At least, I think this is my third attempt.

Over the weekend, I tried to quantify my previous failures to cookiemonger, and I think I had better start there.

When I set out to design a system for dungeon generation that worked like character creation, I started with the idea that “what might be found” in the dungeon would be the most important thing to determine.

So I had stuff like “Secrets,” “Denizens,” “Traps,” and “Treasure.” All the different reasons that a character might brave the depths. It was a goal-oriented system of attributes. It drew heavily from video games.

But it didn’t tell me what the dungeon looked like.

My second attempt at designing a dungeon generator, was from a layout perspective. I concerned myself with encounters and obstacles, plus any number of other things that would guard the contents of the dungeon.

This second method drew a lot from classic video games from the NES era.

Zelda and Castlevania, specifically. (Two games I haven’t actually played.)

Upon reviewing some other dungeon generation procedures, I realized that what I was missing couldn’t necessarily be found in the contents of the dungeon. More, it had to do with the way in which the dungeon was played.

This is what I figured.

There are different kinds of dungeons, and these kinds are based on their contents — but it’s less about what’s in the dungeon, and more about what happens when you interact with its contents.

Because that’s where you get variety.

I was thinking about dungeons in terms almost directly related to video game level design. I’m looking at games like Mario Brothers 3 and Crash Bandicoot, complete with their overworlds and mission hubs, and tons of levels.

Dungeons, dungeons, everywhere!

For that, I needed dungeon attributes that helped me determine how player characters would need to approach the dungeon — scores that clued me into the most efficient method of encounter resolution and treasure extraction.

There’s plenty of room for variation, but I wanted themes.

First, I figured there are the basics:
– Monster lair / infestation
– Death trap
– Labyrinth / maze

Compared to a neutral, “average” dungeon (which doesn’t exist except in theory), each of these types is specialized. Some dungeons have more monsters. Some have more traps. Others have more pits, passages, and locked doors.

I figured from the start, that I would need five.

Why five? Well, so that it’s statistically possible for players to encounter a number of similar dungeons or different dungeons, and draw expectations from the dungeons they encounter. That is an admittedly boring answer.

Seriously, though? I needed enough dungeon types for variation, but not so many players might be overwhelmed. Because if you have too many types of dungeons, no own can create a rational set of expectations. Not players, not GMs.

So I looked for two more.

Bear in mind that we haven’t touched on dungeon purpose yet. That’s actually separate from the “dungeon approach” that we’re looking at here. A dungeon purpose helps define other very important stuff, to be discussed later.

The next type I figured, represented an Existential Threat.

This is a dungeon with a time limit — whether it’s a sinking Atlantis, an extinct volcano that’s about to erupt, or simply a rival adventuring group that’s racing you to the treasure. (In the latter case, it’s the existence of treasure.)

You have to approach the dungeon very differently here.

Now, the last one eluded me for some time. I tried to figure out if I even needed a fifth one, and I kind of figured I did — then when I thought I had one figured out, I tried to talk myself out of it. It’s based on the Stealth Mission.

Sometimes, your dungeon isn’t booby-trapped . . . it has alarms and armed guards. The point is to get in and get out without alerting the denizens at all. This is your Thief, your Assassin’s Creed. Getting caught complicates matters.

It’s a legitimate approach to a dungeon. And it’s different from the rest.

So, I’m dropping “layout” and “contents” for the moment, since neither one of these things is helpful for determining how or why the players will enter the dungeon. This set of attributes should help with that.

Sometimes, you just have to discover a bunch of wrong ways of doing a thing before you can find the one right way of doing it.

Hopefully, this will be that right way of doing things.