Monday, 27 February 2017

[BLOG] The Overly Thematic Dungeon

The common wisdom surrounding designing dungeon encounters has changed much over the years, yet the question of what makes for a good one, or what makes for a good room mixture has never been satisfyingly settled.

A wizard did it
The original approach, developed at the dawn of gaming (and seen in such tattered artifacts as the El Raja Key Archives or First Fantasy Campaign), stressed the game aspect with a very brief key and very sparsely “seeded” dungeon levels. You would spend a lot of your expedition time looking for the carefully hidden lairs and those memorable “special” encounters, and – from our perspective – some of these games might now be described as first-person crossword puzzles.

This philosophy had a relaxed attitude about what goes into the dungeon: anything that’s fun and challenging, and damn those pesky questions about why and how. That’s how Citadel of Fire has an underground tavern on one of its upper level dungeons, how Castle Amber has an indoor forest, how Tower Chaos has an earth elemental named “Stoney” guarding the china room just off the kitchen, and how White Plume Mountain has... well, those canoes are a good start. You can rationalise it, but reason is an afterthought – what matters is the spirit of fantastic whimsy. At best, these adventures are great precisely because they take liberties with realism, and do it well. Without a vivid imagination and the skill to turn imagination into mini-games, the results just feel flat and randomly thrown together (this problem haunts much of the early tournament scene, including, in my heretical opinion, a significant portion of The Lost Caverns of Tsojcanth). The best examples of this approach were always the modules which had a sense of cohesion abound them – vague, hard to explain, but there in the background.

The consequences of lousy housekeeping
Then there is the fantastic realism school, first expressed in a comprehensive manner by an ancient Dragon article whose exact references I cannot be arsed to look up. [edit: Identified by Settembrini, and updated with the correct illustration.] You know the one. It shows a dungeon room in two states: the original way it looked, and the dilapidated, looted and repurposed state the party will find it during their expedition. Certainly, this approach provides a sense of realism, of “being there”, and it is actually more intuitive than stocking your dungeons with random shit. If your dungeon was a temple, you stocked it with religion-related encounters, and if it was a crypt, you sure didn’t put an underground tavern in it (and underground taverns just kinda vanished from the gaming scene). This approach often provided a complete blueprint for your dungeon: if you put in a sacristy, you might as well put in a crypt and a refectory, and how about a bell tower and some stables? It is no accident that this approach, lauded across the game design community, ended up the dominant one for decades, mostly displacing its predecessor. (It was, in turn, succeeded by the modern “return to the dungeon” model, a selective (mis?)reading of gaming history, which suggested that the good old days were all about “killing things and taking their stuff”, while silently dumping the heavy focus on exploration the actual old games had.)

There are many advantages to semi-realistic encounter design, but it can also go wrong in ways its proponents never considered. From my perspective, the most important of these is the taming of our sense of wonder, either by considering the fantastic impossible and an interest therein juvenile – a notion which had been particularly popular in Hungary, and as I hear, Germany – or by requiring the rationalisation of the irrational. This has a corrosive effect on any kind of fantasy game, but it is particularly damaging to D&D. Once you accept that fantastic things are dumb and beneath a serious person's interest, you remove much of what makes D&D worth playing. A “cabinet contents” dungeon of endless barracks with bits of string and mouldy old boots stuck in a succession of footlockers, or the “this used to be a scriptorium, where scribes scribed their scripts” school of pseudo-historical flimflam is often a recipe for a dissatisfying dungeon where nothing interesting happens. It subordinates fantasy to reality, when it should have done the exact opposite. In the end, one gets the idea that these dungeons are not worth playing. “Told you so” say the people who never liked D&D in the first place.

Skulls. Why did it have to be skulls?
Rediscovering the fantastic side of RPGs is an important achievement of old-school gaming. And there is no reason why we can’t learn from multiple design philosophies and take the best they have to offer. My go-to compromise has been to go for thematic appropriateness, an approach found particularly often in Bob Bledsaw’s writings. Thematic appropriateness links its encounters to an overall theme (be it a crypt, desert oasis or teeming fantasy metropolis), but operates on the basis of loose associations instead of solid, step-by-step logic. 

When you say “port”, it says “old panhandler sells musical sea shells with secret messages, 1:6 of ear seeker”. When you say “jail”, it says “Bluto and Balfour, two ogres (Hp 17, 23) administer regular beatings and serve inmates Seaweed Slop; prisoners are Refren, musical pirate, Harko Fum, beggar of the 4th circle, Mythor Flax, last bearer of Princess Yarsilda’s shameful secret”. There are obvious connections here to a basic theme, but also large jumps of logic – somehow, we got from that port to an ear seeker and from a jail to a princess and her secret, although it does not immediately and necessarily follow from the starting point. You have to believe in your ability to jump to make it – you have to let go a little. This is how dreams connect things in our mind and how the better kind of random tables can prod our imagination: by coming up with odd juxtapositions and fantastic things that nevertheless feel real as long as we don’t open our eyes too wide.

This was the conclusion I adopted a bit more than ten years ago. And yet, despite having been well served by the approach in multiple different campaigns, I am finding that it should have come with an important warning: use your themes, but don’t let yourself get bound by them. Most recently, I have experienced this the hard way while experiencing a creative block coming up with encounters for Castle Xyntillan. As straightforward as designing about three quarters of the castle proved, the remaining quarter (and the dungeon level) has proved a tough nut to crack. I found myself in that state where I am too analytical, too much of a cynic to have good flow – I could probably continue through via sheer willpower, but the result would inevitably disappoint myself. What went wrong? A simple creative block would have been a convenient excuse, but after a little self-examination, I came to the conclusion that I let a coherent vision of Xyntillan overpower my idea of it as a loosey-goosey funhouse dungeon with improbable things. The existing structures and ideas of Xyntillan were closing off the range of ideas I entertained at the beginning. My thought process became path-dependent, predictable. All in all, I needed a break – not just for refreshment, but to forget and let myself wander again in directions I am not expected to go. Xyntillan needed to be less thematic to retain its theme.

Which again proves: there is a point where theory ends and fuzzier realms of the imagination begin; and in those worlds, we must often walk alone.


  1. You are not only very right, you also put in in words perfectly.
    I would like to stress the point that even early Dungeons were not played for random shits and giggles. There always was a contextualising Campaign environment providing relevance as well as continuity. Also, a blissful state of flux between Mythic Underworld, Giant's bowling lanes and beancounting logistics of Castle Sieges exists... it is the centre of the Rientsian Retro-Stupid-Pretentious model. Like many British TV shows, D&D throws a monkey wrench at small-minded people.

    1. That's a good point. Why declare that a campaign is about X, Y or Z, when it can be all three at once? A series of adventures, like a good book, can be about many different things while maintaining its own style and sense of consistency.

      To take things a little further, this also applies to the campaign's scope and areas of interest. In First Fantasy Campaign (and the less well known Tony Bath's Ancient Wargaming, which you or Pierce recommended to us back on Disputorium), we see games which expand to encompass the players' growing interests, from fantasy battles (the original starting point) to rudimentary diplomacy, Civilisation/4X-stlye domain management, dungeoneering and tunnel construction (the offshoot that developed into RPGs), sea battles, jousting, trade and who knows what else. It is there in OD&D vol. 3 (which is probably less often read than even old-school people would admit) and to an extent the odd corners of EGG's Dungeon Masters Guide, but it is an aspect of gaming that very few people have picked up, let alone developed any further.

    2. Yes, for me this is the basis of RPGs, different modes and ignoring either always felt like a reductia ad ferraria.

      Nota Bene: Pierce pointed to the tome, but I had been organically playing like that since a very early age, albeit with all the disorganization that early strategic gaming entails.

  2. BTW, I did not know the article and had to look it up: "Let There Be A Method To Your Madness" tDm Nr.10, pp.11 by Richard Gilbert.

    1. What do you know, that's the one! Thanks; updated the post with two extra pictures. Yesterday evening, I spent about an hour trying to scour the Internet for that room picture without results. My best bet was James Maliszewszki's blog, since I expected this article to be his thing ("Gygaxian naturalism" - mine is "Bledsawian surrealism"), but surprisingly, it's never even mentioned.

  3. Hah! So I feel less stupid for not knowing that one and looking it up nased on free association of keywords and the wonderful DragonDex. Took me half an hour btw. <>

  4. Great thoughts, Melan. As I have keyed my Foolsgrave megadungeon (Level 1) for Gary Con, I regularly ask myself, "Is this fun to screw around with?" and let that answer be my guide. I may have a primordial swamp filled with giant bugs from another dimension on the north side, a undead-filled crypt on the south side and frozen, icy tunnels on the east side, where evil dwarfs worship a frost giant. I can promise you that although someone will complain this is not realistic, when I have eight people at my table for four nights in a row, they will have fun and leave thinking the place was magical and weird - anything can happen!

    1. Thanks, Chainsaw! That's the way to go, especially with megadungeons, since they need lots of variety to keep exploration interesting.

      IME the most practical way to stock things is to build and stock large dungeon levels in a dual fashion:
      - have a few more dense adventure nodes which are organised around a common idea, like an underground garden and aviary, an evil temple visited by meditating monks, a river area with a fungus forest; - and in-between, use less dense connective tissue with smaller, more varied ideas (which may take some exploration to find, and some of which can be very well hidden).
      That's a reasonable compromise, and it lets the players experience both the overall style of the dungeon, and the thematic areas you have created.

      I am also happy to report that the creative crisis I reported in the post seems to have passed: after a two month break, we returned to exploring Xyntillan this Sunday, and the ideas are flowing again. :)