actually think it's necessarily difficult to produce genuine emotions in players generally speaking. I think panic or confusion specifically is difficult to pull off given the way the game works (I mean, it's easy to confuse players with bad descriptions, but that's the wrong sort of confusion). But producing emotional reactions and creating emotional investments (player emotions, that is, not just character ones) is, like, the beating heart of satisfying roleplaying for me.

[sparkletwist]

I do tend to hope that "survival" would be the priority of characters as well, but I get that what players would do to survive might not be identical with what their characters would do to survive, and that that dis-junction is what threatens immersion.

I think that even when the player's idea and the character's idea of what to do pretty much align, immersion is threatened when the players are too focused on the problem as a player rather than as a character. For example, in a couple of D&D 5e games I've been in on Roll20, I've noticed that when a fight gets too 'tactical' the players pretty much just stop making IC posts at all, and a turn consists solely of moving their token and making their attack rolls, or whatever action they take.

But this is precisely the point of the most-empty dungeon with the vague, mostly-unseen, and mechanically inscrutable monster, a monster you know could potentially totally outmatch your characters but which for almost all of the dungeon you are not actually facing. 90% or more of the time you're not going to be in direct conflict with the thing, you're going to be soaking up atmosphere and getting creeped out and learning about this thing in primarily non-mechanical ways and experiencing other aspects of the dungeon that don't pertain to hit points or armour class.

I don't know what to say other than my experiences have been different, and this has shaped my different outlook. My experience has been that players faced with a mechanically inscrutable monster want to find out what makes that monster tick, because it's ultimately the numbers that determine how much of a real threat it is to the party. So, yes, they're going to be learning about it, but not in primarily non-mechanical ways, because non-mechanical information isn't ultimately what's going to save you or kill you. I do recall a couple of times in Asura that I described a big, nasty, imposing foe and nobody really reacted-- until I made an attack roll with a gargantuan bonus. Then there was that sense of peril you're after!

If you know for a fact that the monster is going to be reasonably easily defeatable at the first encounter, then there's no real reason to avoid it from the player's perspective, right? You can use compels and other narrative mechanics to reward players for roleplaying their characters as if they were afraid of the thing, but the players will still be aware that if push came to shove they could just dispense with the thing without serious risk of being maimed/killed etc. I can see why this is perfectly fine from your perspective, because the point is getting the characters to act in a certain way to make a narrative that resembles a horror story, in this case, but from my perspective, where as far as possible I want to collapse player-character distance for these sorts of moments, it doesn't really solve the core problem, i.e. making the player identify so strongly with a character that they feel a genuine sense of dread, anxiety, horror, etc.

What you're saying is true, but I don't understand why you had your example for using "narrativist" mechanics involve a monster that is easily defeatable. I mean, yes, you can use compels to reward players for running away from a monster that isn't, by the numbers, an actual threat, but you can also use compels to reward players for running away from a monster that definitely is an actual threat. And then things don't seem quite so disjointed.

For that matter, in Fate, you can use fleeing in terror (and getting into some other awkward situation) as a concession if the players end up trying to fight the numerically superior monster and losing; this is a softer failure state, so there is less "real peril," but unless it's a particularly brutal one-shot or something you don't actually want the monster to kill the PCs anyway, so this keeps the narrative going, which is a good thing!

[Steerpike]

Interesting. We definitely see immersion differently, but that's not really news. I completely agree that the depths of combat are when immersion can be most threatened, but that's why the mostly-empty dungeon minimizes combat.

The Asura anecdote about the scary monster is interesting, because my PCs have dramatically the opposite reaction these days unless they're feeling really, really confident. Because they have come to expect at least the possibility of monsters too powerful to safely take on, they actively try to avoid foes unless given very good reasons to get into combat. As a rule, most of the time they won't engage unless they are worried that if they don't they will be taken unawares or something.

My experience has been that players faced with a mechanically inscrutable monster want to find out what makes that monster tick, because it's ultimately the numbers that determine how much of a real threat it is to the party. So, yes, they're going to be learning about it, but not in primarily non-mechanical ways, because non-mechanical information isn't ultimately what's going to save you or kill you.

How do players learn about the quantifiable abilities of a totally mechanically inscrutable monster without being in combat with it? The mostly-empty dungeon is premised on the idea that actual encounters with the monster are very rare and that characters should be primed to try and avoid combat until they feel confident they can take the monster down, because they know that if they're not fully prepared then they're boned. So how can you really deduce stats of a monster that's not in the manual when you're not currently fighting it?

I mean, I'm fine with eventually finding the monsters strengths/weaknesses through investigation and trial and error and stuff that's actually played out in the game. That's not immersion-breaking, that's a core part of the horror genre. Like in the snake monster example, where I suggested the monster doesn't like fire, the characters don't actually know (and have no way of finding out) what this actually means in mechanical terms until they actually fight the thing with fire. And in that situation, I'm not even sure the mechanics are really threatening immersion all that much, because the characters would also be thinking "aha, fire hurts this thing!"

The sort of thing that worries me is more like - say the monster isn't a scratch-made horror but an Aboleth, and the players know all about Aboleths even if the characters don't, and so rather than trying to find out how the creature ticks, they're not going to go exploring, they're not going to experiment during their encounters with the thing, they just know immediately to stock up on cure disease potions and stuff to protect against mind control, and so instead of the Aboleth being a terrifying eldritch horror, it's just a puzzle they've already solved.

What you're saying is true, but I don't understand why you had your example for using "narrativist" mechanics involve a monster that is easily defeatable. I mean, yes, you can use compels to reward players for running away from a monster that isn't, by the numbers, an actual threat, but you can also use compels to reward players for running away from a monster that definitely is an actual threat. And then things don't seem quite so disjointed.

Oh sure, you can use compels in different situations. My point is just that if you don't use compels for a not-really-threatening monster, there's not really any reason for the characters to run away or hide, because essentially the players have to completely fake their fear. In other words, when you have compels, you at least make their characters act as they should when they face a scary monster, even though the system suggests the monster isn't really a threat (and the players know it). But if the monster really is a threat, yes you can still use compels, but you don't need them for the same reasons, because the character's mindset (Ack! This thing is going to eat me!) already matches the player's mindset (Ack! This thing is going to eat my character!). In other words you don't need to further incentivize running away and hiding, because that incentive is already baked-in to the disparity between monster and character.

[sparkletwist]

How do players learn about the quantifiable abilities of a totally mechanically inscrutable monster without being in combat with it?

Picking up on clues, I think. A lot of what the DM says and does can be translated into mechanics, like noting how many squares the creature moves when it's chasing you, paying attention to the DC of rolls against it, and whatever. Knowledge rolls are useful, too, of course, if the DM is willing to give some mechanical info as part of the knowledge. Personally, I'm of the mindset that the DM probably should give some mechanical info as part of a good knowledge roll, but that probably doesn't surprise you given how I think about the game.

The sort of thing that worries me is more like - say the monster isn't a scratch-made horror but an Aboleth, and the players know all about Aboleths even if the characters don't, and so rather than trying to find out how the creature ticks, they're not going to go exploring, they're not going to experiment during their encounters with the thing, they just know immediately to stock up on cure disease potions and stuff to protect against mind control, and so instead of the Aboleth being a terrifying eldritch horror, it's just a puzzle they've already solved.

Oh, yeah, I hate that.

Oh sure, you can use compels in different situations. My point is just that if you don't use compels for a not-really-threatening monster, there's not really any reason for the characters to run away or hide, because essentially the players have to completely fake their fear.

That's fair, but I don't think that's a particularly good use of a compel. Compels aren't there to "fake" things, but rather to provide a mechanical reward for playing out a character flaw. If the monster isn't particularly threatening, and the players and characters know that it isn't, using a compel to "fake" some sort of reaction is pointless and probably not worth bothering with. On the other hand, maybe the players know a monster isn't a threat but the characters don't-- in this case, player and character reactions aren't in sync, but the players can enjoy the dramatic irony and a compel is likely quite appropriate. A compel would also be fitting for an exceptionally timid character who is frightened by a weak monster that other characters know isn't a real threat, or something like that.

[Steerpike]

Picking up on clues, I think. A lot of what the DM says and does can be translated into mechanics, like noting how many squares the creature moves when it's chasing you, paying attention to the DC of rolls against it, and whatever. Knowledge rolls are useful, too, of course, if the DM is willing to give some mechanical info as part of the knowledge. Personally, I'm of the mindset that the DM probably should give some mechanical info as part of a good knowledge roll, but that probably doesn't surprise you given how I think about the game.

I can see how you'd pick up some of this, but it definitely seems pretty DM-dependent to me.  I'm not sure speed really counts as something "mechanical" exactly, as it's probably the least dissociated/abstract statistic a creature has.

It doesn't surprise me that you'd give out mechanical info, because I don't think you have the same sort of objections as I do. What that means, though, is that in practice it is pretty easy for a DM to keep most of a monster's stats unknown if you put it in the right environment and context. Sure, players may be able to guess armour class or something due to a few rolls, but for me the goal is to try and keep the players' minds off the numbers themselves while still talking about the monster's strengths and weaknesses in a way that will pay off in mechanical terms during the final confrontation.

If the monster isn't particularly threatening, and the players and characters know that it isn't, using a compel to "fake" some sort of reaction is pointless and probably not worth bothering with.

This thing you said is more or less my point. The monster should either be actually very threatening or at least seem potentially very threatening. I don't like compels in this situation either - but I also don't like non-threatening monsters in this sort of environment. My point is that for monsters to feel properly threatening in a Horrifically Empty Dungeon, you need to have established the possibility that the players could be severely outmatched, with nasty fail conditions. If there's a reasonable assurance that monsters will be matched carefully to character ability ("level appropriate" challenges) and/or that nasty fail conditions are off the table, I don't see how a monster can actually be very threatening; and if the monster is not actually very threatening, the only way to get characters to act threatened is using something like compels, which don't actually make the players feel threatened anyway.

On the other hand, maybe the players know a monster isn't a threat but the characters don't-- in this case, player and character reactions aren't in sync, but the players can enjoy the dramatic irony and a compel is likely quite appropriate.

Yeah, this is the only big thing that my approach leaves totally out. I pretty much avoid dramatic irony as much as possible, precisely because it requires player and character to be out of sync. But I can completely understand why you would enjoy dramatic irony in roleplaying, because we're deriving very different kinds of enjoyment from games.

[Kindling]

I love reading through your debates, Steerpike & sparkletwist, not least because I often get the impression that your views are far less diametrically opposed than you both seem to think they are at times. While I'm not going to say you've derailed the thread with it, though (everything you've been talking about has more or less been on topic after all!), I was sort of hoping for a more specific, nuts-and-bolts conversation about dungeon design.

That said you've both touched on some specific principles in the course of your discussion, which I think are helpful little terms for me to start thinking in, for example Steerpike's format of "Characters enter empty rooms -> find vague evidence of monster -> more empty rooms -> brief glimpse of monster (or what could be monster) from afar, or evidence that monster was recently present -> more information gleaned about monster and/or treasure or history of the place -> encounter monster, know they're outmatched because they've obtained information about it, narrowly escape, perhaps not unscathed -> find more evidence of monster, hinting at weaknesses -> find weapons/tools to defeat monster while being stalked and harassed by monster -> hunt monster down and kill it with cleverly executed plan."

I'm finding myself more and more interested in the idea of this kind of thing as a "haunted" or "ghost themed" dungeon, as I think it allows for a lot of more interesting interactions between players and monsters than the lair of living, flesh-and-blood beasties might. They might have fascinating encounters with completely intangible manifestations of the enemy, where neither party is able to physically touch one another and so they are left to simply observe which could give them all sorts of clues. Also when combat does occur, the ghostly nature of the foe might allow more interesting and less flatly permanent end-states to victory or defeat - maybe if the ghost takes you down to 0hp you don't die but it possesses you, and maybe if you take it down to 0hp you've banished it from this level of the dungeon, but it can still manifest itself in the lower depths.

[Steerpike]

Yeah, sorry about that Kindling. I think both sparkletwist and I are like pitbulls with tires when it comes to arguments about these kinds of things.

I think the ghost idea is great, especially since haunted locales often have a lot of ghostly "symptoms" which may not be direct actions caused by the immediate presence of the ghost - taps that run with blood, portraits that scream, doors slamming shut of their own accord. I've always really wanted to run something in a haunted house where the people died of plague, where the house itself has become sick with spectral, quasi-organic afflictions.

In terms of design goals, one thing I think is big is making sure that every given sequence of rooms has some meaningful story-clue, magical item, minor puzzle, or other key feature that rewards characters for exploring, because I think one of the pitfalls of the Horrifically Empty Dungeon is the frustration of scouring a sequence of chambers and finding not just rooms devoid of monsters but rooms devoid of anything.

In St. Severine's Skull, for example, I carved up the dungeon into distinct sub-regions ("The Gatehouse," "The Catacombs," "The Library," etc) and then levels within them. There would usually be only one major foe/threat in each sub-region, but each level would have something important in it. So for the Gatehouse, for example, there's one monster (the Gatekeeper), but then most levels would have some key feature:

Battlements - A mangonel which can be used to punch holes in the castle or attack creatures inside or outside it

Level 5 - Map of the gatehouse, dungeons, and key-ring to various parts of the castle

Level 4 - Ammunition for the mangonel

Level 3 - Supplies, most crucially a store of lantern oil

Level 2 - Winch Controls for the Gate, plus a useful amulet if a thorough search is made

Level 1 - Drawbridge and Gate - useful for keeping enemies out or trapping people inside the castle

Dungeons - A trapped NPC with useful information about other characters in the castle, also a valuable magical mask, well-hidden

Apart from occasional vermin I had only one monster in that 35-room section; the Gatekeeper laid a few traps here and there, but otherwise there were no other direct challenges. But I think if I'd made whole levels totally valueless the dungeon would have been a lot less interesting to explore. I think ideally you want players to start to fit together how a place works as they move through it, so that the exploration is a slow process of understanding.

Lovecraft's At the Mountains of Madness also does this really well, for a more literary example. The explorers move through lots and lots of empty rooms, but the rooms are filled with carvings showing the history of the underground city, and it's only at the very end that they encounter a Shoggoth, which they only recognize as such because of the carvings.