View Full Version : DM Help Dungeon Layering

2015-09-21, 10:34 AM
I've struggled to make good dungeons in my short career as DM. Often, it ends up being an encounter in a rectangular room followed by another encounter in a rectangular room behind an unlocked door.

My dungeons lack fluidity and nuance.

What advice would you all offer to a young DM struggling to create great dungeons?

2015-09-21, 10:44 AM
You could allow the design to inspire you. Start with what sort of place this is, what is its function(s)? Or what were they when it was originally constructed? An ogre mage outpost would be quite different from a hobgoblin burial chamber. Then design your dungeon with that in mind. If it's a ruin, you can leave parts closed off, especially if it's getting too big for your wants.
You can look up plans for various kinds of construction, and you can find lots of dungeons to inspire you.

2015-09-21, 10:56 AM
A dungeon is what happens when you distill the essence of roleplaying into its most basic form: making choices. "Do you go left or right?"

Make sure there are enough clues for the players to make informed decisions. Some guesswork is fine, but if you don't give them any information about what is left or right they might as well not be choosing at all.

Make sure their decisions have consequences. If the players can go left and then backtrack and go right without a hitch, their choice didn't really matter all that much.

Make sure their decisions are followed by more opportunities to make decisions until it's all a confusing mess of lefts and rights and ups and downs and all sorts of encounters in between. Multiple entrances, point A connecting to points C and H, traps that move you about and make you lost, encounters that chase the players around and close doors and flood chambers. That's when you're in a dungeon, not just walking along some corridors.

2015-09-21, 11:56 AM
Put things in those rooms.

A ledge where the goblin snipers can hide with there slings, balconies agile characters can jump from to get away from meele brutes, underwater passages filled inexplicably with maneating eels that are the only conection to hidden terasure rooms.

Make the monsters and the dungeon work together as a challenge. Put encounters at the bottoms of pit traps. Have the last of the kobolds flee into tunnels whose traps she is too light to trigger. Have your monsters ever opened a door to call for friends in the next room, or closed a door to try to divide and conquer?

Players will likely forgive linear paths if you put very interesting encounters along the way.

What they want to do is be presented meaningful choices. Left or right down different-but-identical hallways won't be remembered. Towards the sound of large animals snoring down the hallway littered with blood, or away from it? That can have meaning.

Since encounters take more time than exploration, focus on giving encounters with multiple solutions (although not necessarily equally good solutions) first, and then find ways to make choices matter in the maze.

2015-09-21, 12:17 PM
Unless you're playing in one of those novelty "dungeon world" settings, complex structures don't just spring up into existence. Someone built it, and built it for a reason. Emphasize to the players either the history of this place, or the fact that they don't know the history of this place; in the latter instance, leave clues of various kinds. The dungeons that have been responded to most positively in my group have been the ruins of ancient civilizations the players hadn't even heard about before, with only the slightest clues as to what happened there.
That said, it's important that you know what happened in a particular place, even if the players will never hear about it. Knowing the hidden background information behind something is important, even if it never comes up in actual play, because it helps make informed, reasonable decisions about things that do come up in play: what objects are or aren't in a place, what kinds of writing is scattered around, how NPCs react or enact their short- and long-term goals.

2015-09-21, 01:08 PM
This is nice advice so far, so thank you. I never make dungeons because they always felt so forced, but I took the story-telling aspect of it to heart and am starting to put some cool stuff together.

The dungeon I'm creating right now was created by Cyric's followers to hide the Cyrinishad before he was imprisoned by Tyr and Mystra (I'm kinda editing forgotten realms stuff to my own world). Cyric spread lies and conflict among the gods of the realm while building a following in a massive, mayan-esque civilization in a deep jungle. The conflict came to a close with the deities warring in this civilization and resulted in its destruction and Cyric's imprisonment. Far into the wild jungle, the dungeon sits under a massive tree on the edge of a lake.

So, I'm thinking the surface level - the easily accessed part of the dungeon - will house lizardfolk; their tribe has slowly taken to worship Cyric, subconsciously compelled by the residual magic deep into the cave. It's pretty simple, but it shows signs of more intelligent crafting: support pillars made of cut stone, a pair of slatted iron gates on a thick chain pulley, and Cyric's skull design sculpted into a rock wall. There will be three main rooms and an inaccessible stairwell behind the gates. Beyond the third room, which appears to be a naturally created cave, will be a fancifully constructed room of marble, and a descending circular stairwell.

That's all I've gotten to this afternoon, but I think in the deeper levels it will become a more expansive dungeon. I think given Cyric's nature there will be less straight fights and more traps, puzzles, etc. Maybe if they fail to complete a puzzle they can find an alternate way to raise the gates and continue on that way. I'm thinking that only magically strong beings would have been able to delve deep into the dungeon and fall victim to the Cyrinishad... maybe a coven of Hags? That would be interesting I think.

Given this info, what kind of monsters should I add? Party is nearly level 4. I was thinking inanimate objects that could have survived for centuries: Helmed horrors, animated armor? A basilisk sounds cool. Are basilisks long-living?

2015-09-21, 02:43 PM
Add a secret entrance to the second (or even third) level! This gets you both an alternative point of entry or exit in case of emergency (be sure to scatter hints about this all around the area - "too obvious" is nearly impossible to achieve!) and a convenient explanation for why all these monsters haven't died of old age after all this age.
If you struggle thinking of why this additional entrance would be present at all, here are some ideas: It could have been a secret entrance for the priests of the cult to pull off a spectacular appearance in front of the crowd, or an exit tunnel for the same guys in case the crowd suddenly wasn't so happy anymore. Alternatively, parts of an ancient network of tunnels could have collapsed, opening a new access. A third option would be looters who tried to dig their way past the lizardfolk and guardians (their corpses/skeletons could also provides hints about traps or other dangers in the temple).

As far as puzzles go: Don't overdo it. In addition to easily becoming the game of "guess what the DM is thinking", some players just don't like puzzles at all.
That said, try to combine those puzzles that make thematic sense with an encounter. Figuring out which lever to pull is much more entertaining if something tries to gnaw your foot off while you do it.

2015-09-21, 03:13 PM
A related thought crosses my mind: Is there a good, realistic way to use traps except to torture players?

2015-09-21, 03:30 PM
@b4ndito: The use of traps by intelligent creatures who expect an attack should somehow assist the trap-makers without interfering with their lives. Traps won't stop invaders on their own. What traps do is either (a) delay the invaders long enough for a counter-attack, (b) sound the alarm, or (c) give already-present guards a tactical advantage.

Examples of a type A trap: an obvious pit trap, or those scything-blades-down-a-hallway things. Designed to stymie or slow down invaders. These kinds of traps also serve at deterrents, which means they are the ones most likely to be immediately obvious.

Examples of a type B trap: a kobold trap with a wasp's nest attached to a trip-wire. Turns a would-be sneaky invader into a loud, panicked invader, and draws the attention of the kobold guards. The simplest of these are a plain bell attached to the back of a door, which will ring loudly when the door it opened.

Example of a type C trap: a weighted net woven in with barbed hooks, set in the middle of a wide corridor. Goblin guards are set up at the far end of the corridor with ranged weapons and cover, encouraging attackers to charge at them. They hit the pressure plate and are "attacked" by the net, which stops their movement and prevents them from either attacking or fleeing, allowing the guards to pincushion the attackers at their leisure. These traps can be more complex, but in the end, their only purpose is to change the terrain of a combat site in order to benefit the trap-maker.

2015-09-21, 03:37 PM
A dungeon should not just be cages with monsters lined up.

Come up with a reason for the monsters to be there. You don't have to tell your players, but this will guide the conditions of the fight and the surrounding area.

One monster might be a wandering creature that's set up inside an abandoned section of the dungeon to lay eggs or give birth. This creature would patrol its territory, hunt for food, and generally leave signs of its presence (which lets you ask for checks to slowly clue the players in on what they'll be facing even before they see the creature). Monsters of this type will be territorial and fight other denizens of the dungeon.

A completely different monster would be one that was enslaved or domesticated by residents to guard a place (think the Rancor from Star Wars, or your typical guard dog). Monsters of this sort will occupy a single room, perhaps with weaker, intelligent handlers. They'll be present inside a choke point that is difficult to sneak by, but there may be various tools of restraint (such as a cage that drops from the ceiling, or magic soothing items) that the PCs might be able to find and use to take it down.

Traps are related to this problem. You'll also want to think of what the traps are doing there, but remember to think of why they are there from the meta-game perspective too. If you're adding traps just to have traps, nobody has fun. Traps should serve a purpose more than being an HP tax.

Alarm traps are interesting, because the consequence is interesting. PCs may wish to trip one intentionally to lure enemies into an ambush they set up. The enemy may try to trip them when encountering the PCs, and if the PCs spot the trap, they will have to try and stop the baddies from calling for backup. Traps are excellent as a terrain feature in encounters, but rubbish as encounters in their own right (similar to locks) because the outcome becomes "you spend X time to disarm it and keep going" which is boring. Even in abandoned tombs, constructs and undead provide an excellent trap companion.

2015-09-21, 04:08 PM
A basilisk sounds cool. Are basilisks long-living?

Have you established previously that they aren't? It's your world, and the lifespan of a monster is one of those things which is easily changed between settings without causing problems. A basilisk has a slow metabolism and a method of long-term food preservation, which suggests they aren't particularly short-lived; unless you've already mentioned that they only live for 10 years or something to your players, they can live as long as you want them to.

2015-09-21, 04:24 PM
Since I have yet to DM, I can only offer a mix of personal anecdote and advice I've looked up for myself.

One thing I loved about exploring the dungeons in Tomb Raider: Underworld was the 'accidental' asymmetry. You could see where the rooms would have been identical if not for the ravages of time. I feel like that extra aspect of seeing how the room has changed gives more life to the dungeon.
It is also interesting when what should have been a short walk becomes a challenge, while elsewhere a deadly trap is easily bypassed.

The reverse approach is also possible. Think entirely of the purpose each room serves for the metagame.
You design your map with amorphous blobs of rooms with patches of game features.
I want some rough terrain here ,', the masonry has collapsed.
The room turns here ',' It will force the archer to move deeper into the room.
Of course, you come up with excuses for why the builders made the dungeon as they did afterwards.

Don't be afraid of starting every room as a simple rectangle; it's probably the most common shape to use in architecture.
What you do is deform the rectangle. Maybe they needed one end to be thinner to fit limited space. Maybe the wall between this room and a neighbour has fallen down.

I'd say it doesn't even matter if every room is rectangular; it's the encounter as a whole that gets remembered, and there are plenty of other ways to add variety.

Or just build dungeons as you always did but, when you see how many rectangles you've used, cut off some of the corners. Bam, octagons!

Mark Hall
2015-09-21, 05:29 PM
Consider who lives there, and how they interact. If you have room A and room B, and the folks in Room A get in a noisy fight, how does Room B react, and why? Do they rush to help their friends? Do they set up an ambush?

One of the earliest, best, dungeons is B2: Keep on the Borderlands... specifically, the Caves of Chaos. There are several unconnected dungeons in the area, each housing a different kind of humanoid (two tribes are orcs), including one with an ogre, one with a minotaur, and one with an owlbear. If you look at HOW these groups interact, youc an make a far more dynamic setting than a plain, keyed, dungeon.

2015-09-22, 02:27 AM
One thing to consider with dungeon design is the overall 'type' of dungeon crawling experience you want to create, and how it interacts with the players' tendencies. For example, are your players completionists who will want to go into every room, or will they be satisfied missing some stuff? You can then design the dungeon to either play into that, or create tension against that, depending on how you want it to feel.

For example, there are different layout types that you can use: linear, linear with side branches (in that there's a given 'last room', but not every direction leads there), 'bundled' (that is, linear at large scale but various choices of path that all lead to the same place), 'true' branching (no 'last room', but just many different areas to go to), tightly looped (so there isn't a clear 'forward' direction at any given point), bypass-based (so that there may be some rare ways to bypass large sections of the dungeon), etc.

Linear with side branches is a common design because it combines two useful things: on the one hand, the overall linearity means that the players always can know where to go next to 'finish' the dungeon when they're tired of mucking around. On the other hand, the existence of side branches means that the players have meaningful choices about where they want to explore, which can grant additional benefits with respect to the linear segment. So this suits a group of players that doesn't really love exploration for exploration's sake, but likes having choices.

On the other hand, the explorers' dream dungeon is something like 'true' branching with a bunch of bypasses and loops. It gives rise to lots of moments of discovery, and if the players can have some knowledge of layout ahead of time, they can also make lots of interesting route-planning options ('so, if we go this way, there will be lots of guards, but if we go that way as long as we can fly and deal with the lava, we can get to the troll king's bedroom without going through inhabited areas'). But a group that has trouble setting its own goals and direction will have a lot of problems with this kind of dungeon, and may just decide to leave it because its hard for them to know what there is to do here and when its 'done'.

Another thing to think about with dungeon design is to consider ways in which the overall layout can matter once sections are 'cleared'. Generally its boring to re-run encounters that the party has already defeated, so people usually make it so that the cleared sections of dungeon don't refresh. That means, every room that has been cleared sort of joins into one big node in the dungeon layout which is the 'places the party can get to trivially' node. It can be useful to mess with that expectation in ways other than just the monsters refreshing. For example, one thing you can sometimes do is to have one-way transitions. They're not necessarily as gimmicky as a true 'one-way portal' or 'one-way door', but an example would be something like a low-level party dropping down a long chute without the gear to make it back up. The result can then be that finding a loop or some second way around can actually still be important, because being on one side of the 'cleared areas' doesn't always guarantee a safe path to the other side of the 'cleared areas' through known territory.

Similarly, things in the dungeon which change the layout or change aspects of passability can create large-scale strategic considerations. For example, lets say there's an underground river, and with a little work the party could redirect it in order to flood different sections of the dungeon. The result might be that certain rooms get automatically cleared, but that pathways that the party has relied on previously become closed or inconvenient.

2015-09-22, 08:10 AM
One thing I loved about exploring the dungeons in Tomb Raider: Underworld was the 'accidental' asymmetry. You could see where the rooms would have been identical if not for the ravages of time...

This is really good advice. Your architecture tells a story, think about what it is or it ends up being "we built a dungeon just for adventurers".

I'd say it doesn't even matter if every room is rectangular; it's the encounter as a whole that gets remembered, and there are plenty of other ways to add variety.

This is exactly what I was saying. :-)

2015-09-22, 12:06 PM

Of course, having dungeon designs where you could skip a large number of monsters through careful planning works better for systems where XP gain isn't directly tied to how many monsters you face. Otherwise, the PCs will feel forced to backtrack to clear the whole thing out.

2015-09-22, 12:32 PM
Of course, having dungeon designs where you could skip a large number of monsters through careful planning works better for systems where XP gain isn't directly tied to how many monsters you face. Otherwise, the PCs will feel forced to backtrack to clear the whole thing out.
No PC choice should be between something and nothing. Avoiding a monster could mean having to fight a weaker monster, bypass a trap, or some other challenge that gives an equivalent XP reward.

2015-09-22, 12:58 PM
No PC choice should be between something and nothing. Avoiding a monster could mean having to fight a weaker monster, bypass a trap, or some other challenge that gives an equivalent XP reward.

It depends on how obvious the choice is. I'd say that in general most games need (and will naturally have) choices between two things both of which have upsides and downsides, but there's also absolutely room for good choices to bypass difficulties entirely that other choices would still have. There's also room for dumb choices to add difficulties that the GM absolutely didn't plan on having in the first place.

2015-09-22, 08:08 PM
This has been helpful, many thanks to all who offered advice. I'll post the result of my dungeon after its completion.