Megadungeon Design: Philosophical Extracts

I recently re-read Herman Melville's Moby-Dick, a 552-page American epic that opens with 11 pages worth of "Extracts": a bunch of quotes from a wide variety of texts that refer to whales and whaling practices. Though Melville presents these extracts as if they were provided by a "Sub-Sub Librarian" for the benefit of his narrator, Ishmael, I have always suspected that these quotes represent stuff that Melville himself tracked down as he was researching whaling for the novel. [Note: Melville had wide experience at sea but never went on a whaling voyage.]

So in that same spirit, and since I have been seeking out this material for my own benefit anyway, I wish to commence my 2011 Megadungeon Project by compiling and presenting some of my predecessors' thoughts on the topic of Megadungeon Design.

Megadungeons are BIG, and so too is this post -- please forgive.

+ From Swords and Wizardry Core Rules by Matthew J. Finch, p. 57:

The traditional dungeon, which many people on the internet call a “megadungeon” is a vast labyrinth of underground tunnels, rooms, corridors, and chambers, extending many levels down beneath the surface. It might contain subterranean lakes, rivers, chasms, and cave-ins, and it is certain to contain traps for the unwary, monsters in their multitudes, and most importantly: treasure beyond the wildest dreams of the players — if they can bring it out safely.
In designing a megadungeon, it’s often helpful to start with a quick cross-sectional map of what the dungeon looks like. But the real meat of designing the dungeon comes when you start making the floor plans of the dungeon levels themselves.

+ From Labyrinth Lord Revised by Daniel Proctor, "Labyrinth Scenarios" p. 123:

[T]he Labyrinth Lord will have to develop multiple scenarios for more extensive labyrinths that the characters keep coming back to. The characters should learn more rumors or legends about the deeper levels of the labyrinth as they progress in levels, or find clues throughout the labyrinth about other regions of the labyrinth. In addition, a multi-level labyrinth used for extensive play should be considered a "living" place. The Labyrinth Lord must keep track of how the player characters alter the environment, and how resident monsters may change in number, type, or behavior in response. A megalabyrinth will evolve through time just as the characters will by adventuring there.

+ From Philotomy's "The Dungeon as a Mythic Underworld":

There are many interpretations of "the dungeon" in D&D. OD&D, in particular, lends itself to a certain type of dungeon that is often called a "megadungeon" and that I usually refer to as "the underworld." There is a school of thought on dungeons that says they should have been built with a distinct purpose, should "make sense" as far as the inhabitants and their ecology, and shouldn't necessarily be the centerpiece of the game (after all, the Mines of Moria were just a place to get through). None of that need be true for a megadungeon underworld. There might be a reason the dungeon exists, but there might not; it might simply be. It certainly can, and perhaps should, be the centerpiece of the game. As for ecology, a megadungeon should have a certain amount of verisimilitude and internal consistency, but it is an underworld: a place where the normal laws of reality may not apply, and may be bent, warped, or broken. Not merely an underground site or a lair, not sane, the underworld gnaws on the physical world like some chaotic cancer. It is inimical to men; the dungeon, itself, opposes and obstructs the adventurers brave enough to explore it.

+ From James Maliszewski, "Megadungeon Background":

One of the things I think is very important for the creation of a good megadungeon is a solid background for its existence. By that I most certainly don't mean pages upon pages of details about the dungeon's construction and all the little events that took place in and around it. That strikes me as largely useless. I do, however, find utility in a "framing device" that gives some context to the whole thing -- something to provide a rough explanation for what's going on in the dungeon, even if it does obey its ecological and physical laws quite distinct from those of the world outside.

+ From Joseph Bloch, "Old School Campaign Tent Poles":

[A]n old school campaign should properly have a "megadungeon" as its tent-pole adventuring locale. [. . .] [B]ut it's important to remember [. . .] that just because a campaign has a central megadungeon as its focus, it is by no means the exclusive province of the campaign. Adventurers are free to delve into its mouldy depths if they choose, but they're just as free to poke around in the local town or city, hie off into the wilderness, or explore smaller dungeons that the DM should properly have waiting for those eager for a change of pace.
But what I think is important is that the tent-pole dungeon is there, for those days when the PCs weary of tracking down clues and following whatever plots they have been pursuing throughout the campaign. The tent-pole should, I think, be a sort of refuge from such things, existing in the world but not of it, a place where the PCs can set aside their plot-induced cares and just go someplace, kill things, and take their stuff. In my own campaign, the dungeon of Quasqueton served that purpose; when they needed to take a break from witches, brigands, and conspiracies the PCs would delve into its depths. It helped to bolster them in terms of x.p. and magic items, too, getting them to a place where they could more equally interact with some of the bad guys.

+ From Al Krombach, "Megadungeon Design and Philosophy":

Think Big
Think about the number of rooms you want to detail on a given level (say, 42), and then multiply that by two, or even three. That's how many rooms you really have on that level. I'm not saying you should detail three times as many rooms, but that Megadungeons need empty space. Its important to the whole concept. The Megadungeon shouldn't feel like a series of consecutive rooms, each filled with its own monster, treasure, and trap to overcome. [. . .]
All those empty rooms help convey that feeling of BIG, and they also give you room to flex your improvisational chops a bit. Fun DMing is not, imo, just reading off a bunch of predetermined encounters, but getting to interact with the world just like the players get to. Keeping a bunch of unfilled rooms lets you dig out the randomization charts, add stuff to further the plots that will develop in play via the characters' actions, and let you provide fresh areas to explore for new 1st level parties, whether that be from DMins multiple groups, or because of the occasional TPK!

[. . .]

Connectivity
One common mistake, easily made, is to place too much emphasis on the levels, and not so much on the connections between them. Without vivid, detailed, and well-thought connections between the levels, all you really have is a collection of ten different dungeons, linked weakly together by the odd stairwell.
Its important to get a three dimensional image of your dungeon. Think about where the levels connect, and design the levels around those connections. Think about how massive places in real life are organized, be it your local shopping mall, theme park, business centers, urban districts. Notice how things are grouped around and organized according to connectivity. High traffic areas are important, to vendors, police, even criminals. As you move away from those connections, things get smaller, seedier, more spread out, darker, etc.
You can apply those same grouping tendencies to the connections in your Megadungeons. Areas near the main connections between levels may often feature larger chambers, great halls, etc, with more traffic, special guardians, and depending on the amount of adventuring traffic, may not have as much in the way of treasure. Similarly, as you move away from these connection points, things will get more spread out, traffic will decrease, and many chambers may not have been disturbed for a long time, which means the possibility of greater rewards.

+ From Al Krombach, "Mapping That Megadungeon":

The best way I've found to manage that heavy mapping load is simply to change your perspective on what exactly needs to be mapped. Most DMs I've spoken with running Megadungeons will tell you that they're running it about 33% pre-planned, 33% random charts, and the remaining 34% "on-the-fly". [. . .]
Map your Megadungeon the same way you would your campaign setting. Start with a familiar "home town" area they're going to frequent. In most cases, this is the main entrance to the dungeon and the main connection points. Sketch out a rough idea of the scope of the [. . .] Megadungeon. Then add detail as you need it, and as the characters journey. Map possible areas of further exploration a little at a time, in between sessions, so that mapping goes back to being a fun, relaxing pastime, rather than a chore.

[See Al's whole series on Megadungeon Design -- absolutely Top Notch!]

+ From James Maliszewski, "Empty Rooms":

One of the things I have done religiously while creating my Dwimmermount megadungeon is abide by the rules presented in Volume 3 of OD&D for the distribution of monsters and treasure. There are several reasons I've done this, but chief among them is that these rules ensure that about two-thirds of the rooms on each level contain no monsters. That means, for example, that of the nearly 70 rooms on the first level of Dwimmermount, only 23 of them will have occupants.
That leaves 47 "empty" rooms for the characters to explore. Of course, a lot of these other rooms aren't really empty at all, since they might contain tricks, traps, clues, unguarded treasure -- 1 out of every 6 unoccupied rooms has it according to the rules -- and just plain inexplicable things. My experience over the last 18 months of running a megadungeon-centric campaign has been that it's often the "empty" rooms that are the most memorable, as it's here that the players, through their characters, interact most immediately with the game world. Furthermore, empty rooms help build tension and mystery, both of which are vital to the long-term success of a campaign.
Dungeons & Dragons is, after all, a game of exploration. Dungeon delving is a quest for loot and knowledge and, far from being the focus of the game, combat with the inhabitants of the dungeon is but one possible obstacle standing in the way of the characters' goals. That's why it's important that dungeons, especially megadungeons, have lots of empty rooms. It's a practice I fell out of over the years and whose importance I only understood fully as I immersed myself in OD&D. Now, it's hard to imagine stocking a dungeon that isn't mostly "empty."

+ From Philotomy's "Creating a 'Mythic Underworld' Dungeon":

[S]tart off with the first three levels, and start running it. You can certainly have a framework or general idea of what you'll be placing in the deeper areas, but you don't need to finish (or even map) those areas, yet. You'll develop the deeper levels (as well as continuing to develop and modify the upper levels) as the game continues.
This is a very cool, and very "old school" approach. Your dungeon will evolve in a very organic manner. During play, the players are going to ask questions and take actions that make you think and give you ideas that never occurred to you. Actual play is going to shape the direction and design of your dungeon, often in unexpected fashion. You and the players will be in a sort of creativity feedback loop, and your dungeon will be all the better for it.
When creating your first three (or so) levels, there are a few general concepts that you should keep in mind. First, remember to offer the players plenty of choices. Even at the entrance to the place, don't give them one path to follow, give them four or five choices to make, right off the bat. For that matter, there needn't be only a single entrance. Have several ways in, with a few of the entrances going directly to deeper areas. Maybe new entrances open up or are discovered as play continues. Another important way to give players choices is to offer them many opportunities to move up and down through the levels. You want the players to decide when they want to go deeper. This isn't a video game where you play through the level to the end with the boss monster, then find the stairs. If they're a group of 1st level PCs, but they want to try their luck and skill on the 4th level of the dungeon, that's their decision.
Also, remember that stairs needn't go up or down a single level, and that's it. Give the players ways to go down multiple levels. Some paths up or down may skip one or more levels. You may be leery of including a stair, shaft, or elevator that spans multiple levels, fearing that your players will go down into undeveloped areas of the dungeon. That's true; they might. However, it's more likely that they will be fearful of going too deep, and even if they do descend to a level you haven't developed, they'll be very jumpy and very likely to stick close to their line of retreat.

[. . .]

[T]here is something to be said for eschewing graph paper, entirely, and drawing your maps on plain white paper. This frees you from the constraints of the grid, and you might be surprised to find that your mapping takes on an entirely fresh character, with levels stretching out or sprawling in a much more organic and natural manner. Varying your approach from level to level is another good technique for keeping things fresh. One level might be very maze and grid-like, with relatively thin walls and not much rock, stone or earth between areas. Another might use large chambers, widely spaced, with curving tunnels through thick areas of stone. Trying different approaches to the act of mapping will naturally result in different styles of map, in many cases.

+ From OD&D Book 3, The Underworld and Wilderness Adventures by Gary Gygax and Dave Arneson, p. 4:

In beginning a dungeon it is advisable to construct at least three levels at once, noting where stairs, trap doors (and chimneys) and slanting passages come out on lower levels, as well as the mouths of chutes and teleportation terminals. In doing the lowest level of such a set it is also necessary to leave space for the various methods of egress to still lower levels. A good dungeon will have no less than a dozen levels down, with offshoot levels in addition, and new levels under construction so that players will never grow tired of it. There is no real limit to the number of levels, nor is there any restriction on their size (other than the size of graph paper available). "Greyhawk Castle", for example, has over a dozen levels in succession downwards, more than that number branching from these, and not less than two new levels under construction at any given time. These levels contain such things as a museum from another age, an underground lake, a series of caverns filled with giant fungi, a bowling alley for 20' high Giants, an arena of evil, crypts, and so on.

+ From AD&D Dungeon Master's Guide by Gary Gygax, "Monster Populations and Placement" pp. 90-91:

As the creator of a milieu, you will have to spend a considerable amount of time developing the population and distribution of monsters - in dungeon and wilderness and in urban areas as well. It is highly recommended that you develop an overall scheme for both population and habitation. This is not to say that a random mixture of monsters cannot be used, simply selecting whatever creatures are at hand from the tables of monsters shown by level of their relative challenge. The latter method does provide a rather fun type of campaign with a ”Disneyland” atmosphere, but long range play becomes difficult, for the whole lacks rhyme and reason, so it becomes difficult for the DM to extrapolate new scenarios from it, let alone build upon it. Therefore, it is better to use the random population technique only in certain areas, and even then to do so with reason. [. . .]
Natural movement of monsters will be slow, so there will be no immediate migration to any depopulated area - unless some power is restocking it or there is an excess population nearby which is able to take advantage of the newly available habitat. [. . .]
Alter creatures freely [. . .] after players become used to the standard types a few ringers will make them a bit less sure of things. Devising a few creatures unique to your world is also recommended.

+ From AD&D Dungeon Master's Guide by Gary Gygax, "Placement of Monetary Treasure" pp. 91-92:

All monsters would not and should not possess treasure! The TREASURE TYPES given in the MONSTER MANUAL are the optimums and are meant to consider the maximum number of creatures guarding them. Many of the monsters shown as possessing some form of wealth are quite unlikely to have any at all. This is not a contradiction in the rules, but an admonition to the DM not to give away too much! Any treasure possessed by weak, low-level monsters will be trifling compared to what numbers of stronger monsters might guard.
Assign each monster treasure, or lack thereof, with reason. [. . .] [I]ntelligent monsters, or those which have an affinity for bright, shiny objects will consciously gather and hoard treasures. Others will possibly have some as an incidental remainder of their natural hunting or self-defense or aggressive behavior or whatever. Naturally, some monsters will be so unfortunate as to have nothing of value at all, despite their desire to the contrary - but these creatures might know of other monsters (whom they hate and envy) who do have wealth!

+ From AD&D Dungeon Master's Guide by Gary Gygax, "Placement of Magic Items" p. 93:

Initial placement of magic items in dungeon and wilderness is a crucial beginning for the campaign. In all such places you must NEVER allow random determination to dictate the inclusion of ANY meaningful magic items. Where beginning/low-level player characters are concerned, this stricture also applies to the placement of any item of magic. Furthermore, you need never feel constrained to place or even allow any item in your campaign just because it is listed in the tables. Certainly, you should never allow a multiplicity, or possibly even duplication, of the more powerful items.

+ From Michael Curtis, "Die Roll Dungeons," where he discusses the the Stock the Dungeon table from Moldvay Basic:

I’m a proponent of the randomly stocked dungeon. I follow Gary’s advice to determine where the important monsters and treasures are found, then dice to see what else is down there. For me, this is half of fun of designing the dungeon in the first place. Random determination of contents often gives me results I hadn’t anticipated, challenging me to come up with offbeat solutions or interesting backstory to justify the presence of an object. And, as I’ve stated in the past, the ability to creatively interpret the unforeseen results of a die roll is one of the overlooked skills of a great referee, so it pays to develop and constantly hone this ability.
Still more interesting for me as a referee is that, by using a random method to stock the dungeon, the dungeon itself sometimes informs me of what it wants to become rather than what I had in store for it. My most recent dungeon has done exactly this. I started with the intent to make it a delve suitable for 1st level PCs, but, as a result of several dice rolls, the challenges awaiting the players have made it more suitable for 2nd and even 3rd level characters. And I think it’s a much richer adventuring site because of this. A few of the specials found within it have also added to the history of the campaign world and the background events of the sandbox setting. None of this would have come about if I had gone about deliberately inserting each element into the dungeon and worried about maintain a certain level of difficulty.

+ From Jeff Rients' "Notable Features of the Dungeon of Doom":

Much of the dungeon was generated using the random charts in the back of the first edition DMG, with one slight variation. When a monster was called for I would roll a d6. On a 1 or 2 I would then use the Wandering Monster charts from the DMG to determine which critter to use. On a 3 or 4 I would go with the Fiend Folio charts and on a 5 or 6 I'd use the Monster Manual II tables. The result was that the monster mix was skewed heavily towards the original Monster Manual, but with a good smattering of weirdoes from the other two books. Also, I used whatever creature I rolled, no matter how clumsy the mechanics or how little I like the monster. If the dice said the room had 2 vulchlings, then by Cuthbert there were two vulchlings in that room.

+ From "Stocking the Dungeon" by Sham aka Dave:

Before I begin drawing the map for a particular level, I normally start out with a handful of ‘Major Room’ ideas. These may or may not be fully defined at this point, but the idea is in place. The level itself may or may not have a theme, these Major Rooms will fall into such a theme if one is present. Normally, I draw these Major Rooms on the map first, then I fill in the rest of the map with my normal halls, doors, pits and other rooms. Then, I number each room or space and proceed to keying the rooms. I rarely have any actual ‘empty’ rooms, at the very least a room will have some features, even if it is something simple like graffiti, murals or rubbish and debris. These are still considered uninhabited rooms at this point. Certain Major Rooms will include recurring encounters, and will essentially have the same monsters no matter how many times the dungeon is visited. Such examples I currently use are Goblin Spawning Pits, Giant Centipede Nests, Rat Infested Areas, Insect Swarm Mires, Spirit Haunts, Sacrificial Pools, etc. The monsters spawn at these locations, and it is nearly impossible to destroy the source of their seemingly endless numbers. Other than these specific, recurring monsters, everything is, at this point, considered uninhabited.

Step 1, Manual Placement of Treasure: Easy enough to understand. I determine areas where I know I want a treasure trove, and I proceed from there. Nothing random at all, just picking and choosing.

Step 2, Check For Monsters: Before I dice for all of the uninhabited rooms, those with no recurring monster source, I will decide if an individual room will even have a chance of a monster inhabitant. Some rooms will be left as is, depending upon their description, these simply do not have a chance of monster inhabitants. The remaining rooms have a 2in6 chance of a random monster inhabitant.

Step 3, Random Monsters: Rather than using the actual table in Vol. III, I instead refer to my customized Wandering Monster Table I write up for each level. These normally have ten monsters types, and a small chance of a special or NPC type encounter. So, 2in6 chance of one of these encounters in each uninhabited room. If diced for, I determine randomly the number of such monsters, again from the Wandering Monsters Table. One key difference between actual Wandering Monsters and these room inhabitant types of monsters is that Wandering Monsters carry a random amount of coins detailed in that table, while room inhabitant type monsters will be carrying or guarding both those random coins, and possibly an additional treasure as determined in Steps 4-5.

Step 4, Check For Treasure: Now dice for treasure in those same rooms. 3in6 if a monster is present, and 1in6 if none is present. As detailed in the OD&D guide, unguarded treasure should be hidden or protected in some manner, either locked away, behind a trap, or both.

Step 5, Random Treasure: The original guide does not give a method for determining whether results in the Gems/Jewelry column are Gems or Jewelry, so I will use a simple system for this random determination, as follows: On levels 1-7, or when using a d6 to determine the number of Gems, a roll of 6 represents one piece of Jewelry and five Gems, lower rolls represent Gems only. On levels 8+, or when using a d12, a roll of 11 represents one piece of Jewelry and 10 Gems, while a roll of 12 represents 2 pieces of Jewelry and 10 Gems, lower rolls represent Gems only.

To determine the quantity of Magic in a particular treasure trove, I will use a similar percentile approach, adding the dungeon level to the roll, and referencing a custom table. Lastly, once the quantity is determined, the individual Magic item types will be determined by a similar roll, using percentile dice and adding the dungeon level.

[See also "OD&D Dungeon Stocking" on Sham's blog.]

+ David from Tower of the Archmage discusses his megadungeon:

There are certain things I'm trying to do with my megadungeon, and one of those things is to keep it old school. Graph paper and pencils to map it, with interesting shapes, 10 foot per square scale, and increasing deadliness as you go down deeper. I'm keeping in mind the weirdness of dungeons, and I'm not going to shy away from magic pools, talking statues, and other assorted goodies. I'm also keeping in mind that this isn't some basement, or castle prison... this is a Mythic Underworld! Not everything will make sense, and in that vein I want to randomly roll what is in it, and try to stick to the rolls I get.
That was the theory anyway. In practice that turned out to be a lot more difficult right from the first level. While I understand that things aren't necessarily balanced, and in fact can be rather highly unbalanced, I ran into a problem that basically forced me to make major adjustments to the results of my rolls.
It's pretty obvious that a series of somewhat improbable rolls could make a level unplayable without the DM stepping in and tweaking things. As it is things are still going to be pretty interesting with some really tough potential encounters [. . .]

+ From Michael Curtis' five guidelines for Megadungeon Design:

#1: “Stop worrying and love the dungeon”
#2: “Balance Realism and Fun, but when in doubt, Fun always trumps Realism”
#3: “The Fantastic, when cranked up to eleven, somehow equals the Realistic”
#4: “Never be afraid to say ‘no’ to the dice, but also never be afraid to say ‘yes’”
#5: “Plant many seeds, but only tend the ones that grow”

+ From Michael Curtis' addition to his initial five Megadungeon Design guidelines:

I have a personal design philosophy when it comes to building dungeons. Inspired by the wild and woolly days of this hobby's beginnings, I've learned to let the need for realism and rationality to go on holiday if I think it's going to stand between me and a memorable dungeon experience. I summed up this philosophy some time back with the phrase, "Stop worrying and love the dungeon." That mantra still influences my efforts. [. . .]
I'm adding a sub-clause to the Society of Torch, Pole and Rope design philosophy of "Stop worrying and love the dungeon," which states, "Go crazy, just 'cause you can."