How to build a dungeon?

A little question by Sam about how to create an adventure (as an excuse for using the Random Esoteric Creature Generator) for LOTFP lead me to write this answer. Sam’s point was about “creating an adventure that is a garden for emergent story”. I was in this exact situation a few months before and this question is still an open one for me. Thus I don’t pretend to be any expert in this – but here what I have done as my first answer.

Making the dungeon

1) considering the regional context. Mine is the English Civil War. From this I had two obvious possibility for a first dungeon (there are more, but those are obvious): abandoned monastery/churches in various degrees of ruin (because of a royal decree of dismantling abbeys a century before) and the same with castles (because of the war)

2) putting the dungeon somewhere. I thought an underground Stonehenge maze but finally choose “the Tomb of King Arthur” based on the legend of Glastonbury Abbey who is in ruin. I thought it would be inspiring to think about how the Arthurian myth mix with LOTFP weirdness while building the dungeon, without any plans

3) picking an interesting map. I just stole the Holmes Basic 1977 Sample dungeon because I liked it – after wondering about a few Dyson map that could fit ver well too. Some variables guiding my choice: not to linear (you can directly go into three directions from the entrance) – there is no obvious path choice. Also the room are huge, so when you enter most of the room with torches, you don’t see the walls just by entering a room – you need to explore it further to see them. There is a 4-doors room with a “statue” that open the door it face. I liked it because I imagined players using that as a way to avoid the terrible Esoteric Creature I would use.

4) I made the creature. I ended up with a six legs/four arms skin-translucid wolf-centaur (instead of horse) with a wolf’s head who was capturing things in its lair to impregnate them and that could spit biting leeches (you could see them living in a nest in its translucid organ). I called that thing “The Lady of the Lake”. I also put the Abbey near the sea without any consideration for the real Glastonbury Abbey consideration

5) I used the real map of the Abbey as a “surface level”. I stole Ron’s idea and put monk lepers living there in the remaining buildings. I used a mental remake of those two pictures :

Une image contenant roche, extérieur, nature, rocheux

Description générée automatiquement

Here is the sample dungeon map remade by Dyson :

We are getting in the meat of it now

6) I opted for more than one entrance. So I decided there was a pit, an entrance from the lake, and the “main” one of the map (near C) was the crypt of Arthur in the crypt of the Church itself. Then I decided where those entries lead in the dungeon map

7) I made a random tables of wandering things. Not all of them are necessary monsters or things to fight. Just thing there and we’ll see in play with a reaction roll or any procedures. I selected possible social groups in the region and historical period. It’s the english civil war so I putted: 2d6 Royalist Cavalry and infantries, 1d3 Roundhead (Parliament) cavalry, 1d6 lost/captured Quakers?, 2d3 Crazy Monks, 1d3 Ill impregnated monks, The Lady of the Lake, The Lady’s Babies (from dead monk bodies), The Lady’s Lair, some insect swarms, and bizarre stuffs but not too heavy surnatural monster that would put the Lady in shadow – that I picked up from Veins of the Earth (like ultraviolet papillons).(modifié)

8) I made a random table of “colour” things related to this table or picked up from things I’ve read and like. A chained skeleton with 8 arms, Monks with open ribs cage, an infinite chasm leading to hell (you can hear the screams). So I would “fill” empty rooms with those things that are just colour and not have just empty rooms.(modifié)

9) I made a similar random table for treasures with things that could fit this particular dungeon

10) I used a procedure Tommi sent me and described here:

In summary, my explicit choices of variables were those: non linear map, huge rooms you can’t scan just from the door, regional and historical context as inspiration for random tables, random tables specific to this particular dungeon for encounters and treasures, different social groups in presence that could lead to multiple interactions (with some antagonism between them), two levels but only on interesting on, many ways to get in and many ways to get out.

How it played

Ross asked me how did it play – and I’m happy to answer.

It was a first timer for new roleplayers who never played a dungeon and on of my old “cautious player” buddy. We played two session of a few hours. It took a lot of time to enter the dungeon when they encounter the lepered monk. Character creation and entering the dungeon took 2 hours for 4 players.They got out and tried to overthink, plan and strategize to the point I pushed them to make a decision. I’m wondering if, for this group, knowing they were new and knowing the tendency for cautious play of my buddy, it would have been better to put them directly in the entrance from the crypt without this surface threat. Between the first and second session we had a discussion about this with some players.

They saw fleeing Royalist Cavaliers fleeing from the surface and two of them thrown in the dungeon from the pit. They ran through the ruins and found Arthur’s tomb in the crypt which lead to the “Main entrance (C). I kept four skeletons in the room B from the original dungeon, hidden in niches in the wall (it was still experimental) – there was a combat but not an interesting on. They met a Royalist Cavalier who had seen the Lady of the Lake and was afraid to death describing the lair and the thing. They made a few rooms and ended up in on of the biggest (K) where the saw a weird altar made in an old meteorite stone and impregnated monks praying there an image of the Lady (grotesque statue, no way to know the were impregnated). They stole a weird lantern with crystalline chrysalis that seemed to have value. The Lady appeared (from a random roll) and they all an as hell in the opposite direction – they ended up where they start so they got out of the dungeon with the Royalist Cavalier who was able to leave. I told them if they want to come back, the Royalist would tell the army and the army would raze the thing in 2D6 days.

I think it was a learning game for them: “you don’t need to agree with the other players to make a choice”, “interacting with the situation is rewarding”. For instance, my “cautious” buddy (a mage) basically carried the torch for the group the whole thing, while other players, after the “planing” thing and a few discussion about that interacted with rooms and actually got the treasures without difficulty (but lots of anxiety).

Some things helped. Ron’s description of how he played the wolf fight in its Ottoman Empire LOTFP game, and how to describe things after the roll. I used that in the first combat and even if the encounter in itself was not incredible, it was not boring and I think it helped to get a sense of how narrative authority was distributed. 

Between the two sessions, one players told me he felt “pushed in a direction” – something I didn’t feel doing myself. He told me some things like “I wanted to run in the crypts, but I could not leave the group, it would not be cool for them”, “those monks was a huge threats and were getting around us, so it was like we see the threat then could not react” (they weren’t moving or doing things). I asked a few a question lik “why could you not make your choice ? Do we need to agree with the whole player group about every character’s choice ?”. At the second session – two players were not anymore discussing but entering the rooms while cautious players were always “waiting and observing what happens”.

The random table and reaction rolls were really interesting for me. They met this affraind Cavalier and at first were threatening and intimidating – but one of the player described at character creation that he had a Royalist Cavalier armor, so the NPC saw them as a way to get out of the dungeon and acted and interacted with players by pushing this only agenda (“gold is nothing if  I’m dead, let’s get out of there”). At some moment it created a tension because some of the players wanted to explore and he wanted to leave – leading to a Tarantinesque scene with weapons pointed, but it was defused by one of the player.

I’m really interested to know other answers to Sam’s question. I’m also really thinking how to do it with D&D 4e, having watched the three first Barbaric Psychedelic videos and I’m amazed of what Ron did – but really don’t know how to do 🙂


13 responses to “How to build a dungeon?”

  1. Building a Dungeon!

    This is a really great post, thank you for taking the time to do it. It made me realize that I need to focus less on populating each room of the dungeon I have been working on, and let the dice handle that with some custom tables. What I want to focus on is making sure that any random thing that could be encountered isn't just thrown in there for more content–I love how your random table comes straight from the setting concept and the larger events of the world. The thing that I think seems really key is actual people with opinions who do not have any predefined relationship with the characters. I am kind of tired of the "here are the good guys, they help you; here are the bad guys, they try to fucking murder you on sight" thing. Like, we have monsters for that. 

    Once I am able to complete the dungeon and play it I will make a reply to this post describing the process. 

    • Sure!


      I use the random thing as a skeleton then I try to make sense of everything – a bit like Circle of Hands does, or Legendary Lives character creation.

      So it's not just "*roll* 3 crazy monks in a room and a treasure *roll* a valuable lantern", but "3 monks are living in this room.. So I'll put an Altar, and some couches, with wood box and cloathes with junk on it, a lantern with a weird crystalline chrysalis that seems to be valuable but also shine cast a light that affect the mood. They believe that they are chosen by the Lady but they are despised and terrifying for the monks in the surface – who throw one of them sometimes to appaise th Lady. They pray in the morning, they write weird painting in the afternoon and they do this and this after that [So I will know more or less what they are doing depending on what time the PC have chosen to enter in the dungeon and they took to arrive there]".

    • I am kind of tired of the

      I am kind of tired of the "here are the good guys, they help you; here are the bad guys, they try to fucking murder you on sight" thing. Like, we have monsters for that. 

      I'm using LOTFP procedure for that, and I must say I'm sometimes influenced by Holmes's reaction table (which is more "they are ready to negociate something" – you could hire the goblins!)  – so sometimes, even the Monsters are not a threat for everyone. If it's not obvious that the Monster encounter is a fight – for instance the hungry Monster just ate something and the character are not obvioulsy swinging weapon with rage, a Monster could even get a reaction roll.

      Or a weird f*kin monster could avoid a chaotic character while trying to eat the Lawful Cleric.

      I realize a thing is missing in my procedure for this dungeon: traps!

  2. A cross-reference

    Some companion reading: thoughts on what the term "dungeon" even means, in the context of the extensive Fourth Edition game we played a couple of years ago: Dungeon is as dungeon does. It's a good context or framework for this post, I hope. – wow, re-readng the comments, Ross totally kicks ass there.

    For practical comparison (with useful similarities), see also The House of Wisdom, in which I was GM and Sean was a player using Lamentations of Flame Princess; the post includes my notes and maps.

    • Oh! And if anyone here isn’t

      Oh! And if anyone here isn't familiar with it, Tony Dowler's How to Host a Dungeon is a building-system for these things which inspired a wide range of other authors. The linked site includes a free demo version.

      Another point of comparison is James Maliszewski's Urheim, as he takes this cultural/historical approach seriously.

  3. I thought an underground

    I thought an underground Stonehenge maze but finally choose "the Tomb of King Arthur" based on the legend of Glastonbury Abbey who is in ruin. I thought it would be inspiring to think about how the Arthurian myth mix with LOTFP weirdness while building the dungeon, without any plans.

    I've been steeped in Arthuriana all year, starting in January with Malory, and I have to say, this has me hot to trot. From the perspective of the myth, actually finding the tomb of Arthur would be extremely important and exciting and disappointing and all kinds of crazy from any number of interested parties. Just a really great idea for an adventuring location. It can be simple or very socially complex, depending on the details of the setting.


    I ended up with a six legs/four arms skin-translucid wolf-centaur (instead of horse) with a wolf's head who was capturing things in its lair to impregnate them and that could spit biting leeches (you could see them living in a nest in its translucid organ). I called that thing "The Lady of the Lake".

    You are a bad person, and I like you.


    I opted for more than one entrance.

    Alternative entrances to dungeons are of course a hallmark of the genre, but sometimes I wonder how often they are discovered. In my play, in these types of situations, unless there are strong rumors hinting about where other entrances are and why one might take them, the players are keen to just get on with the meat of the game and go with the obvious entrance. Was the alternate entrance telegraphed at all? Did they have reason to seek it/a reasonable chance to stumble upon it? Or is it more there as an alternate exit?

  4. To be in it, one must be in it

    Having leaned back into Barbaric Psychedelic for the upcoming convention, I've thought about the lessons learned about this particular species of scenario preparation since I started playing 4E about ten years ago.

    1. A little is enough. Huge dungeons or dungeon-style design may be great creative efforts, and they certainly tap into one's private enjoyment of a larger concept for setting, but they are typically not accessed in play.
    2. Actually play in the thing, not near it. In narrowing circles, "How we met," "How we get there" (possible subroutine 'why'), and "What do we do here on the periphery," are all paralyzing.
    3. In combination, #1 and #2 are practically a recipe for not seeing or being in most of what's been prepared, and arguably not playing much at all.

    Looking at these, I see a couple of points to clarify. We're not talking about an inch by inch examination of every aspect of my beautiful prep, and clearly seeing one area or thing without seeing another is a feature, not a bug. That's why I used the word "accessed" – if I prepared 80 rooms and the group encounters two of them, it's fine, if that two-room experience really nails the aesthetic power and dangers at hand, i.e., "why we play."

    So, two points arise from there.

    First, if those two rooms really nail it that well, then why bother preparing 80 of them? Given that two of them provided us with a substantial period of play (one session, three, ten, whatever), then enough for context and to make choosing/finding those two into an event rather than a program should be plenty. One's notion of enough is certainly individual, but going well over the scale of something like The Cursed Chateau, or closer to home for me, an adventure location in Trollbabe, should rightly be considered excessive and unnecessary for any imaginable purpose of playing right now with these people. (The in-fiction scale isn't the issue; if play includes moving smartly over large distances, then my point still stands for continent-spanning prep, because I'm talking about how much has been prepared in terms of activity, and how interrelated it may be.)

    Perhaps it is rude to point out that the recent Blueholme game which ended with Bartle the skinny and rather useless thief wonderfully surviving, only poked at one or two spots well above and far removed from most of the elaborate map and content provided in the Lost City preparation. None of the underground preparation mattered a damn … nor imaginably could in terms of continuing to play, unless you found a way to nab the player-characters and put them down there.

    Looking at your preparation, Greg, I'm thinking about my original play-experiences with Barbaric Psychedelic, which used preparations far bigger than the ones I used for the game you're reading about. The players were bewildered and exhausted by it, thrown off by their contact with this or that piece rather than energized to drive in. Even if they had, the resource mechanics of the game simply would not permit that kind of extended foray; they'd be killed within another layer or series if they tried.

    Therefore I went much smaller for the later game, with dramatically different results. A preparation like the Kzekk's head setup is huge for purposes of play, although by the standards of any published dungeon, it is tiny. And the dungeon-y portion of the second adventure is a single Dyson geomorph, with one room.

    Second, I feel your pain regarding the players' behavior, as well as enjoying a little poetic justice, considering I felt it directly from you in our Lamentations game. But let's not say, "bad, stupid players" – which you did not, Greg, but I'm sure you can see that is nearly the guaranteed interpretation of what you did say. Let's say "given all this prep and my enthusiasm for you to share and enjoy it with me, why are you treating it like a root canal?"

    Jon and I talked about this in Conversation: D&D play culture, which I hope you'll review in order to pursue the topic here. For now, the point concerns player contact with the important part of what's been prepared, not merely hints or side-effects of it. We'll have to develop this in conversation, I think – what I'm seeing here and have observed so many times is this: as soon as they can, the players go into escape mode. "The only way to win is not to play," essentially. It's as if their characters went up to the edge of things because they had to (being fictional), but now that they have any autonomy in that fiction, first they delay and plan and flinch in order to avoid touching anything or encountering anything, then when they encounter anything despite their best efforts and take a hit or two, they cower, scamper, and flee.

    But why? I don't ask in terms of player motivation or presumptions about play or delusions about what "dungeon play" is. I ask in terms of what is placed before them in immediate, in-fiction terms, for which, apparently, this visceral response is the only one possible. Let's presume the players are not being entirely unreasonable about this, so that, if they are being unreasonable (and frankly despicably wussy) in terms of what you prepared, well, they can't see what you prepared, and they are not unreasonable in terms of how they see and understand what has just been described to them.

    Something about that description, or about many interrelated aspects of starting to play which culminates in that description, is going wrong. From GM-side, not player-side, what could that be?

    • Perhaps it is rude to point

      Perhaps it is rude to point out that the recent Blueholme game which ended with Bartle the skinny and rather useless thief wonderfully surviving, only poked at one or two spots well above and far removed from most of the elaborate map and content provided in the Lost City preparation. None of the underground preparation mattered a damn … nor imaginably could in terms of continuing to play, unless you found a way to nab the player-characters and put them down there.

      This is a fair point. Using a prepared dungeon is slightly different than making your own, but I do not think it matters in the context we are speaking here. Except that The Lost City is a product that is supposed to supply hours of play before you buy the next product. It comes with the built in idea that you are going to gain enough gold to convert into experience and level up. It has far more content than it needs to do this and largely does this to meet outside expectations.

      But the characters had found enough treasure that even if divided in the most fair and strict manner, they would have leveled up without issue. And this was done with bypassing huge spaces of the dungeon. If I had made a straight-line dungeon with all of the same content in one room-corridor-another room combination, the result would have been the same.  


    • A few month ago, I wanted to

      A few month ago, I wanted to answer to this comment. My answer was something like "I don't understand, could you try to rephrase". But Noah's comment made me understand the things and I tend to agree.

      Should I try again this type of game in a dungeon, I would go basically with the same assumptions:

      • Start direct in the "dungeon" (wether it's a region of wilderness or a real dungeon)
      • just a few interesting rooms, 3, maybe 4. When I designed this dungeon, I was wondering if I need empty rooms. I finally put empty rooms with "color" elements from a specific table, assuming that it could be used during play in some way I don't imagin (running from a creature, for instance, I don't know). But definitely, "just give me the good rooms" seems to make better sense.

      In other words maybe: keeping the dungeon + the fantasy, but excluding the assumptions behing the crawl. 

    • In reviewing this discussion,

      In reviewing this discussion, I think I over-stated the "two rooms" concept. A few empties seem necessary to me too. The notion of the dungeon, ruins, tower, whatever-it-is as a place matters a lot, and if it's just a few cubbyholes with a "great encounter" apiece, it's not a place.

    • I’m surprised by and interested in your question:
      “First, if those two rooms really nail it that well, then why bother preparing 80 of them?”

      I have three answers from my experience.

      First is that two rooms often isn’t enough. Play might involve tactical situations throughout a dozen rooms in one setting in one session, particularly if the PCs are much more powerful than the inhabitants of the rooms. (I know, you said you were being hyperbolic about this, but it deserves a legitimate answer.) Part of the challenge is preserving limited resources across many encounters.

      Second, I revisit most locations again and again over the course of many sessions of play, and mastering the location, which is too big to be readily understood at first glance, is also part of the challenge.

      Third, the 80 rooms which the players don’t visit give context for and help me prep and play the rooms they do visit. They let me put in a whole living situation for the gang of orcs, or new enemies for the goblins, or a place for the bandits to retreat to, or reinforcements for the demons. Etc. And the players might visit them; I don’t know in advance.

  5. Prepping a Dungeon for 4e

    Here is a video discussing the prep that went into my first dungeon for D&D 4e. This post and the discussion in the comments were really helpful while prepping.

    Also, HUGE apologies to both Greg and Sean for misremembering the original author of the post — I only caught it after the video was done processing, but I called out the misattribution in the vid notes.

    • All good. Greg’s post was

      All good. Greg's post was fantastic and I am glad he got a shout out.

      In the fiction, the next adventure just happens. This is part of the literary tradition and to some degree part of the "dungeon" tradition. I say that, not to direct you there but acknowledging that it might not be satisfying or desirable to do it. But you also do not have to be afraid to do so.

      You can think of the wilderness area as a dungeon unto itself. Especially making use of the Skill Challenges. 

Leave a Reply