Friday, August 26, 2011

Open Friday: Megadungeon Formats

To date, not a lot of true megadungeons -- as opposed to just really big dungeons -- have been published. I'm starting to wonder if the reason for that is that it's difficult to present a megadungeon in a format that's both comprehensive enough to do justice to such a campaign tent pole and yet easy to use in play. Then there's the additional factor of how much detail a megadungeon product needs to be usable and whether too much detail undercuts the very essence of a megadungeon.

So, today's question is this: assuming you're interested in megadungeons -- if you're not, please don't use the comments section to express your disinterest -- what would be your preferred format? Feel free and assume that there are no limits and go with what you would consider to be the ideal format for presenting a true old school megadungeon.

I'll admit I'm very curious to hear what people have to say, since it's a topic I've thought about myself at some length and that I'll share in an upcoming post.


  1. Sandbox style. A dungeon big enough that it's not really a dungeon, but like a fallen city with different venues for exploration and retreating to different strongpoints. While Undermountain was a bit of a megadungeon for its own time, it had a lot of ties to Waterdeep and Skullport.

  2. I think the megadungeon should be at its heart, a sandbox of sorts myself. A large scale place that may take years of actual play time to fully explore with lots of different features and various factions, etc. I am currently working on a megadungeon project called Dreadrock, which I've been posting about on my own blog. The first quadrant of level one is packaged in a nice PDF one page dungeon format, and the remaining sections will be done in a similar fashion.

    My thought was to give minimal descriptions to each area, but enough variety to allow DMs to go in various directions as far as a storyline is concerned. The megadungeon truly is a campaign setting in of itself, and building one is a fairly daunting task, which is why we've seen so few of them over the years. I've been enjoying working on Dreadrock, however, and in the playtests I've ran, it's gone very smoothly so far and the players really seem to enjoy it.

  3. A megadungeon is almost a campaign setting. It should be like a sandbox, but modular and open enough to withstand DM tampering. Though I'm loath to admit it, the dungeon in "Diablo" is close, though smaller than I'm thinking. It's not so much a static dungeon but a series of elements that can be incorporated into a dungeon.

    A megadungeon needs to be compelling enough to draw the characters back for years of adventuring.

  4. Modular.

    And by that I don't mean that the whole megadungeon is one big module -- but that the megadungeon itself is composed of many small modules.

    Each module is highly detailed, but describes a relatively small area -- no more than either a single whole level or small parts of a few levels.

    And every module is designed such that it can connect to any other module.

    In other words -- each module is a sort of geomorph writ large, complete with full descriptions, keyed encounters and a random encounter table.

    So the megadungeon ends up being whichever modules get put together in whatever arrangement they get put together.

    And it can keep growing and growing by adding more modules wherever the PCs go.

  5. I have players who love em and those who are opposed to them. As a DM I like megadungeons b/c it doesn't require much set up and you "feel" like you get more accomplished in a night's game session even though it was searching a bunch of empty rooms. Go fig.

  6. Like a campaign setting but more contained. Maybe more of a sub-setting type of thing so it could be put in an already existing world.

    Also something bare bones would be preferred. Detailing every room down to the deleterious would be just tell me that it was a product meant to be big so it could sell for a lot. A few rooms should be detailed like the main entryway and some set piece type areas but otherwise besides maybe a quick one to five word description like "dirty" or "signs of recent fights" to provide hooks for me to hang stuff on I only need to know the room is there.

    Basically it would end up being a product mostly filled with interconnected maps and the occasional optional set piece. The only thing I would say for the actual structure of the dungeon is that the first level should be mega in and of itself with lot of place marked as possible entrances though this is just be my opinion on megadungeon construction in general.

  7. It should be a setting i.e. differents area and different levels would have their reason for being with interesting interactions between them. A few areas in details and the rest in a minimal dungeon format.

  8. This might sound a bit schizophrenic, but I'd say that Stonehell Dungeon does a pretty good job of getting the formula right, even though SD isn't really a megadungeon I'd enjoy playing in. I think they did a good job presenting the locale in a useful and usable way regardless of whether what they put in it was my cup of tea.

  9. Its a hard call. While Sandbox has its merits, there is a degree of non ramndomness that you expect from a fixture in a larger setting.

    You might craft a huge scale hex map covering the 56 miles of the Kue Gobelin tunnel and detail key sites that are nodes on that dungeon (the Mines of Radlebb, the Bugbear Caves (a southern exit), the gobelin caves (a northern exit), and the Ruins of Korizegy.

    There will always be some obscure little cave network you can add along the tunnel giving priority to what is most likely in this region - Bugbears at the southern exit, Goblins at the Northern Exit, and a Possible Lich Lair somewhere in the middle...

    The Cleaves (a megadungeon accessible only by failed dimension door spell from anywhere in the setting at any time) is made up of room tiles that are selected from a card deck randomly. The limitless Dungeon is randomly drawn - the reason the Adventureres are looking to create stability in it so they can explore and locate a solution to their escape. Effectivly the deck can be expanded perpetually and the dungeon has no limit - every 'room' a node and the concept of hallways non-existant.

    So both have their merits...

  10. On a level design, I always though those Arduin Dungeons had a megadungeon feel to them. by the way Hargrave would draw lots of halls and room onto a single sheet of 8' x 10' graph paper. Even on a small scale, he was able to achieve this effect as he did with" Lancer's Rest" , which was made for his friend Lance Mazmanian back in '87.

  11. Admittedly, this might be biased because I'm writing my own megadungeon for use in the web-game I run, but my ideal megadungeon has to have a mixture of sheer, random dungeon dressing, combined with genuine feelings of living and community of the various races populating it. That can build up into rivalries and alliances between the settlements, which then leads to 'deadzones'.

    As the various zones are constructed and populated, the different groups and dungeon areas build up organically and by their nature allow the players' actions to have a direct effect on populations and attitudes.

    The megadungeon, as has already been commented, has to feel as much alive as a city can. On top of that, I like to think that any 'true' megadungeon, to function as a tentpole, has to have nearby settlements or, if you want to be fancy, settlements within it to allow the resting and replenishing of PCs and thus allow the megadungeon to truly feel all-encompassing for as long as the players want to plumb it's depths.

  12. Personally, I'd like to see a published megadungeon with, say, the first three levels fully fleshed out, so it can be played straight out of the box; after that, though, if bit by bit less and less detail was provided for every subsequent level, with tips on how to provide that detail included in the product, that would be awesome. Sort of a megadungeon product that teaches you how to make your own megadungeon.

  13. Not sure what exactly James means by "format" in this question, but I would argue for a fair amount of detail. There is nothing more disappointing than plunking down money for a product, only to find that it's up to you to provide the product. (A grievance I still nurse against module B1, by the by.) The original Ruins of Undermountain boxed set was a reasonable presentation of (the first three levels of) a mega-dungeon.

  14. I think I'd present it in two volumes. One would have the maps and key with mostly brief descriptions of rooms, but some key monster lairs, traps and specials, not to mention important treasures that might be the subject of rumors or maps. Each entry would have some white space at the bottom for DMs to write in their own notes or changes that happen.

    The second book would have suggested monsters for filling out each level, random encounter tables, and information on factions and how the different groups within might react to the actions of the PCs.

    The heavy lifting would be done, but it's up to each individual DM to flesh it out and make it their own, and then try to bring it to life in play.

    And yes, I think this would be pretty tough to pull off well.

  15. For media format, I would like to see an extensively hyperlinked html file that cross-references everything and make extensive use of tooltips. That way, you can keep room descriptions down to a minimum and allow the DM to expand the level of detail as needed for mechanical issues, saving mental space for narrative issues.

    Otherwise, something like Lord Gwydion above suggested would be close to ideal.

  16. A big thick book with the map on one page and the key -- with some detail but a fair bit of wiggle room -- on each facing page. I don't know how the published Stonehell did it, but the original version was in such a format and I thought it worked well.

    I'd also make use of modern graphic design techniques to aid navigation through the book. Colour-coded page edges to pull together related levels, easy-to-understand icons, stuff like that. Many rpg products -- both within and outside the OSR -- tend to be stuck in the 70's and there are surely better ways to present the information. Well, Zak's Vornheim shows there are, so that's kind of thinking I'm looking for.

  17. At the front: a couple gigantic maps that show the whole thing--with a few "landmarks" picked out for orientation.

    Then zoom in and have every "quadrant" done in pretty much the map-facing-comprehensive-but-efficient-key format that the dungeons in Vornheim have.

    I like the idea of "modular" or sandbox-inspired design in theory, but I feel like if you're gonna go ahead and publish one big megadungeon, you might as well have levels and zones that were designed as a piece and make some sense together. Otherwise why bother making a whole new product?

    Maybe the modular bits could be in the appendix.

    Also I feel like an efficient and fan-fiction-free summary of the different subsections and the way they "work" in play would be soooooooooooooo helpful. Too many big dungeons make you pretty much read every room before you get "the big idea" behind any given section.

  18. Here's what I'd like to see in a megadungeon project that I'd pay >$20-30 for (pdf version):

    1) High-quality maps

    I want to purchase maps that are attractive with interesting design features! That way, I can re-use them for other purposes if an author's vision for a level doesn't fit my own.

    They can be hand-drawn, professionally inked or created on a computer. What I dislike are maps that look like they've been created using a random dungeon generator.

    I'd also like a keyed and unkeyed version of the maps in the pdf.

    2) Images

    I want to see (and show the players) what the new creatures and cool features of the dungeon look like! There don't have to be 100s of pieces of artwork, but I want to give the players a feeling for what this environment is like. I'm willing to pay for this.

    3) Factions & motivations

    I want to have a clear sense of why NPCs/intelligent creatures are on a level, what they want and the politics of achieving their aims.

    Any MD can easily be turned into a hack n' slash if desired. It's much harder to go from a place where everything-must-be-killed/cleared to combat-is-a final-option-when-all-else-fails.

    4) Options

    Yes, I want to run it out of the box, but we all know it doesn't always work that way. So I would like options.

    That could take the form of suggested alternatives, random tables, etc.

    NOTE: I re-read Stonehell Dungeon yesterday. I really like how Michael managed to capture a lot of feeling in his dungeon in relatively few pages. There's a good sense of who, what and why. The dungeon dressing is really good. It feels easy to swap out creatures. I'm not a big fan of the maps and some of the levels don't work for me conceptually, but I feel like I could lift whole levels and place them in a new environment with better maps and have fun.

  19. My ideal megadungeon would be EXTREMELY huge, and presented in three volumes:

    -A book describing the setting, relevant places in each level and/or area, the history behind them, the denizens, the monsters, how the relations between the different communities of monsters work (if at all), a section with possible loot and treasure specific for the setting, and a little bestiary if needed, etc.

    It should be detailed, but not too detailed: giving just a general and brief description of each level should be enough, along with a detailed description of a single particular locale in each level, something unique.

    -A DVD/CD/download with the maps for all the levels, with index, links to quickly access any area and with all the relevant places "bookmarked", should be possible to zoom in/out, etc.

    -A third volume for the encounters. Detailed, "iconic" encounters for each level keyed to some rooms or areas, along with some suggestions on what should be encountered in each level. The last part of the book should be full with random encounter tables for the different levels (Table 1 - Levels 1 to 5, Table 2 - Levels 6 to 10, etc.), random loot, etc.

    Then put it all in a box, throw in some dice and graph paper, a poster sized map of one of the levels and call it a day XD

  20. "I like the idea of 'modular' or sandbox-inspired design in theory, but I feel like if you're gonna go ahead and publish one big megadungeon, you might as well have levels and zones that were designed as a piece and make some sense together. Otherwise why bother making a whole new product?"

    With modular design, you wouldn't publish one big megadungeon. You'd publish individual modules that can be assembled in any combination and any arrangement so people can easily build their own megadungeons from them.

    And each individual module would have levels and/or zones that were designed as a piece to make sense together. That's exactly the criteria that'd be used to decide what'd be an individual module.

    So the advantages of modular design are:

    1) Designers wouldn't have to complete whole megadungeons before publishing them. They could start with just single modules.

    2) Modules could be designed by any number of people, just like geomorphs are. So megadungeons could be built with modules from many different designers.

    3) Buyers wouldn't have to buy whole megadungeons at once. They could buy modules just as they need them.

    4) Buyers could pick & choose which modules they want to use in their own megadungeons.

    5) Opportunities & guidance for players to design their own modular sections are built right into the system. And, if they do it right, using their own modules in their megadungeons won't interfere with continuing to use commercial modules either. In fact, they could even sell their modules so other people could use them too.

    6) Just like with geomorphs, some people probably would get into designing & posting megadungeon modules online for others to use for free.

  21. If you are going to publish it in parts, then it is important that the first part contains a precis of the entire dungeon, including revealing any secret or ultimate goal of the dungeon, as well as general design philosophy.

    It should also include sufficient information of all other entrances to the dungeon to allow them to be reconnoitred by the players (with the implicit reason as to why they may be unsuitable for low level characters to use).

    Ideally the complete dungeon should be published as a single volume, which obliviates this requirement.

    One nice approach is the Diablo style approach, where certain core encounter areas are provided in good detail as set pieces, whereas the surrounding area is often much more loosely described. Navigation could even be in terms of these set pieces rather than an explicit map, allowing gamemasters to customise these between places.

  22. My take would be something in the lines of a campaign setting... o maybe halfway between a single module and a boxed set. Like a campaign setting, it'd be some areas very detailed and others more sketched (as when you have the "Burning Desert" overview and later you zoom to a street-by-street description of the City of Pilars, etc).
    Also, I'd like to see a living megadungeon, with an entire chapter devoted to history, factions and internal conflicts.
    Given the huge size of a megadungeon, It'd be interesting to have an "adventure book" with especific missions, quests and suggested adventures (complete with hooks, patrons and NPCs): just in case the DM and players do not want to explore the entire place, the quests will break up the megadungeon to easy bits, taking the PCs to a single or a couple levels per mission (also, reinforcing the "themes" of each part of the complex).

  23. Zak: "Also I feel like an efficient and fan-fiction-free summary of the different subsections and the way they "work" in play would be soooooooooooooo helpful. Too many big dungeons make you pretty much read every room before you get "the big idea" behind any given section."

    Yes. Yes.

    I also won't mind sparse footnotes or sidebars with a sort of "director's commentary" for the less obvious stuff ("This relates to the plot of the snake people in area 4a. If the PC's find the MacGuffin in this magic hiddy hole, the cultists will be unable to use it to seize..." or "Most groups I tested this with could not resist eating the cake, so be prepared for...")

  24. Level by level is the most straighforward to me. Include all the special keyed areas as well as your open TBD random areas.

    The structure must support ease of play for DMs.

  25. This is what I want in a product...

    Map hand outs for players
    Side-view maps
    Top-down maps
    Local maps with icons for interesting features.

    Wandering monster tables
    NPC tables for a particular area
    Rumor tables for a particular area
    Ambiance tables for a particular area

    Keep the descriptions to a minimum, a la the one-page dungeon format. For some areas allow the tables to provide all of the "description" except for extremely special locations. Keep background and history sections to a bare minimum.

    Examples of play:
    Sporadically through the product provide examples of play (e.g., GM: You see a fountain with a face spewing a glowing, purple liquid into a basin. Fighter: I run up and splash some of the liquid under my arm pits.)

  26. Doesn't do pretty much everything being asked for in these comments?

    Maybe they need to start up an OD&D version of it.

  27. James and all,

    Would you consider a large module, say SPI's Enchanted Woods or Chaosium's Griffin Mountain, a mega-dungeon, being a sandbox/self-contained area of scenarios/encounters/sub-settings?

    Am personally not fond of huge underground settings, but certainly surface mega-dungeons (like ruined cities, etc.) would be of interest.

  28. meg·a/ˈmegə/
    Adjective: Very large; huge:

    I'm sorry, but I get very confused on this: why isn't a 'really big dungeon' a 'megadungeon'?

  29. An idea I've mulled over in the past is that something more like a "dungeon workbook" might be a practical and interesting way to do such a product. So you get the maps, random encounter tables, brief descriptions of what the areas look like and of any fairly static features or challenges therein. Each keyed area comes written in a box with plenty of blank space, so that as the DM's running the dungeon s/he can make notes about what happened in each area: creatures killed, treasures looted, new monsters which have moved in, and so on.

    This would allow the dungeon to grow and change over time, while still allowing the DM to run it from that single book, rather than having to maintain and cross-reference a separate collection of notes of what's happened after twenty sessions of adventurers meddling.

  30. All of the above methods have their good points. Essentially, you could take any single book, from the AD&D DMG to Zak's Vornheim City Kit, and go crazy for days randomly mapping out and stocking a megadungeon. You wouldn't even need a Monster Manual, just breed critters as tough as you need them.

    My own preference is to have a really big area mapped out, with grey areas to be added as needed. Its one reason why I am working on detailing my own version of Greyhawk City, to be like an urban megadungeon. I want a general idea of what the whole looks like with enough canvas to paint as I wish. A city has nobles, clergy, and other groups and guilds aplenty that work with and against each other. An underground megadungeon needs that same sort of ecological flow.

    Take a few hints from Gygax. His Castle Greyhawk dungeons changed to represent adventures and happenings within. Levels and sub-levels grew where needed, with the occassional re-entry point added for quick access to a level when a party would replenish its supplies.

    Stonehell Dungeon and Joe Bloch's Castle of the Mad Archmage are both good examples to emulate.

  31. First, please keep room descriptions semi-short but not point-form short! I hate this format (ala Stonehell Dungeon), it basically requires me to read each room, look at the room dimensions on the map, and quickly try to think of something on my feet. It gets exhausting to me fast and takes longer than reading a line or two anyway. I never feel like I can relax and focus on "fun" things like throwing curveballs at the players or working on dynamic situations (ie. the orcs in room 12 are stalking the PCs and want revenge). Sure, keep them semi-short so I can embellish them but please, please, please, provide a sentence or two so that I can give the basics and leave it at that when I choose to.

    "You enter a 30 x 30 room with grey stone walls that glisten with moisture. An overpowering smell of mildew greets you as you open the door and in the north west corner you spy a stack of what appears to be rotted leather-bound books".

    I'd also like one that doesn't try to tell a story and slowly reveal things to me, the DM, as though this is some sort of sub-par fantasy novel. I want one that basically "speaks" to me as though it's my gaming buddy or DM-mentor. Cut to the chase, tell me the secrets for running this level well:

    "This level highlights a war between the goblins and kobolds. The goblins are much more likely to negotiate with the PCs and may even attempt to hire them as mercenaries, trading free access to their "realm". If the PCs do make peace with the goblins record their contacts (the Goblin king on level 4 will want this information if they descend that far)." etc. etc.

    Lastly quality maps are very important and I'd love pics of some of the main rooms and unique monsters.

    Finally lots and lots of tables (even for fluff things).

  32. @ Nick: "meg·a/ˈmegə/
    Adjective: Very large; huge" - Mega means Million (x10^6), Giga means Billion(x10^9), so a megadungeon would have a Million Rooms.

  33. I'd like to see a megadungeon product that was just the maps of levels and sublevels for almost the whole dungeon. Despite my interest in it, one of the reasons I have yet to pick up ASE is that it only presents one level, leaving me to do the rest of the work. It still seems like it has a lot of good ideas so I'll probably pick it up at some point.

    Anyway, back to what I'd like to see: I think the best product would be just the maps and special encounters. Then there'd be discussion of the dungeon "factions" and what level they can each be encountered on. I think after that it should provide advice and charts to help the referee stock it on their own.

  34. I think a megadungeon should be presented by:

    1. Engaging background that can be learned by the PCs
    2. A homebase
    3. Tables of rumours and myths
    4. Really good maps
    5. Well developed Saturday Night Specials
    6. Sidebars about how the Saturday Night Specials fit together, NPCs interact, etc.
    7. Useful tools for stocking/dressing the areas between the Saturday Night Specials.

  35. So I think what yellowdingo is saying is that someone needs to start on the *million room dungeon project*...

  36. So in summary we have:

    1. sandbox setting with background/history
    2. modular/related areas/zones
    3. rooms with some description (1-2 lines)
    4. monsters/factions & motivations
    5. decent maps, keyed & unkeyed
    6. tables, tables, & some randomness

    Okay, let's get started...

  37. So, if each of Grognardia's followers design 1,000 rooms we'd have our Mega(million-room)dungeon! :)

    Almost too easy.

  38. I love the concept of Megadungeon, it's just that I think I may have developed a certain attention deficit syndrome about them since I stopped having regular classic dungeons in my games (for the most part) many years ago. But, my fascination with the mythic underworld remains.

    I think to keep me happy, and therefor my players happy, I would need to go with the modular approach. Mostly for the reason that I would like for the party to be able to leave the dungeon (at least the upper levels) fairly easily for town visits or other side non-dungeon adventures above ground. The easily pick up where they left off. This would also be helpful for one-shot games at game days or cons.

    I think that modular method would work best for me.

  39. DungeonADay is my favorite megadungeon. As a hyperlinked web page, it is amazingly easy to run from a laptop. Sure, it is for 3.X. But really, converting to older editions has always been relatively easy.

    One of the features I really love from DaD is the revisit sections for rooms. Suggestions on what monsters might move in and claim the territory after the PCs have defeated the original inhabitants. Goes a long way toward making the dungeon feel a vibrant, ever-evolving place.

  40. @yellowdingo

    So in other words ... there have been no megadungeons. Which means the object exists solely in a theoretical state. ...Good to know.

  41. I'm pretty fond of the way that Fight On! has been releasing a Megadungeon level by level. The levels aren't tightly tied to each other, so each dungeon layer is useable by itself or as part of a grand dungeon, and each section is being worked on by different people and so they've got different styles of presentation to go with the different dungeon levels.

  42. I would go with not so much a "finished" megadungeon (is there such a thing?) but a toolkit (a la Zak's Vornheim) for continuously generating and improvising the megadungeon.

  43. As a lot of others have said, step one is a map. I think the best way to do this map would to make it some kind of computer-based document that would show more detail in a sidebar when you hovered over landmarks in it. A little bit like Roger Burgess and Zak S. said.

    Step two would be what Ed Dove said, make the whole thing modular, to allow for better customization.

    Step three: it has to make sense. There can be areas of total funhouse if it seems suitable, but let me tell you why, after initially falling in love with Castle of the Mad Archmage, I quickly found myself frustrated by it: what are these human berserkers doing in a previously sealed complex, full of hostiles, a great many stories underground?

    Essentially not only should the monsters seem to have logical relationship to each other-- they need to have a reason that they're in the dungeon and a method of ingress. This goes double for anything sapient or that normally lives on the surface.

  44. A lot of published adventures conflate multiple tools into a single reference system. It's sub-par, but they get away with it because of a small scale. When dealing with a mega-dungeon, however, you've got to break out the tools and make them easy for the DM to pick-up and use. This means that, in general, you need to separate:


    STATIC ELEMENTS are mostly the architecture, but may also include some stationary monsters. (Like an otyugh who never leaves his garbage pit; or a stone golem guarding a specific door.)

    FACTIONS are most of the inhabitants in the dungeon. By separating this information and then cross-referencing it to the map key, you make it MUCH easier to handle non-local reactions (reinforcements, etc.); re-deployment between sessions; interactions between factions, and restocking (among other things).

    EVENTS/DEVELOPMENTS refer to how the dungeon evolves over time. Obviously a lot of this will depend on what exactly the PCs do, but in a published product giving some options for general possibilities makes sense. But having this information mixed in with the encounter keys isn't useful: It's too difficult to reference (since you've spread it out randomly all over the place) and it makes it more difficult to use the material you DO want to access at the table. So split it off into its own section. (If you've got clear factions, this also makes it easy to organize this information.)

    Finally, I recommend indicating some clear ZONES within the dungeon. Basically, this is the "faction A controls zone 1" and "faction B currently controls zone 2 and zone 3" stuff.

    Basically, it's about organizing discrete elements into larger chunks that it's easier for the DM to both understand and manipulate. Grouping inhabitants into factions gives you one set of chunks; grouping locations into zones gives you a different set of chunks. Cross-referencing the chunks gives you an easy-to-manage dynamic.

    Compare this to, say, Temple of Elemental Evil: Not nearly large enough to be a mega-dungeon, but still too dizzying in its complexity to run efficiently without a lot of extra prep work.

  45. To more or less echo what Zak and others have said, the elements I would most want out of such a product are
    - a straight-up, to-the-point backstory and summary of the megadungeon.
    - a straight-up, to-the-point summary of each level and and an explanation of how the various levels fit together.
    - level by level descriptions of the major denizens, landmarks, and special features of each level.
    -maps of each level that show the major lairs, landmarks and special features, but that leave plenty of space to be filled in as needed by the GM.
    -random tables for monsters, treasure, traps and additional specials, unique to each level, that can be used for filling in those empty spaces when the GM is tired or low on prep time.
    -a couple of pages about the experiences of the playtesters, GM's included.

    Other cool, but not necessary features you might include in a product like this might be
    -a short list of NPC's who have their own reasons for exploring and exploiting the megadungeon, whom the PC's might choose to befriend or make enemies of.
    -a short list of monsters that the PC's might inadvertently release from the dungeon that then make their above-ground lives more interesting.
    -a secret cow level. (Ok, not really.)

  46. So one thing I think that people forget when they talk about how Gary ran Castle Greyhawk is that he was running many simultaneous groups with tons of people going through it. Is anybody else doing that? I run with one group - Gary's problems don't need a solution for most of us. So the re-stocking part isn't really a huge deal - personally I start hand-waving some areas after the umpteenth time through.

    Now, what I look for (well put into, since I'm writing Anomalous Subsurface Environment) in a megadungeon is tons of flavor - things that make the players try to connect the dots, whether there are dots to connect or not.

    I'm also much bigger on traps & "specials" that the players need to work around, rather than combats. That's a flaw with the one-page dungeon format - how do you dump in a lot of interesting traps & tricks when you've got a sentence to describe a room? It's a formula that leads to a combat-heavy dungeon, in my opinion.

  47. I am enamored of the idea of megadungeons, but you're right that they're very rarely presented in a good format.

    Ones like "The World's Largest Dungeon" *shudder* throw way too many numbers at you with very little fluff and, worse, a map too big to navigate effectively.

    Night Below, back in 2nd Ed, at least had multiple poster maps and broke the campaign up into smaller chunks.

    I'd actually love to see a megadungeon presented as an application that a DM could click through. Have a large visual map that the DM could click on to bring up the details associated with the room.

    It would eliminate the need to thumb through a bulky hardcover book and do away with over-sized poster maps that are confusing for DM and players alike.

  48. I like the World Largest Dungeon. It's very much in the spirit o those 70's third party OD&D dungeons like "Deep Delve" or "Palace of the Vampire Queen " were the GM was expected to mess around with it so it would fit into his campaign ( as any good module should) It might not be a good for someone who wants to play something out of the box without putting much though into it, but fo a GM who likes to tinker with his adventures, yeah, it's pretty least for myself.

  49. @Pat: All my megadungeons (starting on my third now) have had multiple parties (or individuals) adventure in them. It's definitely a lot more fun. [It's also my preferred way of running continuing campaigns, but it gets harder to do as people become less available due to other time commitments, whereas people can usually make time for an occasional dungeon delve, especially since you don't have to go with the same party each time.]

    @Nick: I'm presuming here, but the focus of a campaign with a megadungeon is the dungeon itself, and the surrounding world generally exists only as an adjunct and in support to it. When the outer world becomes just as, or more, important, you simply have a very large dungeon.

    They can even exist simultaneously in the same world. The archetypical example of this is Castle Greyhawk and the World of Greyhawk campaign. In the first the focus is on the castle itself (with side adventuress involved in getting back to it if you got to the very bottom), and so it is a megadungeon. In the second the focus is the events that are happening in the world around it, and so CG becomes simply a very large dungeon in the campaign world (possibly a source of side-adventures), and not the focus of the game.

    The Ruins of Undermountain on the other hand is simply a very large dungeon located in the Forgotten Realms campaign.

    In my first megadungeon almost everything outside the entrance was relatively undefined (and the entrance was only reasonably defined because people had fun there). The second was located on a remote tropical island, so adventurers had to come in by ship (and were allowed to create whatever they wanted as background for the world "outside"). The third, in a homage to Runequest's Pavis is located a difficult trek through hostile nomads who believe that it is evil and that anyone who attempts to exploit it must be killed.

  50. I'm sorry, but I get very confused on this: why isn't a 'really big dungeon' a 'megadungeon'?

    A megadungeon (or "campaign dungeon") is a dungeon complex that is both the focus of a campaign -- a "tent pole," as explained in the post to which I linked above -- and too large to ever be "cleared" by even several parties of adventurers. Megadungeons are designed to be an ongoing and permanent feature of a campaign, not used and discarded like most dungeons published in the past.

  51. @ Rach's reflections: If I may, I think you're missing the entire point of something like Castle of the Mad Archmage. You ask "what are these human berserkers doing in a previously sealed complex, full of hostiles, a great many stories underground?"

    Well, at the risk of sounding trite, that's up to you, as game master, to decide. Maybe they were sent to that room in CotMA by Odin when there wasn't enough room on the mead benches in Valhalla. Maybe they're cursed, not realizing just how much time has passed. Maybe they know exactly what they're doing, and are gearing up for an all-out assault on the level that never quite materializes. Maybe they're Zagyg's personal guard, kept entertained here until they're needed elsewhere.

    As far as the other hostiles on the level go, maybe they know just how tough it would be to take out 50 berserkers, and give them a wide berth. Maybe they've struck a bargain with them. Maybe they don't even know the berserkers are there. If so, why? Figure it out.

    The whole point of a mega-dungeon module like Castle of the Mad Archmage is that it is replete with gray areas like that. If I had put in every motivation for every NPC, and a justification for every trap, treasure, monster, etc. the whole thing would collapse under its own weight. It would literally be impossible for any single GM-- myself included-- to internalize enough of that information without having to endlessly consult the written page. By providing just enough information, and placing a plethora of tantalizing "hooks" like the Mead Hall in there, I allow you to breathe life into the dungeon and turn it into a very very different place than it is when I run it.

    And that's what I want to happen. Make it your own! Don't see an explanation? That's on purpose! It's room for you to explore and expand off the cuff if you want, or with a little planning and forethought it you work better that way. Castle of the Mad Archmage is the framework only. You need to fill in the details, by design.

  52. And I should add that some of the best explanations for such things come from the players themselves. While they engage in agonized speculation about what those berserkers are doing down there, don't be afraid to take a good idea they float and run with it. Or turn it on its head. You're not cheating. You're improvising.

  53. I have to say Joe, I go the other way. My preference is somewhere between your CotMA and Castle Zagyg: Upper Works.

    I like my dungeons to make sense in some way. Logically hang together as to what is where and why, and how all those elements and creatures form a little ecosystem. I look at it as part of the puzzle or story for the pc's to solve. On some level it has to make sense why the room of bugbears is across the hall from the room with the roper in it, and how they are able to actually live together, as well as survive individually.

    As a player I like to try to figure out the dungeon's internal logic, or story, in order to gain some advantage over it. As a DM I can't just design random shit that makes no sense to me. Someone designed the thing for a reason. That reason may have been lost in time and different denizens are there now, but both the original intent and the current sentup should be inteligible to the players and DM. Even the maps have to make sense as to why they are laid out in such a way.

    Your dungeon falls short in that regard, for me, as for what I look for.

    On the other hand, Castle Zagyg give it to me in spades. Too many damn spades. The descriptions and backgrounds and interconnections are too intricate, there is too much written about it all in a lot of places. I don't want a travelog, just a module.

    The whole reason to play the module in the first place is because I have no time to make something up, or I want to give the players a chance to jump into someone else's creation, and experience something created by another person. I don't want to "make it my own." I just want to play in their sandbox for a while.

    Here's what I want: a dm description box to read to the players, no more than one or two sentences. I like these for a couple of reasons. First, they give a quick glance at the room, give the players something to chew on while I as DM quickly scan the dm info section. Second, they set the tone or flavor in a way that the author's voice is expressed. If I buy a module by Rob Kuntz, for example, I want the ability to run the module for my players as if their DM was Rob. The only way for Rob's voice to be there is thru the captions.

    Next, I would like the next section under the caption to be a short paragraph giving the dm all he needs to know to run the room if the players quicjkly decide to do something. Also, it gives me a way to quickly refamiliarize myself with the area, as in "Oh, this is the medusa room with the swan." As DM's we can prep for a game, but no amount of prep will let me memorize a megadungeon.

    After that, give very brief details of various objects/monsters in the room that are of note.

    At the start of each distinct section of the dungeon complex, give me a paragraph which summarizes what it was when created, what it is now, and how its denizens interact with the areas around it and how they are able to live and breed and survive. Basically, give me its mini-atory.

    Random encounters must make sense. On a level dominated by demi-humans, why would there be zombies if there wasn't a necromancer's lab or graveyard nearby? Or a random poisonous spider encounter when every room is fully developed and there are no spider lairs anywhere? I don't want to make up a BS reason, as I think you lose credibility with your players that way.

  54. part 2

    On the maps, give me enough info by way of syumbols that I don't have to refer to the text of the module to read the contents of each room. Have the map make sense in layout, as to its original creation purpose. Show me a real live place that someone lived in, let me imagine what it was, in a way that makes sense. In other words, put the latrine near the barracks, rather than next to the king's quarters or the alchemist's lab. No latrine? Fail. Everyone shits man.

    All that being said, I don't like megadungeons because you never get to see the world around you and never get to conquer it. You go into a hole in thye ground, and some months later you come out a badass, with occasional trips to town to resupply and sell. That's just boring as shit to me. My brother ran a group of friends and I through Return to the Temple of Elemental Evil, and I tried to sabotage the fuckin' thing every chance I could get just to get out of the goddamned hole in the ground, so I could enjoy some of the most fun levels in a character's career above ground in the world.

  55. Just to clarify, I don't mean to portray Joe's Castle of the Mad Archmage as a compilation of random encounters. It's definitely not that. I like it. I ran it at a con last year in conjunction with Castle Zagyg. That's where the stark contrast really hit me. There are a lot of good parts to it that do hang together and tell the mini-story, but there are some that don't. Some of the stuff in Archmage is damn brilliant and I steal it for my own stuff. But I like it all to hang together, but not as much as Zagyg in terms of the details. If you could throw Mad Archmage and Zagyg together in a soup pot and see what comes out, it would be the perfect megadungeon for me.

    It's all about personal taste.

  56. "I don't like megadungeons..."--JoetheLawyer

    Did you not read the part of the original blog post where Mr. Maliszewski said "please don't use the comments section to express your disinterest"?

  57. Disinterest is different than dislike.
    Someone who is disinterested doesn't take the time in the middle of a fuckin' hurricane to post 7000 words on it. Mr. Maliszewski is free to delete my shit if he doesn't like it. It's his blog.

  58. "Disinterest is different than dislike."--JoetheLawyer

    So you think that, when Mr. Maliszewski said "please don't use the comments section to express your disinterest", he meant to allow for people to express their "dislike" of megadungeons?

    Stop acting like a lawyer for a moment and just admit that you did something wrong.

    Now, back to the actual topic...

  59. It’s unfortunate that, apparently, almost nobody is interested in any megadungeons that don’t have at least very extensive (if not complete) maps right from the start because that’s likely one of the main reasons why so few megadungeons get published. If enough people would just accept the idea of building megadungeons manageable-piece-by-manageable-piece, then there could be lots more of them to choose from. And we wouldn’t even have to choose from whole megadungeons either. We could choose only the pieces we like best and use them to build our own megadungeons. And chances are that some people would even post free megadungeon pieces online too. But, because, apparently, almost nobody is interested in any of that, it probably won’t happen. And that’s a shame.

    I wonder if the assumption that megadungeon projects should have at least very extensive (if not complete) maps right from the start is based on the assumption that they should have ”a précis of the entire dungeon, including revealing any secret or ultimate goal of the dungeon, as well as general design philosophy” right from the start too? After all, if you don’t assume that there should be ”a précis of the entire dungeon” or ”any secret or ultimate goal of the dungeon” or a ”general design philosophy” right from the start, then there’s not much basis for assuming that there should be at least very extensive (if not complete) maps right from the start either. Right?

    And, if you don’t assume that there should be any of those things right from the start, then there’s no reason to assume that there should be things like a ”backstory and summary of the megadungeon...summary of each level and an explanation of how the various levels fit together...level by level descriptions of the major denizens, landmarks, and special features of each level” either.

    And, once free of all those assumptions, it’s possible to build and run megadungeons much more quickly, not just think about them endlessly.

  60. I got to the paragraph in joethelawyers post that began with "All that being said, I don't like megadungeons" and I was all "erm, well Why did I just read the previous five paragraphs??"

    You coulda told me at the outset at least... :)

  61. Personally, I think megadungeons are fundamentally non-sensical - and I don't mean that in any negative way. I just think that much of any effort spent trying to justify and rationalize the "hows and whys" of them is wasted. To me they are basically vortices of chaos and the rules of the normal world don't apply within their dark corridors. I think they work best if you just embrace the gonzo ftmp and don't worry too much about what keeps it all going.

  62. I thought of another question about why apparently almost everybody interested in megadungeons wants them to have things like at least very extensive (if not complete) maps, ”a précis of the entire dungeon, including revealing any secret or ultimate goal of the dungeon, as well as general design philosophy”, and a ”backstory and summary of the megadungeon...summary of each level and an explanation of how the various levels fit together...level by level descriptions of the major denizens, landmarks, and special features of each level” right from the start...

    Are people who want those sorts of things mostly interested in designing or reading about megadungeons -- not actually running them?

    I wonder because it's so much easier to run a megadungeon that's presented in manageable chunks rather than as a massive thing that's difficult to digest & use.

    So why does apparently almost everybody interested in megadungeons want them presented as massive -- difficult to digest & use -- things?

    The only rational possible reasons I've been able to think of are that they either find the challenge of designing such things fun or want comprehensive megadungeon tomes to read but not actually use.

    Are there any other reasons why people who want such things want them?

  63. @Ed Dove: I don't think that simply because a published megadungeon is large it necessarily needs to be either non-modular or difficult to digest and use.

    If something were presented in a series of a dozen small modules, rather than a single large one that contained exactly the same information, would that somehow be better?

    To take my own as an example, if you just read through the first level of Castle of the Mad Archmage, you could do so easily and run the level with little difficulty (the comments of folks like Joe the Lawyer notwithstanding). Once the party found a staircase to level 2, you could then read the chapter on level 2, and so forth.

    That's not to say, of course, that there aren't any connections between them. There are, and the relevant information is contained within. There are also, of course, sections that interact with one another more vertically than horizontally (intentionally so, to go against some of the "each level is an isolated unit from the rest" mentality that such levels can sometimes, unconsciously, foster).

    But if I'm reading you correctly (and I certainly admit I might be misunderstanding your point), it seems like there would be little difference in just acting like a large published megadungeon was, in fact, published in smaller chunks, than how it could be approached if it actually consisted of dozens of smaller modules. Am I missing something?

  64. @Joseph:

    What you suggest sounds right in theory, but -- in my experience with things like megadungeons, city books (ones like CSIO & Ptolus, not Vornheim), campaign settings & supermodules -- doesn't actually work well in practice.

    The impediment is the physical format -- a book -- that's harder to use sections of than if those sections were completely separate modules with cover folders with maps on them that are bigger and easier to use than maps in books are, and booklets that are easier to navigate than books are -- and that lay flat & stay open more readily than books do too.

    So the only reason I've been able to think of why anybody would rationally prefer things like megadungeons, cities, campaign settings & supermodules in book format is because they mostly just want to read them, not actually use them.

    And, of course, anybody who's gone to all the effort of creating a whole megadungeon, city, campaign setting or supermodule probably will prefer to sell the whole thing all at once in book form for a high price rather than sell it in more useful pieces and hope people keep buying the pieces (even if they could possibly make more money from selling it in pieces if people bought all the pieces).

  65. Dove: I doubt the audience of this blog or the OSR in general has money making as their first prerogative. They try to cover costs, and some are trying to make a go of it as a source of income. But even those that do see producing these works as income are (if they are realistic and have some sanity about them) doing it for the process and the end result, not the money.

    There's a different format between individual booklets and the hardback: the box set.

  66. @Ed: Sounds like your beef is more with the physical format of books, rather than the conceptual design of megadungeons per se.

  67. "I doubt the audience of this blog or the OSR in general has money making as their first prerogative. They try to cover costs, and some are trying to make a go of it as a source of income."

    I'm sure you're right. But even just the desire to merely cover their costs will lead people to produce books instead of other formats if they think they'll be more likely to cover their costs by selling books. And, in cases where other formats would be more useful than books, that's unfortunate for those of us who want to use these things, not just read them.

    "But even those that do see producing these works as income are (if they are realistic and have some sanity about them) doing it for the process and the end result, not the money."

    I know. That's exactly what I meant when I speculated that they might "find the challenge of designing such things fun".

    "There's a different format between individual booklets and the hardback: the box set."

    Yes. And boxed sets are an ideal format for complete megadungeons (and cities, and campaign settings, and supermodules). But they still require completing the whole project before producing anything though. And they also require buying the whole project instead of just parts of it too.

    "Sounds like your beef is more with the physical format of books, rather than the conceptual design of megadungeons per se."

    Not exactly.

    While I do think that megadungeon (and city, and campaign setting, and supermodule) products would be much easier to use (and also to produce) if they were produced in modular formats, I also think the reason why they're not produced in modular formats is because the predominant assumptions about the proper conceptual design of megadungeons don't allow for that.

    So you're right that I don't have a beef with "the conceptual design of megadungeons per se". What I have a beef with is the assumption that megadungeons should have any conceptual design at all.

    That's why I advocate ditching that assumption.

  68. I forgot to mention that, despite all I've written advocating & explaining modular design, I think the best idea presented here so far has been:

    "...not so much a 'finished' megadungeon (is there such a thing?) but a toolkit (a la Zak's Vornheim) for continuously generating and improvising the megadungeon."--Matthew Miller

    Though I also think that...

    "A series of 32 page pop-up books."

    ...could be pretty cool too. ;o)

  69. Not to belabor the point counting how many angels are on this pin, but...

    A megadungeon box set box could include the general books such as setting assumptions, global maps and the like as well as the first set or so of modular components. Each additional modular component (32 page pop-up or fold out map cover, etc.) could be made to fit within the box or used on its own.

  70. "A megadungeon box set box could include the general books such as setting assumptions, global maps and the like as well as the first set or so of modular components. Each additional modular component (32 page pop-up or fold out map cover, etc.) could be made to fit within the box or used on its own."--Red

    That could be really cool!

  71. I know that I am a little late to the party and that I havent read through every comment but I will put in my (probably redundant) 2cp anyway.

    I would like to see a MD presented in 3 major parts:

    1) Maps - Detailed maps of the entire MD as well as the surface lands around it. These maps should detail pretty much only the unique and "fixed" physical aspects of the dungeon (ex. Moon Pool, Cleric Tree, Portal Room, etc.), perhaps, with some misc details that can be added to rooms or one line seeds for use in non-unique rooms in case the ref gets stuck (heck, it IS a big dungeon with lots and lots of rooms!). Also, I think that having some details on physical elements (I am thinking traps) that can be placed as needed would be useful here.

    2) NPC & Monster Manual - This should be an encyclopedic treatment of all of the NPC that exist inside and surrounding the MD as well as all of the potential monsters that exist within your world (i.e. your ecosystem - at least, in how it varies from the typical Old School ecosystem). I would also include stocking/encounter tables here too. Seeing the number of comments in this post that indicate people want their monsters keyed to levels, perhaps a small, separate appendix/supplement that details the potential MD inhabitant ecosystems per level.

    3) Background - This should be all of the material that, when combined with 1 & 2 above, make the campaign setting into a unique world. Deities, races, factions, back story of the world and its peoples, etc. Additionally, this piece should have a summary of the dungeon and its "secret" (if any), details about how the levels fit together within the story material, as well as the "evolution" of events within the world assuming there were no PCs involved (if this actually fits with the materials you have developed).

    By splitting the MD up into these three major parts I feel as though you cater to more audiences. For those who want to add a MD to their existing or already detailed world they can simply buy part 1. For those what want to start a MD campaign in a unique setting, they can buy parts 1 & 2. For those what want to actually run the Dwimmermount campaign itself...they buy all three parts.

    I think this is the most reasonable way to present the material since it is the closest approximation of how the material was created to begin with; you created some maps, you created some NPCs, you created some basic ideas about the settings story. Then through play all of these things got put together and fleshed out. Having these three elements separated out will allow refs to put the pieces together how they see fit, or if they prefer to use Part 3, how YOU saw fit.

    I look forward to seeing your thoughts on the matter!