Before I even begin, let me first say I’m using the word “dungeon” very loosely here. By “dungeon”, I could mean an actual dungeon, a cave complex, a set of ruins, a patch of jungle or forest or desert, or even an isolated village. The basic premise is simply a self-contained location, intended to be explored (likely cautiously) by the PCs, as a series of encounters (many of those likely violent).
The way I think of “dungeons” is that they fall into two categories: dynamic and static, and while there are exceptions to every rule, the general rule is that the two do not mix well. They are each governed by their own special set of rules to maintain their internal logic, and internal logic is key here: immersion is about maintaining pace and the suspense-of-disbelief. Anything that makes a player pause to think “but how…” is counter to good gameplay. Dungeons come with their own sets of implied rules, and they have to adhere to them, lest “but how…”.
Dynamic Dungeons
Back in first edition of Dungeons and Dragons, modules were little more than a map combined with a collection of rooms and their contents. As much as we all like to remember “the old days” with the rose-colored glasses of nostalgia, the fact is that when you go back and review these dusty old modules, you’ll likely come away disappointed when you realize two crucial things: first, just how “bare bones” the information is contained therein, and second, how little logic is contained when you take a step back and consider the overall picture. One could kick open a door and find a fight with some orcs, and then kick open another door and find a larger chamber with a fearsome dragon. How did the dragon get in there? It’s sheer size would make it impossible. What is the relationship between the orcs and the dragon? If they are friendly, why didn’t they retreat to its protection, and if not, why do they live RIGHT NEXT DOOR? These are the sorts of questions modern players ask, while we – back in our youthful, unsophisticated bliss of earlier editions many moons ago – perhaps did not ask. Keep on the Borderlands has every type of humanoid imaginable living in caves with entrances a mere stone’s throw apart, yet little explanation as to why the trolls (e.g.) haven’t simply eaten all the goblins already.
The first module I remember being really impressed by in this regard was a dungeon in The Rod of Seven Parts in second edition. There was an orc lair guarded by a chained-up Ettin, and then farther below, other forces. Their lair contained a kitchen and bathroom, explained how and when the smoke vented out the top of the hill (a possible point of entry or exit for a creature small enough), what they hunted, and so on. It detailed the day/night cycle and resident locations. It was the first time I read a module that approached the process of detailing a dungeon not merely as a collection of contents, but as a living, breathing ecosystem.
For me, it was an awakening that would guide every future dungeon I would write or run.
I believe that in order for a dungeon to be memorable, it has to be immersive, and to be immersive, it has to feel like a believable ecosystem. It’s not simply a container that houses a scattering of unrelated monsters resulting in unrelated combat events. It is an ecosystem into which the players likely intrude, and it reacts and attempts to reestablish homeostasis when they do. If the players retreat and rest, the inhabitants move around and shore up defences. As the players invade, opponents retreat and call for assistance from their brethren, or attempt to lure the group into traps.
The Sunless Citadel is an excellent example in this regard: it houses multiple types of enemies, includes non-combative youth and elders. It contains opposing factions and takes an otherwise mundane crawl into something more memorable by weaving a story into that collection of rooms and those opposing factions. It can play out as a simple hack-n-slash crawl, or as something far more complex, but either way the players will be left with a sense of having invaded someone’s home.
How to Build an Ecosystem
Creating an ecosystem doesn’t mean things have to be simple and boring. The forest doesn’t contain only mice and owls and nothing else. It contains a whole variety of wildlife in a complex, dependent structure. A Kobold warren shouldn’t contain only Kobolds but its other contents should make sense within the context of it being a Kobold warren.
In order to build a dungeon as an ecosystem, I start out by selecting its primary inhabitant(s). If there’s more than one, I need to be able to justify their co-existence and explain the nature of their relationship. Then I build out from there, exploring the lives of these creatures and possibly adding additional creatures in response to the needs of this ecosystem, in way becomes an iterative, recursive process, as I explain to myself how they all fit together.
Take as an example that Kobold Warren. They’re the primary inhabitants.
I ask myself questions about how they live:
- What do they eat?
- What do they drink?
- Where do they sleep?
- Where do they go to the bathroom?
- Do they have children?
(If you’re familiar with Maslow’s Hierarchy of Needs, it’s a great place to start.)
For a Kobold Warren, the answers to these questions will lead to the inclusion of a kitchen, dining room, some sort of primitive barracks, and the oft-forgotten toilet. There should probably be a nursery of non-combative youth/hatchlings.
For creatures that are sentient, there may be a few more questions as well:
- How do they make their shelter safe?
- What do they use for construction materials, tools, or storage?
- Do they clothe and armor themselves?
- Do they have a currency? Do they read and write?
Perhaps these Kobolds have taken over an abandoned Dwarvish mine, fashioning some weapons from bronze that they smelt and cast with their rudimentary knowledge of such things. They’ll need a room for that. Maybe they mostly use flint weapons. Perhaps ceramic. Hide armors can come from something that they breed and/or hunt.
Kobolds are known for traps and poisons. Where do they get all the poisons they use? These ones milk snakes. Simple, primitive. Add snakes to the list of inhabitants. Perhaps there’s a room where the snakes are kept. Perhaps outside the warrens the PCs could happen upon a “snake-collecting” expedition of kobolds with baskets, a foreshadowing of the snakes and poisons that will follow when they attempt to invade.
- Air? (If they’re cooking their food with fires, where does the smoke go?)
Answering these questions causes one to reflect on “a day in the life”, and conceptually you begin to see the Kobolds as a functional system with inputs and outputs. Then, a short step farther back to see how it functions – that it functions – in the long term (Economics? Persistent food sources?) should round things out into something sensible and alive.
And as other creatures become a part of the picture, it becomes a recursive process: where do the snakes reside/come from, and what do the snakes eat?
Before long I have a tribe of Kobolds who venture from their abandoned mine to scour the desert outside collecting snakes, desert rats (to feed the snakes), cactus fruit, and water. They use riding lizards that they’ve domesticated because having chosen a fixed location, they need to range a little farther now to collect the things they need (desert life is harsh!) Perhaps this is why they’ve recently drawn the attention of the town that sent the PCs to find them. They mostly range at night, of course, because of the heat of the day. These lizards also serve as guards and as sources of food and hide armor. They collect and milk small venomous snakes using small ceramic pots which they make in a primitive pit-kiln, along with their ceramic arrow heads and ceramic blades for weapons. They collect what firewood they can, and their fiercest champions are gifted the handful of bronze weapons they’ve fashioned from leftover ores they found, as they’re not particularly inclined to mine the maze-like veins themselves, being far less industrious than the long lost tribe of Dwarves that founded the place. The caves include the entry area (well guarded with alternating patrols on day/night cycles or perhaps on rotating 8 hour shifts), kitchen, pantry, dining hall, barracks, living space, pottery room, nursery, toilet room, throne room, chieftain’s quarters (he doesn’t sleep with the rabble!), and armory. If invaders make it past their first line of defence, they can always release the Dwarven zombies – last remnants of the mine’s former inhabitants – which they keep contained in a locked pen near the entrance, ever vigilant and never need be fed. The hallucinogenic spores they collect from the Myconids in the back caves deeper in are received in exchange for bringing the fungus-men various bits of offal. These spores provide great entertainment and their agreement is a way to dispose of their waste, but the spores also make for some interesting poison effects when used by their most skilled shaman, who hurl tiny pots of concoctions to create small but potent gas clouds. The caves are rat-infested, which the Kobolds don’t discourage, regarding them largely as pets and a secondary source of food, as well as a great way to keep their snakes and riding-lizards fed. These vermin carry Sewer Plague because of the squalid conditions, but the Kobolds and rats themselves have long ago grown immune over the course of generations.
Now it’s not just a “dungeon”. It’s not simply a cave with some rooms and Kobolds. It is a living, breathing ecosystem, and it is immersive. We know what the Kobolds are coming with. We know what they’re carrying and wearing and how they made or acquired it. We even know what happens when the PCs stop for a rest and drink the water: sewer plague is the price of being an interloper.
Static Dungeons
Dynamic dungeons are living ecosystems, but what about those lost ruins that are not inhabited by living, breathing things, but instead are locked away from the world, snapshots in time from a forgotten age?
These are the static dungeons, and unlike the dynamic dungeon crawl, these places are intended as a series of time capsules. In these dungeons, the various bits do not necessarily interact. They can remain hidden away from the rest of the world, timelessly, because they do not require such interaction.
Static dungeons should therefore include only those things capable of existing for extended periods without fear of starvation, dehydration, or rot. These should be primarily populated with undead, constructs, and traps, particularly in their deepest parts. Anything living should more likely be nearer the entrance, on the premise that it comes and goes to the outside world, and merely lairs there in safety without venturing forth into the perilous depths. Also near the entrance, it makes sense for the PCs to encounter a few (inanimate) skeletons amid already-sprung traps: a warning of what’s to come. Perhaps the early rooms are already looted. The real loot surely comes only after the traps and inhabitants that haven’t been bested already.
Provided things aren’t airtight, one good exception to this exploratory timeline into the past is vermin. Tiny creatures that come in swarms can fit through tiny openings, so it’s very conceivable that insects, or maybe even rats, could come from world outside world to feast on the decaying flesh of whatever remains below.
This question of airtight is also important in another sense. The question of airflow also impacts moisture, and whether the area is dry or damp should impact descriptions of the conditions they find. Is everything rotting and covered in fungus? Or is everything desiccated and covered in cobwebs? This is something that could vary, if there is a reasonable explanation for these divisions: the Shriekers and Lacedons from the underground lake area are separate from the dusty old Mummy crypt by a pair of large, wax-sealed, air-tight stone doors. (Perhaps to keep in the poison gas?) But if you want to maintain the immersion, these separations must be explainable, not a random Hodge-podge. Examine the final map: there shouldn’t be a secret door to the sealed area, unless that’s got a wax seal too.
And in a broad sense, as the group delves deeper, the things they find get deader as well as deadlier.
SPOILER ALERT: EXTREME SPOILERS FOR CURSE OF STRAHD
The Amber Temple of Curse of Strahd has both good and bad examples of this at work. The barbarians near the entrance who use it as shelter but have never ventured inside? This works. Two pairs of untouched doors lie between them and the dangers inside, and their story is that they’re afraid to ever enter.
Vilnius works, but only to a point. The premise of his being there was to be that he escaped Neferon and went into hiding after Jakarion died. That means he must have arrived roughly at Day “PC-Arrival minus 1”. If the PCs keep taking long rests within sight of the entrance, his story grows more and more preposterous. Given how deadly Neferon is, that Vilnius was let live is already a stretch to begin with. If the PCs are taking their time exploring the place, a good DM would be well-advised to perhaps have him appear and make a run for the entrance.
The Barovian Witches downstairs do not work! Why would Neferon tolerate them? Why are they there in the first place? The same is probably true of the Death Slaad. Neferon, given his nature and as a sort of landlord of the place – this makes sense within the context that he’s long-lived (assuming Yugoloths live thousands of years). It’s a prize. It’s a feather in his cap. He’s an Arcanaloth and this place is a massive collection of the Arcane, and he’s claimed it for himself. But why is a Death Slaad simply chilling out in a side room, waiting for…. waiting for what? For my run, I’ve transformed the Death Slaad into an escaped Vestige, still bound by magics to remain near the room with the broken sarcophagus it escaped from (to Neferon’s amusement) and I’ve changed the Barovian Witches into an Allip and Will-O-Wisps. [Note: I believe I read that last bit suggested on a Reddit thread. If I can find it again I’ll give credit and link it here.]
The Castle Ravenloft of Curse of Strahd – almost comically – stands as both a tremendous exception and tremendous example to what I’ve laid about above. In the Castle, one can expect to encounter living souls – those in the employ of Strahd – as well as the undead – also in the employ of Strahd. Given Strahd’s nature this seems at first blush to make sense. But the moment the PCs are given an inch of room to breath, they may quickly realize it stretches the boundaries of disbelief: I will be completely unsurprised when one of my players jokes about how it must be perilous working there, such as being an accountant and having to move about the building with all the mindless Shadows lurking in the hallways. Vampire Spawn makes sense. Even Wights possess an intelligence that means they could follow Strahd’s instructions. But how are the Shadows kept at bay?
And from back in the 4th Edition adventure path, I still want to know how that Centaur Ravager got to the 2nd floor of Sarshan’s tower, Mr Cordell.