Friday, December 27, 2013

The Hybrid Dungeon

This is the result of a long (looong) series of emails with one of my friends about dungeon design.

He makes the argument that having a linear (or at least, minimally branching) dungeon allows for fewer arbitrary decisions and more GM control over how the encounters develop.

He's got a point.

Unscripted encounters sometimes suck because they are so random, and scripted encounters can be pretty epic when done right.

I mean, we're all familiar with the maze-like dungeon that has a set-piece lich encounter at the end that always unfolds the same way. We use a linear presentation for liches and epic shit in dungeons even though we use random encounters throughout the rest of the dungeon, because we always want the lich to be epic.

It's a conservative choice, though.  The scripted lich is usually pretty cool, but the most epic encounters I've ever had have resulted from the collision of random tables.  But I guess DMs don't want to risk their final encounter relying too much on a random encounter.

Anyway, I wrote this.  I expect it to be at least a little bit controversial, and that's fair.  I'm still rolling over some of these ideas, and I have yet to put a lot of them into practice.

How Organic-Random-Variable Should a Dungeon Experience Be?

This relies on the assumption that organic, random, and variable are all the same thing.  Organic dungeons lean on simulationist ideas of construction (instead of gamist), and as such, tend to be laid out with lots of many ways to enter, exit, and travel through them, because forts often have many ways to enter and have few linear parts.  Many loops.  Many ways to enter the dungeon.  Because there are so many ways to move through the dungeon, there will also be a lot of variability for how players experience the dungeon.  Some groups will have a much easier time than others, simply by getting lucky in choosing which way to go.  So organic dungeons are correlated with more random experiences, which are inherently more variable.

Purely organic dungeons can suck for a variety of reasons.  They can be confusing.  They can be too easy.  They can be too hard.  They can be difficult to balance.  

But these are all symptoms of bad organic dungeons. Because they tend to make more internal sense within the game world, players can use more common sense when navigating them, instead of relying more on DM explanation. And they are not random. Players that think before charging in will have a better chance of finding pathways more amenable to their goals. Excessive branching can be cut down, and extraneous branches can be filled with easy encounters or non-damaging encounters, so that players aren't unfairly depleted when their reach the boss. Keys and chokepoints can be used to cut down the variability of a dungeon path, so that the "easiest" and "hardest" paths are comparable.

Organic dungeons have many advantages, too.  Because they are based more on simulationist concerns and less on gamist concerns, they react better to unforeseen stresses.  Like when the players decide to siege, flood, infiltrate, or sabotage the dungeon and its denizens.  They are also able to react more appropriately when players attempt unusual strategies that the designer did not foresee.  By including organic considerations like where the dungeon inhabitants eat, sleep, and excrete, the players have more options to formulate effective outside-the-box plans.  (Attempts to poison their water supply always fail if there is no water supply, for example.)

But there is no reason why a dungeon should be hobbled to a commitment of being purely organic.  Consider a hybrid dungeon, that contains both scripted and organic elements.

The hybrid dungeon has a well designed set of paths through it.  These are the most obvious ways to navigate the dungeon to any goals within it.  These are balanced, fair, and fun.  These are the orthodox pathways, and they should be readily available to the players who attempt them.  This should be a complete dungeon, and it should be fun and awesome, and not lacking in any way.  A player should be able to play through the orthodox dungeon happily and successfully.  It will have branches, but these will be designed so that either (a) the player usually knows when they are choosing a side passage instead of the direct route to their goal, or (b) short enough and not so brutal that the dungeon will become unwinnable if the player chooses all the non-direct passages.

Around this orthodox core, we should build the shadow dungeon.  These are the things that are not obvious to players.  In fact, many of the players will not discover or even care about some parts of the dungeon.  These are the secret passages and subtle shit.  These are there to reward players who look for lateral solutions to things, who would rather interrogate than fight directly, or would rather step off the well-scripted core of the orthodox dungeon and get into a more chaotic dungeon.  The shadow dungeon will usually make the dungeon a little easier, but not always.  If a player spends time and resources finding a way to climb onto the roof, they should (usually) be compensated for their good sense and hard work.  The important thing is that the shadow dungeon is neither necessary nor obvious.  In a way, it's not part of the dungeon at all.  It's anticipating the players will want to hack the game, twist the rules, and look for an easier solution.  The shadow dungeon is there because the designer anticipates the the players will attempt to hack the dungeon, and has considered them when designing the dungeon.  In this way, the shadow dungeon is a dungeon of contingencies.

Remember that the shadow dungeon is not always physical rooms the way the orthodox dungeon it.  The shadow dungeon is putting in considerations for what happens if the party allows themselves to be taken prisoner, or decides to look for a map in town, or decides to parachute onto the roof.  It is simply anticipating that players will try to hack your dungeon, and planning ahead in a way where good ideas are (usually) rewarded.

The shadow dungeon is necessary because some groups will always look to jump off the tracks that you have laid for them. If they spend time and effort looking for a way to circumvent the orthodox blockade you've put in front of them, they should find one.  Yes, this does mean that it takes more work to design a dungeon.  But certain players will look for shortcuts and hacks, and it is better for the designer to plan for these things than for the DM to scramble to improvise when the unexpected happens.  Remember that not all parts of the shadow dungeon are easier than the orthodox equivalents. They might be much more deadly, but that is the risk you take when you decide to do a gamble such as allowing yourself to be taken prisoner.

And lastly, around the orthodox dungeon and the shadow dungeon, we place an organic wrapper.  These are the toilets, grain stores, and the ecologies that allow dungeons to function when players stress the dungeon in ways that no one could have predicted.  Hopefully.


  1. While I appreciate the attempt to offer the best of both worlds, I think the big sticking point for me here is the assumption that a dungeon must have a goal / boss / scripted endgame. The main unifying feature of LotFP modules is that they don't have a true boss, or sometimes even significant monster presence; they're just funhouses/deathtraps to play around in. The whole point of a megadungeon, even with a boss monster somewhere in its depths, is all the tinkering around the party does over time. Building a dungeon around an endpoint means that you've decided in advance what the party should do, which can be problematic.

    Example: I once stocked, tweaked and ran the Crypt of Luan Phien. The party, wandering around in an attempt to catch their bearings, ran into Luan Phien, decided they wanted none of that, and quickly left the room (to his amusement). This is specifically not a problem, and didn't render the whole dungeon pointless, because defeating him wasn't a requirement -- they could have fought with, negotiated with, or avoided him entirely and the fact that they had an adventure in an interesting environment would have remained unchanged.

    In brief: don't worry about trying to "hybridize" dungeons. Different dungeons will have different purposes and different design and stocking specs anyway; the only requirement is that they're engaging in play.

    1. You're 100% correct. I should have added a disclaimer at the top that we were talking specifically about dungeons that aren't just dungeons for their own sake, but rather pieces in a larger plot arc.

      Funhouse dungeons and megadungeons are fun in their own right. We were only trying to drop some theory on goal-oriented dungeons.