Tuesday, May 29, 2012

Designing Undercities

Giant subterranean cities piled out far under the sun-baked city streets. Byzantine, twisty layers of rubble-choked tunnels (nee roads), cavernous temple fanes, forgotten palaces piling on top of each other aeon upon aeon.

One of the elements I deeply love about Tekumel is that hot and forgotten world's vast undercities. In the core play area empire of Tsolyanu, a long-standing tradition called ditlana in which cities—at least in theory—are razed every 500 years and built over. Part sprawling megadungeon, part still-active underbelly of city life, they have had a fascinating pull on my imagination. Not surprsingly they have been long something I wanted to incorporate into my classic D&D campaign, yet I have never been quite able to pull it over the conceptual hump.

Why? Undercities on a Tekumelyani scales are seriously spatially-challenging beasts. I've had small-sized versions for many years, neighborhoods of cities conveniently bracketed off with cave-in's or very compact and sealed off little settlements (let me count the number of domed cities sitting in lakes I stole from the Holmes side-view example). But the truly sprawling undercity that matches and supersedes the larger environs of the city it sleeps under? Intimidating.

Just to give you an exact sense let's take a passage that Victor Raymond wrote in his ever-useful introduction thread to playing EPT “out of the box” about Barker's underworld under the port metropolis Jakalla:
“...The first level of the Jakállan Underworld is drawn on a 17” x 22” sheet of graph paper, 10 squares to the inch–and each square is ten feet in measure(!). Thus it is assumed to cover an area roughly 1700 feet by 2200 feet – almost 1/3rd by 1/2 a mile in size, centered largely under the ruined Temple of Hyáshra in the City of the Dead. If you consider the map of Jakálla, each hex has been said by Prof. Barker to be 50-100 yards across. To be fair, Prof. Barker has also said that the map is “semi-representational,” i.e. more important buildings appear larger on the map than they really are. Even so, the top layer of the Jakállan Underworld would require several 17x22 sheets to cover the entire city. This suggests that it would be difficult to actually map the entire Underworld...”
It's just not the horizontal hugeness of such a beast, you also have any number of other complications to conventional mapping to deal with: miles and miles of “empty space” (old roads, side-tunnels, etc.); the vertical dimension of historical layers that increase the total size many times; large numbers of entry points from the surface; and of course the potential active use of the upper levels by humans and other surface dwellers (thus also giving it the more complicated social dance of urban adventuring).

One could find a massive piece of architect's graph paper and laboriously fill each exacting section of your undercity in piece by piece, all the while trying to allow for the mental trick of knowing that your map still doesn't reach. I'm not going to do that.

What follows is less tutorial—as you can tell I am no expert—and more of me grasping for a method in designing undercities on a grander, more thoroughly thought-out scale.

Part I: Mapping the Layers
As much as I hate timelines and all the attendant setting-bloat whoha when tackling a undercity it undeniably makes for a practical starting point. I work better answering a series of leading questions while futzing with some kind of schematic.

How old is the city?
How many civilizations flourished here and for how long?

I am going to start with a young city by Tekumel standards, 5000 years, as my world is not quite as ancient. I am going to say perhaps five different civilizations each spanning a convenient millennium.

Who first founded it?
How did they order their city? What did they think was important?
Were they displaced or did they just evolve historically?
What lead to the city being buried? Was it a fiery cataclysm, a more gentle abandonment, or intentional process? Did anything survive on the surface?
How large of a surface area was it when it was abandoned? What structures survived being buried?
Who replaced these city-dwellers?
[Repeat question set above again and again until you have finished with each succeeding phase of the city.]

I take a standard piece of graph paper and start to draw out length wise the layers as I answer them. My squares I am going to say are roughly 250 feet a pop to give me room to work with. I would ratchet these upwards and downwards in scale depending how vast of a city I want, but the exact ground scale is really not so important to me as I am mostly just trying to grasp the overall vertical, horizontal, and historical relationships with this process.

The founding layer is naturally the bottom and I place it there. I color code it for use later (blue in this instance) to help distinguish it from the top layers. I am going to go with my perennial “lost civilization” favorites from the Hill Cantons, the Hyperboreans. They constructed a fairly compact city with cycolpean walls. The city was submerged by a vicious sorcerous deluge a 1,000 years into its life leaving the large stone structures encased in a preserving 30-foot thick layer of mud.

The original citadel, high on a bluff, survives and was incorporated into the next phase of the city (I mark it on the right and make sure it is visible in my next layer) by the Latter States who built a broader, yet less grand city on the site. I draw that layer on my side-view with an orange color and no space between the layers due to the relative shallowness of the layer.

Exactly 1000 years later, a Space Elf host burns and ravages the city near completely with only some of the larger, more durable. This city is larger and far grander than the other layers with giant plazas, massive pyramids, aqueducts, etc. Foul serpent women (from the future) encase this city again a 1,000 years into its life in a giant bubble of amber and collapse a mountain over it for good measure. I draw in a thick layer to represent the deep burying.

For mysterious reasons known only to their serpentine minds, the serpent women hollow out a space in the mountain rubble and construct a massive underground ceremonial space (colored in maroon) only to abandon it nearly intact after a millennium. The current human city of Dobre Rajetz rises on this spot.

And this is what I am left with (click to enlarge) a simple but functional cross section--and more importantly a conceptual idea of what structures, flavor, and size each layer of the undercity has (and roughly where they fit in relationship to each other). 

In Part II, I take up exploring how to use my pointcrawl ideas to capture both the sweeping horizontal and vertical dimensions of the space.  


  1. I hope you will take a look at Tony Dowler's "How to Host a Dungeon" game. He addresses many of the same considerations.

    Also, in your second paragraph, I think you want "razed," not "raised." In ditlana, a city is not lifted up, it is torn down and rebuilt.

    1. Oucha, the longer the gap widens between when I stopped being a professional editor and my current lackadaisical state, the sloppier I seem to get.

      Yes, I've played How to Host a Dungeon. It has it's own metric though, one hardwired with its own setting assumptions (and mini-game). I was looking for something more freewheeling and specific to the kind of undercity I want to design.

  2. Thanks Chris - this helps free my mind up a bit. I think I'm always a bit hampered in such endeavours by my archaeology background; realising that most such cases in our own world only have a few feet to show for each layer, as they are all basically smashed.

    1. Where's the fun in that? It may take some finagling with imaginative reasons for why there is so much larger strata between, but I think it's mostly worth it for the dungeoneering feel.

  3. I was working on something like this in my Fenrecz campaign, where the titular city was built on top of old versions of itself, inspired by the Undergrounds of Seattle and Portland, among others.

    I'm curious to see where you go with the next post and I suspect it will be similar to what I came up with, where points of interest in the Warrens were separated by sections of convoluted tunnels/staircases/roads that essentially acted as wilderness areas, with places that needed to be mapped using Zak's neighborhood layout system from Vornheim (which wound up being a little more 2D than was called for; although it served its purpose). What I didn't implement was a system for getting lost.

  4. Some inspiration.. http://www.slideshare.net/jamesfurlo/underground-city-of-derinkuyu
    Seems like (rightfully so) most are stone-cut cities; which are the only structures that would support that much weight on them for centuries.

    1. Those are awesome pics of Derinkuyu. I like all the multilevel balcony/gallery shots. That's exactly the sort of thing I tried to accomplish in my old Caves of Chaos re-work.

  5. "Were they displaced or did they just evolve historically?"

    I'd add for purposes of gaming--were they displaced at all or are they (or their descendants) still living in the undercity?

    A few more questions:

    How easy is the undercity to travel in? How much of it is exposed? How much of it does the average citizen know about? How easy is it to access? How many access points are there? How well preserved is it?

  6. "miles and miles of “empty space”"

    I'm sure this is where you are going with point crawl. But, one way to tackle this (both in how can I possibly start and how can I possibly finish) is to map areas, locations, centerpieces. Linked together via hand waves. Add a set of "random locations" for the long boring stretches and when players go where they shouldn't (i.e. places you haven't mapped yet.

    Dumping "Doing it right", is the first step in Hex/Dungeon/etc crawl.

    1. Well I was imprecise "doing it right" is not meant to be "doing it correctly" (whatever that may be), but "doing it in a interesting and satisfying fashion".

  7. I live near a city which has an undercity that can be explored (and there are guided tours, of course). It does happen that the streets are raised on occasion, but only in particular circumstances (in Seattle, because of frequent flooding).

    This post is timely for me, as the home city of the campaign setting on which I am working is one that has layers of ancient cities underneath. I haven't yet worked out who built the lowest layers, but those look like some intriguing inspirations.

  8. For a minute there I thought you were telling us that space elves and serpent women were part of the deep background of the Hill Cantons...?

    I like what you're doing here but my melodramatist's brain wants to reverse the whole process. Why do you want undercities in your game - what do they do for you? Are they deniable underworlds - sub rosa connections for criminals/grey markets/subconscious desires and urges which the surface can't live with or without? Are they completely secret, honeycombs in the foundation of reality? Are they sealed Tutankhamen's Tombs waiting for the PCs to open? Doors into the dungeon dimensions?
    In every one of these cases I would rather use the Lost writers' method and discover them together with the players as we drill down. About the only time I would want to start at genesis is if the whole city, upper and under, is a giant balkanized Jacquays dungeon.