Gus’ Megadungeon Thoughts

Gus wrote a thought-provoking article about the node-based megadungeon I developed a few months ago.

I agree entirely with his post. As an adventure, the node-based megadungeon I presented is woefully incomplete. At the least it should include the elements he describes (maps, monster tables, etc.).

What it really needs, I think, is to have the various identified nodes expanded on. I would be inclined to apply the same techniques as I have so far, but at higher resolution. The method I use is more or less fractal — just as I started with about a dozen high-level nodes identifying the regions, then broke each region down into areas, those areas can be broken down into encounter locations (‘rooms’, if you like, though an individual encounter location may span several). I might expect that different encounter locations have monsters or other things to interact with (set pieces) while the regions have the encounter tables and the like (including entries to pull from other regions — when in the Clockwork Hell you might run into Aristothanes or something else from his Sanctum, you might run into the dwarves or something else relating to the Dwarven Safehold, and so on).

I talk about next steps in Node-Based Megadungeon: All Regions Outlined, and they include

  • Creating encounter lists for each region, identifying likely creatures and events that could be expected to come up in play.  First pass would be a list of creatures and events, second might place them into specific nodes and/or wandering monster tables.  For instance, in Clockwork Hell you might find dwarves almost anywhere (random encounter) but there will almost always be dwarves found in the ‘Dwarven Crafters’ node near the ‘Repair Bay’.
  • Polishing this model.  Some of the nodes can be collapsed and merged, some of the relationships are a little dodgy in retrospect.
  • Creating a nicer map.  GraphViz is a great tool for initial layout so I can see the relationships between the nodes, but they are not all arranged the way I would like to see them.  For example, the Dwarven Safehold and the Fungoid Cavern should be ‘more horizontal’, the Wolf Den should ‘surround’ the Goblin Warren, and so on).  If I were rendering this for publication I would try to make each region more representative of its actual shape, and possibly construction.  I think creating a map in the fashion I would like to see it requires artistic chops I simply haven’t developed.
  • Creating detailed maps for the regions, showing the actual physical connections between the nodes and the rooms (or whatever other subareas exist) within each node.

As I said, this is fractal. You could identify major adventure sites, dungeons and megadungeons, and link them together in a fashion similar to this. The same techniques apply up and down in levels of detail.

I wasn’t aiming to go to that level of detail, though. I was mostly doing this as an exercise to show how the methods I use work. I did some math that indicates that I spent perhaps 25-30 hours on this series as a whole, and that to fully flesh this out might be ten times as much… rather more than I would aim to spend on a demonstration.

If I were to run this again I would likely reset it and start over, developing the regions and nodes as I’ve described above, as it looked like I was going to need them. I left myself inadequately prepared to run this, and I am too far out of practice as a GM to really improvise well.

One thing I do like about it, and that was mentioned in Gus’ article, is that this provides a framework of an adventure that I could give to someone else to flesh out. Someone who has read the entire thing has some idea of what is there, and can be considered to “have heard stories” about it, with no real chance of having detailed knowledge that could destroy the suspense. Sure, it might not be a surprise that Shalthazard is at the bottom of the dungeon, and that he has ties to (manipulates, rather) several groups within the dungeon… but while this knowledge helps fix the tone of the place, there is enough room to mess with expectations.


  1. Incidentally, I am impressed and amused that while I last really worked on this series four months ago, and last played in the megadungeon more than three months ago, I remembered a surprising amount about the megadungeon as a whole and about the series describing it. The hierarchical structure of the graphs, combined with how they are developed in a way that focuses on relationships, makes it surprisingly easy to retain the structure in memory. I would have to look up the names of the various areas within the regions, but I can remember to a fair degree about what each region is and the major links between them.

    I would totally rebuild the Abandoned Tower, though. I’d keep the graph, the node structure, but the content would be entirely done over. I let myself get caught up in mundane design when I could have made it much more complex and interesting.

    On the other hand, I suppose if I had done that I might not have ended up with the mobile smelter and the T-Bag, the swinging hipster goblin….

  2. Just last week I started thinking about implementing some of the megadungeon things in WRPS as a practical experiment to help me think about stuff, so I had just noticed myself that it was underspecified a bit. I am therefore making stuff up as I go along!

    I commented with a few random ideas for sprucing up the Abandoned Tower.

      • GreyKnight

        An RPG I’ve been working on for longer than I care to remember. I keep changing things so I don’t have a link to the most recent rules right now (although I intend to do some writing over the weekend). It is getting more stable though, and I think I have finally cracked a long-standing niggle with attributes… sadly fixing it means I have to abandon the SPQR acronym for attributes, sadface.

        It initially started as the offspring of GURPS and D&D 3.5, but I think that it’s changed sufficiently that this isn’t a useful illustration anymore, really. It’s also borrowed a bit from Echelon, for one thing.

        I have a rough conversion guide for ACKS->WRPS and back, so it should be FLAILSNAILS-capable. Probably I should define conversion directly with OSRIC for greater reach?

  3. David Lamb

    I think Gus is being a little unfair. I found your original abstract outline a very useful step for getting an idea of what is going on. His suggestions are a possible next set of details in making things more concrete, but I think the original description of overall connections was useful in and of itself.

    I’m sure a software geek like Keith knows about architecture vs detailed design vs implementation!

    • I don’t think Gus was particularly unfair. This is an incomplete work, in the sense that it is not prepared for someone to use in play. I took it only as far as I had intended, which didn’t included the detailed development that would ‘finish’ it.

  4. Pingback: Path Extensions: Transitions in Node-Based Design | Keith Davies — In My Campaign - Keith's thoughts on RPG design and play.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to Top