Category Archives: Dungeon Design

20 Architectural Features for Memorable Dungeons

King's Quest 6 Labyrinth Copy ProtectionThere’s more to crafting a memorable dungeon than the room descriptions. To be really great, it should be an interesting space to move through on its own merits. The way the rooms connect to one another, and to the outside world, is a fundamental part of any dungeon’s character, which I have too often ignored in the past.

I’m no Dyson Logos or Stonewerks. My maps are best described as “serviceable.” And while I doubt I’ll ever share their artistic acumen, I’m trying to do better for the sake of of giving my (amazing) room keys nicer homes to live in.  To that end, I’ve set myself the goal of having 3 architecturally interesting elements in every dungeon I put together. Things that give the dungeon space an inherent complexity for the players to struggle with, or manipulate.

It is essential to note, that complex room shapes are not interesting dungeon design. They are a lazy way to make a map look interesting, without actually being interesting. Unless the shape of the room has some particular impact on play, more-or-less squarish spaces are all you need.

  1. The entrance to the dungeon is in the center, with rooms radiating out around it. Rooms might interconnect freely, or form distinct “wings,” with only occasional connections between them. The player’s options are maximized from the start, and if they are stymied in one direction, they can easily try another.
  2. The entrance is perilous, preventing quick egress. Perhaps getting in and out requires climbing 100′ of rope, or slowly wiggling your way through a narrow crevice in the wall. The players are thus limited in what they can bring in or out, and will not be able to flee to the safety of the outside world if they are being pursued.
  3. The main flow of the rooms form a ‘figure 8’ pattern; meeting in the center, and forming two distinct loops. This is a nice simple way of making the player’s path through the dungeon nonlinear, without making the rooms overly interconnected. Each loop, of course, could and should have little offshoots.
  4. A river of water, lava, or just about anything else flows through the dungeon, intersecting with multiple rooms. Not only does it serve as a point of reference, and as a way to make the individual rooms it passes through more interesting; it could also be an alternate way of moving through the dungeon.
  5. A space from which another space is visible, but not obviously accessible. When keying, this latter space would have some interesting feature the players would want to interact with, but be barred from doing so until later. This could be accomplished with bars, with a sharp change in elevation, through unbreakable windows, walls of force, large uncrossable lava pits, etc.
  6. Connections which are spatially impossible. Doors that lead to the other side of the map, or hallways that loop back on themselves without ever turning. This need not be presented as a “gotcha,” where the players explore for a long time before realizing what is happening. Reference points set early in the dungeon could clue them in very quickly, and give them an opportunity to use the dungeon’s geometry to their own advantage.
  7. A room which vertically intersects multiple dungeon levels. Perhaps with bridges spanning back and forth across it, allowing players to reach other parts of the dungeon by finding some way to get up to a higher bridge, or down to a lower one. Or perhaps the room is filled with water, and players have to swim through it. Or any number of other possibilities.
  8. Areas on one level, which only connect to one another through a different level. So, for example, the first floor might have 10 rooms, but only 5 are immediately accessible. To reach the other 5, you’ve got to go up to the 2nd floor, adventure through those rooms, and discover stairs back down to the other half of the first floor. Vertical movement is often unexplored in dungeons. Usually, there’s only a single stairway between one floor and the next. Rather than the players making interesting decisions about their movement in vertical space, this method basically creates multiple dungeons strung together end-to-end.
  9. Combining natural and crafted spaces. This is already done with a lot of maps, where the dungeon is built on top of caverns. This particular arrangement, though, has become a little cliche. It’s more interesting to have a walled garden, open to the sky; or have a crafted corridor that opens out into a natural cave, where the players can choose between a few natural and crafted exits.
  10. Secret doors which do not lead to secret areas, but rather, lead to other areas of the dungeon that could be easily accessed conventionally. This was actually much more common with dungeon cartography in the ’70s and ’80s, but is not as common anymore. Which is a shame! Yes, rewarding players with secret treasure is good, but it’s also good to reward them with secret connections they can use to move about the dungeon more sneakily.
  11. The dungeon lacks any foundation. Perhaps it is flying, or suspended over a chasm by chains, or floating on a fantastically buoyant sea. The floor of the various dungeon spaces have frequent openings, which serve as hazards for navigation, a means to dispose of any foes the party encounters, and perhaps even as a secondary form of navigation if the party is bold enough to try and cling to the underside of the dungeon.
  12. The dungeon is moving. Getting in is a challenge, as you either need to find some way to catch up to it, or you need to predict its route and jump on with perfect timing. And, once it’s time to leave, you have no idea where you’ll be. Such a dungeon could take the form of a castle-on-wheels, a giant walking robot, a structure carried on the back of a titanic beast, or carried in the talons of a massive ancient bird.
  13. That a dungeon should have multiple and varied entrances is advice I remember hearing years ago. But it’s still good advice, and I so rarely see it followed. Multiple ways in and out of a dungeon offers a lot of interesting gameplay. For example, if multiple entrances are known, players can investigate both, and make a choice about which they want to explore further. If the extra entrances are hidden, players may discover them from the inside, creating a natural sort of “save point,” now that they can return to town, and re-enter the dungeon further along than they started. Dungeons may also exit out into new areas: vast caverns beneath the earth, mysterious forest groves, surface temples, or deserted islands.
  14. The players cannot get out the same way they got in. Perhaps the door seals behind them, or perhaps the entrance to the dungeon is a trap door down a slippery shaft. The crawl takes on a sense of urgency when it’s no longer possible to leave at your own choosing; and resources normally taken for granted–food, water, light–are transformed into timers, counting down how long the players have until their work becomes exponentially more difficult.
  15. The dungeon has some internal means of rapid conveyance. Perhaps there is a train, or a set of color-coded teleportation pads in the first room, or a pipe which–when touched–causes a person to merge with it and flow forward until they will themselves to separate.
  16. The rooms, when mapped together, form some kind of shape. This shape is a clue to solving one of the dungeon’s mysteries. As a simple example, perhaps there is a set of buttons in the shape of a circle, square, and triangle. Pressing the right one opens the door to the treasure room, pressing the wrong one kills you. The correct one is the triangle, which the players can guess based on the fact that the dungeon is a triangle.
  17. The dungeon has an open-air layout, with sections of it being completely physically separated from one another, despite having internal continuity. Players can enter any room of the dungeon they wish, but some rooms can’t be fully realized until other rooms are investigated.
  18. A specific calamity which has changed the dungeon’s layout. Many dungeons have bits of crumbled wall creating openings here, and collapsed corridors creating walls there. More interesting, I think, is to determine an event which caused structural damage to the dungeon, and allow that event to alter the whole layout, rather than just a corridor or two.
  19. Puzzlebox dungeons, with big moving parts, where whole wings are locked off until some challenge is resolved. Something like a big cog or water wheel, which the players must discover how to turn; or a button they must weigh down. These, in turn, cause a statue’s mouth to open, and more rooms to become accessible. This is the sort of thing commonly found in video games, which would be all the more interesting for being presented in a situation where the players have real agency.
  20. A dungeon which serves as the habitat for large groups of some benign creature. Preferably, one with some notable effect that will have an impact on how the players navigate. Perhaps the creatures can be ridden, or perhaps they screech loudly when they see light, or maybe they create an anti-magic field around themselves, or they cause magic to be amplified, or they produce weird results when eaten.

Flux Space in Dungeons

Nonspecific Dungeon SpaceBefore there was On a Red World Alone, there was Dungeon Moon. I spent hundreds of hours working on on that game, and to this day I love it as much as anything I’ve ever done. For years, friends have been gently pestering me to get it written up and published, which I would like to do someday. The only problem is that Dungeon Moon was an unplayable mess.

My whole approach was based around striving towards this false ideal of a fully realized megadungeon. Even the least important bits of space beneath the flagstone surface of the moon had multi-paragraphs long descriptions. “It’s a kitchen” was never good enough for me. I had to figure out if there was something special about the pots and pans, or if maybe there was a secret passage that led to a trapped treasure vault.

Likewise, the map was as much a labyrinth for the referee as it was for the players. I spread it across a stack of graph paper a quarter in thick, with alphanumeric codes written on the corner of each page to identify which “column” and “level” it depicted. More than once, a single room had to be spread across two pages just to maintain the geometry of the thing.

All for what? Despite my extensive prep work, a group of adventurers could walk off the edge of the map within an hour if they made the right choices. And as bad as that sounds, it would probably make the game better. If the players are off the map, the referee would have to improvise, and whatever they come up with in the moment has gotta be better than pausing for 5 minutes at every door to cross-reference maps and read room descriptions.

I certainly seem to have made it work back in the day. (People don’t show up every week to play in a game that they hate). But it could have been better, and if nothing else, the excessive notes killed my own enjoyment. It was murder, trying to keep up with the pace I had set for myself. If I’m ever going to run it again, Dungeon Moon needs to go easier on the referee.

Dungeon Moon Mk1 MapIt’s not just a matter of drawing simpler maps and writing shorter room descriptions, though. The problem of scale is inherent to the setting: its a dungeon which literally fills the entire internal structure of a moon. Constantly branching pathways and infinite expandability are built into the premise. A more manageable size would ruin the setting just as surely as my bad notes did.

Yes, dungeon moon needs shorter, table-ready notes, and a map that doesn’t have to be laid out across the kitchen floor to be viewed properly. But it also needs to feel huge and interconnected.

Recently, I was puzzling over this problem, and recalled a conversation from years ago. I must have been complaining about the issue, because  Gus L. told me I should try running Dungeon Moon as a point crawl. At the time, I kinda blew the idea off, because I didn’t want to waste all the work I’d already done on my maps, but with a few years of distance, the idea is way more appealing. Sorry for blowing you off, Gus. You were right.

In a point crawl, the referee maps a wilderness environment as a series of locations, and paths between them. The players don’t just wander straight towards their objectives, instead following roads, or deer paths, or whatever else they can find. At intersections, (the titular “points,”) they come upon something interesting. They can choose to engage with what they found, or continue on past it to the next path.

Flux Space is a way of doing the same thing for dungeons. Megadungeon feel without megadungeon effort. The big difference, though, is that the paths aren’t just direct connections between points. They’re a series of corridors and rooms just like everywhere else, but they’ve been abstracted to keep the game moving and to make the referee’s job easier.

Like a point crawl, there are two basic building blocks here: Locations, and Flux Space. The locations work the same as any dungeon: there’s a map, and there are notes which describe the map’s locations.

To keep things manageable, a single location should be able to fit on a single sheet of graph paper. Locations also should not directly connect to other locations. Some exceptions can be made for areas that play with vertical space, or for secret shortcuts to other locations. But, for the most part, players should need to go through some Flux to reach a new location.

So what is Flux Space? It’s an abstracted section of the dungeon that exists to connect locations together. It’s mostly hallways and empty rooms, without any specific layout. Each section of Flux has three elements: a description, a size, and an encounter table.

The description is a vague idea of appearance, which remains consistent throughout. Something like “Worked stone,” “Oozing walls,” or “filled with garbage.” This gives each section of Flux a distinct personality, which the referee can use anytime they need to improvise some specifics for it.

The size of a section of Flux Space is just a number. In order to pass through to the next location, players will need to roll that number of encounter checks. Once they have, the referee randomly determines which of the connected locations the players emerge out into. Usually, there shouldn’t be any chance for players to wind up back at the location they started from. Optionally, though, that could happen if the players roll a “lost” result on their encounter die.

The encounter table for a Flux is like any encounter table. It has some locations on it, some wandering monsters, and probably some creatures from its connected locations. (If the Hall of the Gnomes is connected to a Flux, that flux will have some gnomes wandering around in it).

Once a group has encountered everything on the encounter table, that space is considered “mapped.” Players can move through mapped fluxes with only a single encounter roll, and may choose which of the attached locations they emerge into.

The overall dungeon is depicted as a sort of spiderwebbing flowchart, showing how all the locations and fluxes connect with each other. Hopefully you can do a better job of coming up with reference codes than I did in the example above.

As a final note, I want to make clear that when I say Flux Space is ‘mostly empty,’ I do not mean that literally. What I mean is that it’s mostly devoid of tricks, traps, monsters, or treasure. There may be bedrooms, or gymnasiums, or warehouses, but none of it is valuable, or interesting, or trying to kill the players. (Unless it’s on the encounter table, of course).

So if your players ask about where they are and what they see, don’t tell them it’s empty. It’s not empty. It’s just boring compared to moving on to other locations.

Related Posts Plugin for WordPress, Blogger...