Megadungeon from Scratch: A Retrospective

July 04, 2024

This spring I ran a megadungeon. As far as campaign statistics are concerned, it was not so impressive. I think we played 7-8 sessions, had a stable of 8 players, some of whom would not show up every session. Still, it was a fun and educational experience.

The genesis for this campaign was listening to the "Into the Megadungeon" podcast by Ben Laurence (available via internet search on a variety of platforms). I was particularly enamored by Luke Gearing's lunchtime megadungeon sessions.

Sometime later on a flight, I had the idea for the premise of the megadungeon. The original campaign pitch was the following:

< pole. It is the last truly wet place. The future was once buried here, in the form of millions and millions of seeds for posterity. But now they have sprouted. They have grown together, genes splicing, mixing, and blending. They have grown, both up and down and into other beings as well, creating a bizarre ecosystem. The old vault now lies open, its depths waiting to be explored.

This is hope. Urth may be restored from its xeric state. This is wealth. Some of these plants hold water very efficiently. This is danger. The proliferation risks overrunning every other living thing on the planet. This is history. Many of these plants have not flowered for epochs. This is the Orelian Gardens, an open table megadungeon set in Vaarn.

Inspiration:

  • Annihilation (Jeff Vandermeer)
  • Dune (Herbert)
  • Chimera Ants (of Hunter x Hunter)
  • Nausicaa (Miyazaki)
  • Svalbard global seed vault
  • childhood reading of plant and animal guidebooks
< /> Things to pique or drain interest:

  • obstacle/dilemma i.e. "traps"
  • faction play!
  • many (hopefully) interesting items
  • strange plants and creatures!

I ran it in the Vaarn system. This post is to analyze my attempt, considering much of my design was 'from scratch:' I didn't use anyone else's procedures, maps, etc., just the vague idea of how it should work. For example, I didn't use the x-in-6 room stocking procedure (monster, trap, treasure), but knew I should restock rooms as needed after each session.

First off, the map. I drew a map of the first level by hand in a dot-grid notebook, in a doodling fashion. I took opportunities to create loops, avoided dead-ends, and had 2-3 staircases down to next level. 

For stocking, I made my own random tables. Vaarn's monster creation is a one-number decision (it determines hit die, stat bonuses, and some attack details in one correlated way), so I just made tables for what kind of monster, and gave it a level depending on a d6 roll. This makes variations among different monsters of the same type, and creates potentially unbalanced encounters. 

I also stocked each room with "wallpaper," mutated plants that had grown in the room. I just made a d66 table of plants and roll twice, combining the result.

Treasure was interesting. Vaarn, it seems to me, seems to have overland travel more in mind than extended dungeon delving. Accordingly, the standard rules give a whole level, not individual XP rewards, for treasure. I changed the level granting to finding a pack of pure, un-mutated seeds that can be sold and advance a level. This makes less granular XP rewards, and less predictable based on player actions. I did have the pure seeds as part of the room stocking, but it's a bigger deal to find individual treasure.

I think a lot of this preparation got thrown out the window once play began. The first-level map was obviated when the player characters asked if there was an elevator shaft from the entrance, and I responded yes. Then the player with a plasma sword cuts open the doors and just descends to level 3. I think this is where the most fun happened, but I just made up room arrangements on the fly. Not ideal. It creates a big mismatch between the detailed level 1 and the totally improvised level 3.

The stocking also didn't quite go into prep. After running through some standardly stocked rooms, it got stale after a few. I like brainstorming interesting situations, so I just made the next room a more interesting encounter. This was particularly helpful for level 3 (the improvised level) as I can pack several sessions into a "sticky" room, where there's a lot to do. This gives me space between these sessions in order to patch up the map, stocking, etc.

But this became a habit. I found my pregenerated rooms to be kinda dull, and putting in an interesting encounter spiced it up. I think I could have done better stocking overall. My stocking was actually automated, by a short python script. I think doing by hand would allow me to inject more thought and creativity into the situations. But I just read a list of room contents during the session and made something up. This flopped most times.

Another thing about my map was that it was also WAY too big. I made level one something like 40 rooms??? I should do 10-12 at most, I think.

Despite the mismatch between preparation and play, I think this was overall a success. Everyone had fun, and though interest fizzled out after a hiatus (I traveled a lot over a few weeks and put the game on hold), I think I have a better idea of what a megadungeon is like, and how to prepare better. 

I think one thing I learned is to think about the relationships between the different dungeon denizens more. I needed to do more than just plop monsters and treasure in a room, and do it consistently.

Another thing I learned was that dungeon levels feel like progression, and so it should be happening more. Forty rooms on one level stagnates pretty quick. I think the elevator shaft is a great idea though, skipping ahead to more "difficult" content should always be possible. 

And I think another thing I want to do better is to have predictable progression, in levels. I think it's a more fair game if everyone can know how they get rewarded and what they need to do to progress. If GM simply hands out levels, it's a guessing and appeasement game for the players to figure out when they've done enough. XP-for-gold (or XP-for-seeds, in this dungeon), would have made the game more fair.

Hope this was enjoyable, even helpful.

<>

Comments

Thinking Small: The Fictional World has a Surprising Amount of Detail

March 13, 2024

Empty Rooms Should be Just That and the 5 Room Dungeon

September 01, 2024

Thinking Small Adventures

February 05, 2024