Monday, December 19, 2022

Dungeon Design, Process and Keys

With my decision to work on Dungeon23 coincides my starting a public Crystal Frontier Campaign and being dissatisfied with the progress I've been making to various new projects. I've got three large, rather experimental dungeons about 1/4 - 1/2 finished (including art and layout), and a smaller one of about 20 rooms 2/3 done, but they've just refused to come together this year. Hopefully at least the small one will be out sometime. So I've had lots of reasons to thinking a bit about how I personally design dungeons and adventures again - not as a theoretical exercise, but because I need to write some new, satisfying dungeons.

Below are some notes on my personal quick technique for getting something together for play at my table and (after a lot of additional polishing) for publication. I hope they can be of help to newer designers thinking about giving Dungeon23 a try (or really just writing up a dungeon to crawl).  As always they are for the classic dungeon crawl style of play: exploration supported by procedural turnkeeping, supply, and randomized risk. They are likely the entirely wrong way to write an adventure that will make really good live action Youtube or help you run a Vampire the Masquerade campaign, I don't pretend to know how to do either of those things. 

I’m a naturalistic, or maybe even ‘organic’ designer.

This has nothing to do with whole grains. As a style of adventure design, what I’m calling organic is an expansion of what's often described as “Gygaxian Naturalism” because Gygax discusses it in “The Campaign” section on pages 86-88 of the 1st edition Dungeon Master’s Guide. While the focus is usually on Gygax's creation of a fantasy ecology, I'd say he goes further and offers an approach to making dungeon adventures that form a logical whole.

It’s also obvious to me that this was Gygax’s personal style of adventure design, meaning naturalistic design was extremely important to the success of Dungeons & Dragons and roleplaying games more generally. Organic design shows through clearly in Gygax’s best work, his most memorable adventures such as Keep on the Borderlands (B2) and Against the Giants (G series). While these adventures might not appeal to everyone today, in praising them it’s important to look at them compared to other design options at the start of the hobby. Arneson’s original Temple of the Frog as included in the Blackmoor supplement for OD&D and Wee Warrior’s Palace of the Vampire Queen by the Kerestans are great counter examples.

The naturalism and coherence of Gygax’s adventures sets them apart, he focused on the dungeon space as interconnected by logical relationships, sometime ecological, but often political or historical. His contemporaries produced dungeons that were far closer to a board game sytle series of encounters, or as the afterthought to a larger scale political and military conflict (which Arneson did design quite naturalistically). The mead hall of Gygax’s giant chief makes sense as a location, structurally, thematically, and as an adventure. Its rooms have clear uses and a sensible layout within the larger fictional space. Its inhabitants relate to each other and have uses for the spaces they inhabit. While Gygax's dungeons are far from realistic, and can become odd at times there’s almost always a thruline of sense and purpose that can only come from the conscious effort to build a dungeon around its inhabitants and a theme. As affirmed in his Dungeon Master’s Guide, Gygax doesn’t want to overthink the logic of his dungeons as a simulation, but to create a plan and structure a location that are comprehensible to and and exploitable by the players. You can poison the giant’s stew in the Steading of the Hill Giant Chief because the kitchen and feast hall exist as part of a sensible and coherent whole—that’s Gygaxian Naturalism (though the term is also used sometimes to describe a group writing process). Jennell Jaquays, tentatively with F'Chelrak's Tomb (1976), and notably in her later adventures Caverns of Thracia and Dark Tower (less effectively), expands on Gygaxian Naturalism quite successfully. Her designs start the practice of layering history and increasing the density of description, the level of interactivity, significance of faction relationships, and spacial complexity of dungeons making them more functional for exploration.

In some circles, Gygax’s other major contribution to adventure design is more popular - procedural generation. Appendices A-H of the Dungeon Master Guide are one of the first (they date back to Strategic Review/Dragon, Issue 1 as a way to play D&D solo) efforts to define this technique - providing a means for a dungeon to build itself through random dice rolls.

I’m not much of a fan. Here’s why.

I’ll acknowledge that procedural generation can be useful as a springboard for ideas or to fill space in a hurry when nothing better is available (such as when your players move off “the map”), but randomly generated rooms are either too vague and disconnected for anything more than board game style play, or require such complex tables that the designer might as well just produce a much larger keyed adventure with the amount of space and ideas. Relying on random stocking almost always means that the details and complexity need to be filled in, and the random design expanded and rationalized. You’ll still end up describing, keying, factionalizing and connecting the parts of your dungeon for it to function well. All procedural generation does is add the step of randomly generating elements of the space that have to be revised to give the dungeon coherence.

These are the basics of what sort of dungeons I want to write. Classic keyed spaces with a high degree of coherence, interactivity, variety, a layered history (useful and discovered via clues in play), navigational puzzles, and both description and themes that go beyond those of typical Gygaxian vernacular fantasy or the contemporary expectations of standard fantasy settings.

In the Dungeon Masters Guide, I take Gygax’s point as the claim that it’s both easier to design and to play an adventure that makes sense on an intuitive level. This is just the old dungeon design practice of rational room placement. The troll kitchen is next to the troll refuse heap and the troll prisoner stockade, not the dragon lair. In a more advanced for however, it’s making decisions about the theme, history and secrets of your dungeon adventure. It’s valuing coherence as the primary goal.

Of course, and this is part of the innovation in Gygax’s dungeon design, the goal is an exciting and comprehensible dungeon adventure, not a simulation. Naturalism only needs to work at the superficial and easily understood level where players can use the information to inform decisions, guess further information or perhaps better overcome obstacles. You don't need to include every bathroom in your adventure (though a chute is a classic dangerous escape rout and toilets are fun to place treasure in), but considering how the monsters get food and water can lead to a more interesting dungeon design.

There's the advantage of creating involving and playable adventures, but this approach to design actually encourages and enables dungeon creation. Designers who think through the past and purpose of fantastical spaces can both more easily and more coherently key them.

Again, one doesn’t need to simulate things or explain all this background, (especially in a published adventure--please don’t!), just describe the immediately visible results of this past and connect them to create an internal logic to the dungeon. Once I work through an idea in my head, imagining its structure, its effect on the world around it, how it came to be and what it is now, it’s easier to map, stock, and key. In a sense an organic approach to location design asks the designer to approach the fantasy location a bit like an archeologist in reverse, deciding the meaning of things within the space, determining how they have transformed and evolved over time, and then only cataloging the outward manifestations of these pasts in your keys. The interpretation is left to the players, and correct interpretations can connect directly to rewards. I understand that archeological discovery and practice is also part of what encouraged Jaquays to write Thracia in the way she did.

So I don't think I really have anything new to offer, and I’ve talked a bit about this approach in the past, and even provided another example of how I used to do this more formally long ago.

I often start my designs very quickly with a few ideas, but given the size and complexity of the mega dungeon that Dungeon23 envisions I’ve thought about the process more in depth.

I go through this process in three broad stages:


The process I use when writing my own work is not this formalized, when I’m designing many of these steps happen concurrently or out of order, mashed together, and without the specificity here. This is especially true of THEME & SCOPE which is extremely informal. I tend to think about the ADVENTURE GOAL or INSPIRATIONAL IMAGE first depending on how the idea gets started. Then I quickly move on through the other early concerns before thinking briefly about DESIGN and start writing with either KEYING or the MAP depending on which makes the most sense. I switch back and forth between stages of the process, editing and polishing. I find that the early concerns tend to inspire and focus the design, while jumping back and forth between map and keys can refine it.

Stop! The absolute first thing is to decide if you are writing is for your personal use or publication.

There’s a huge difference in how we need to design for ourselves, when notes and vague ideas are usually sufficient, and how we need to design to produce a usable and enjoyable product for others. Published adventures must do a sufficient amount of work to justify buying them. There are hundreds of adventures for people to buy, but they bought yours and you owe it to them to make running their game easier. You want to provide something accessible that stirs their imagination and captures something unique while while offering enough content that the referee running the adventure won’t suddenly be stuck having to make all sorts of unexpected improvisations.

At the same time, I think it’s bad design to seek too much designer control as an adventure writer. Engaged referees will adapt and transform even the most complete adventures, so there's no way, and less need, to force your morality or sense of how the adventure should progress on referees and players. At best you can describe your own intent and hopefully make it interesting enough that the referee using it won’t discard everything except the map. Provide every inspiration and convenience you can, while also accepting that it won’t be used by everyone and that you certainly can’t force people to use even the parts you love the best.

Another pre-adventure writing consideration is what system you are writing for and what demands it places on an adventure with it’s mechanics and procedures. This is related to what you want the play experience of your adventure to be about.

For the dungeon crawl style of adventure and the classic D&D based systems I write for system considerations inform design in many ways. A megadungeon as a form is a way to emphasize navigational puzzles (finding optimal routes over multiple sessions), give lots of exploration options and threaten the characters survival with supply depletion. If you are designing for a system that doesn’t support these with mechanics, the dungeon will have to be designed to do something else. Even if it does, and by writing classic dungeon crawls I think I have an advantage with Dungeon23, there are lots of considerations around play that still need to be incorporated into dungeon design. Not everything is immediately obvious even from researching or past successes.

For example, it is common in 2022 to play shorter adventure sessions then old D&D was envisioned for in 1975 or 1980. Rather than 10-12 hour games, a lot of modern games are played online for 2 to 3 hours. This is a material limitation as well as a cultural one. The past three years have taught everyone except school administrators that it’s far harder to stay focused in a video meeting than around a table. I still think it’s worth it to play with distant friends.

It’s not just an interesting evolution because of technology, reduced session length also changes play.

The issue of session length becomes acute if one maintains the “expedition style” of game--the tradition of ending sessions outside the dungeon (a near necessity for higher lethality megadungeon play) so that each session is an expedition and can have new characters and deal with missing players better. It also has other effects. First, party size tends to be smaller, as it’s harder to manage large groups online. Second, the rules for encumbrance and how they interact with supply mechanics (such as torch duration) are designed for longer sessions. To adapt to these changes the dungeon's layout also needs to adapt, with new ideas about: distances, number of entrances, monster strength, and treasure placement/amount. To make a megadungeon for online play one has to depart from the way it used to be done, because people are unlikely to be running the adventure the same as they used to.

These are the collection of early concerns I have thinking about a new adventure. There’s no real order to them, they sort of frame everything else and allow one to get into the actual adventure design, the answers should be quick and don’t need to have much detail -- details and complexity grow organically in this kind of design.

IA. SCOPE & ADVENTURE GOAL: What sort of adventure is it as a game artifact? Lair, dungeon, megadungeon? What levels and what kind of design will I be using? How do I envision this being used either in my own game or by others if published. Part of this is seriously asking yourself the question “What am I getting myself into, do I have the resources for this, will my enthusiasm for this hold out, and what limits can I impose on the scope here to help finish.”

For Dungeon23 this is a big one. It’s a huge project.

IB. INSPIRATIONAL IMAGE(S): Did you have a weird dream, did you have an image in your mind about this place, something arresting that made you want to know more about the space. Usually for me this ends up incorporated as the cover art for the adventure, but that’s not important.

IC. SETTING RELATIONSHIP: How does this space relate to your setting? To any setting? What can you take from the setting (even if that setting is simply “vanilla D&D fantasy”) that helps build more information into the space?

Relatedly, do you need to add any exterior aspect to your dungeon, a haven nearby is the most common, but you could describe the lands around briefly if it's important to your project. The further an adventure departs from existing settings the harder it is to play them without some setting elements built onto them. Likewise hooks and rumors are a useful and related addition, though it's far too soon to think of them now--they come together best when you know what's in the dungeon that might encourage an expedition or be worth a rumor.

ID. CENTRAL CONCEPTS: Not imagery and story so much as a central situation of concept for the location. A consistency that ties all the elements of the location together and explains why any distinctions and departures from exist. A few core concepts are the basis of the theme and help inform DESIGN questions while it evolves with the answers to them.

Again this is one to highlight for Dungeon23 as with a 300 room dungeon there should be a lot of distinct nodes, spaces, themes, environments and variety - both to keep the designer interested in finishing and to keep the players interested in exploring. The mines of Moria are great inspiration, but no one wants to actually play 300 keyed rooms that all amount to “goblin infested mine” (though I suspect a lot of people might claim they do).

With a theme and an understanding of scope, take a few more minutes to think about the adventure’s design--how you expect it to play out, and how you can use the theme to build a more compelling adventure. This is absolutely not a “story” in the sense contemporary traditional design uses the term, this is a way to check the situations, treasures, inhabitants and map of your adventure to make sure it’s playable. Generally I think of a few questions about the location:

IIA. WHAT’S VALUABLE: Why would you brave such a horrible place? If the goal of play is to make your character more powerful by finding magic items and gaining levels by finding treasure--what treasure is in the location? How did it get there?

On the grand scale of a megadungeon it's likely best if there's something far more than mere wealth in there. If it someone is expected to run a dungeon for two years of sessions, it better have something more impressive then some gems and a pile of silver cutlery hidden away in its depths: immortality, a god killing sword, something to save the world or transform it. Most parties and campaigns will never reach it, but not only might this offer a satisfying ending to the campaign, but it’s a reason to keep exploring, and something to imagine.

IIB. WHAT’S IS IT DANGEROUS: If a location is filled with valuables, especially if it’s widely known to be, why hasn’t it been looted yet? What’s happened to the attempts to plunder it before, and what protects (intentionally or not) the treasure? Is it just creatures, or are there special conditions, traps, and puzzles that have made it impossible for past expeditions to succeed?

IIC. WHAT ELSE IS THERE: Related to the questions above, but not entirely the same - not every inhabitant of a location needs to be dangerous, and even those that are raise specific questions. For large dungeons with lots of varied inhabitants and many factions these considerations take on more complexity and importance. It’s still good to have mysterious departures occasionally, but for the most part naturalism creates a baseline that makes those departures interesting.

So what else is in your dungeon that’s not dangerous? Why are/is they/it there? What are they doing? How do they survive in the location and how will they react to outsiders showing up? This is important both with prewritten monsters (why might something with a known ecology be in the specific location) or new ones (how do you emphasize the new creatures and give players clues to understand them in the adventure)?

Similarly, not everything in the dungeon is dangerous or valuable, but it should all mean something (it doesn’t have to be complex) so that it ties into the rest of the adventure in some way. Valueless but interesting discoveries, environment and decoration--dungeon dressing--work not just as flavor and to create a sense of an actual space, but help make player decisions more interesting as they offer something to interact with that may not offer danger or treasure, but still take time (and thus create supply loss and risk of random encounters) to investigate. Dungeon dressing makes player decisions about what to investigate more compelling because of the time cost and associated risk involved. If everything is important then there’s never a question of what to search and what to ignore and the players never have to make deductions about the location or its denizen’s habits. As part of this remember to dressing contain clues and secrets that are useful for solving puzzles and determining what’s important and what’s not - e.g. secret doors with latches hidden behind a repeating symbol.

IID. FACTION RELATIONSHIPS: It’s better to have more than one creature or group of creatures in an adventure because it creates potential space for conflict and interaction. This means the designer needs to know how these factions relate to each other, because it will be essential information for the players. How are the inhabitants aligned with each other? What is their relationship to the overworld? What do they want, what are they afraid of, and how can the party ally with them or make them enemies?

IIE. MAP: I put drawing the map here, but it’s something you can do first, last, or whenever the location is sufficiently solid in your mind to make one. A lot depends on the size of the adventure, because the bigger it is, the more keys, the more it will change as you write it and the more likely you’ll need to redraw the map. This is why I never start with much more of a sketch, and this is also why I don’t use other people’s maps.

Maps aren’t just a convenient way to string together encounters and obstacles, they’re an essential part of classic dungeon design, offering the primary puzzle that the players need to unravel to succeed--at least in larger dungeons. This means maps have their own set of fairly complex considerations. I might get to my mapping techniques and ideas at another time...when I finally get around to drawing some for Dungeon23 perhaps?

The most significant and satisfying part of a dungeon is interactivity, and this means details. A dungeon crawl may not be possible in five or six rooms, as no overall navigation puzzle or supply pressure can really exist, but even in small lair adventures the problem solving and investigation of interactive elements of a classic design are available. A dungeon that has contents to mess with, puzzles, inhabitants and scenery, that encourages players to visualize and act in the fantasy space will provide more fun at the table then something that is bare and closed off except for encounters.

Much of interactivity comes down to keying. While keying is like other writing, a stylistic choice, there are also basic formal elements to it--like the infamous “five paragraph essay” or like some of the more rigid forms of poetry. Dungeon keys have a set of rules that work to limit the author and also produce something immediately recognizable as a key. Adventure keys are first functional writing, meant to be used, not just enjoyed as an expression of wordplay or vibrant poetics. Also the amount of space in keys is limited. No one wants to read a three page key to a single room, written in the style of Undermountian or a 1993 Dungeon Magazine adventure--or at least no one wants to run a game with it.

III A: INDIVIDUAL KEY THEORY: There are a lot of maxims about individual keying, and OSR in its later period became a bit obsessed with creating sets of rules for writing keys. While some of these specifics (which I’ve listed below) are extremely useful, nothing is absolute. Keying is still writing and so a sort of art. Yes it’s functional writing, conveying needed information in an efficient manner and constraints help that. But keying is also poetic writing—description that conveys images and ideas in ways aimed at helping the reader to examine and visualize them.

This second, creative aspect of keying is especially important in dungeon spaces that depart from the expectations of vernacular fantasy or from the backgrounds and material given in rulebooks. Still, even within the confines of the expected and traditional setting of fantasy adventures, seemingly cliched spaces like a mushroom forest require a more poetic and longer key then the basic dungeon location of a dank stone-block cellar. The elements that require the most space in a key however are interactive items: triggers for a secret door concealed in the grout or special mushrooms that make the eater grow smaller. This variability based on both setting and key uniqueness and complexity, as well as personal style, skill, and adventure type all play a part in deciding how to key.

Currently there’s a revival of minimalism among dungeon key authors, which to some degree is a good thing. Some minimalists like bullet points and outline style structures, others want one sentence of references to rulebooks. I’m fairly ambivalent to both, though either can work for the right sort of adventure. The essential and valuable aspect of keying minimalism worth remembering (though not worth the orthodoxy or nastiness some promote it with) is that keys can’t be too long. It's been tried  and they really don't work.

In megadungeons this is even more important as the adventure is so large. Saving space in any book, saving yourself from excessive work, and letting the reader understand a few rooms at a time (nodal dungeon design is excellent for megadungeon) by keeping them each fairly simple are all worthwhile.

I write long keys. Too long. Part of the problem is that I naturally write a lot and write densely (something that RPGs allow me to indulge, unlike my professional writing), another part of it is likely that I started writing up locations in the 90’s when the keys in most RPGs were huge. I’ve mostly managed to curtail this tendency over the past several years of designing adventures, and learned that keys have real limits.

The person running the adventure has to be able to hold the images and essentials of the room in their mind along with all the rest of the ideas they need to run the game … and that space is limited. Keys longer than five or six sentences can work, with a good editor, and an attention to ordering their contents, but that’s a good basic limit. Longer keys are even needed at times for complex puzzles and npcs, but basic description shouldn’t really get beyond a paragraph.

Five sentences isn’t minimal by most keying standards, but overwriting isn’t the only way to waste space and referee attention. One thing thing I also aim to avoid from experience is formalism in dungeon keys. This can be as simple as always starting a key with a list of exits and dimensions, but more elaborate efforts also appear sometimes. The most only repeating structure you need is a room name (I also add a note about the lighting afterwards as a way to promote caring about light sources … but even that might be going to far) and a key number.

1) Number Your Keys!
Use letters for your subareas (you have more then an alphabet of locations sometimes). It's odd, but I see adventure sometimes that don't discuss rooms individually, and fail to split them up into keyed areas, describing a location as if one were walking through it in a series of paragraphs.  No matter how literary or immersive (largely a meaningless word in adventure design) this might feel, don't do it. Numbered keys provide order and make intuitive sense when coupled with a map.

Also note your key numbers on the map and check they're in the right place and all there. Watch them when you move things around. I'd add that I think it’s pretty reasonable to break up your numbering system by level, especially with large dungeons, but I suppose that’s personal choice.

2) Listing Dimensions and Exits is a Waste!
Maps! Use them.
Avoiding this unnecessary info dump is part of what a map is for. Every line in your keys matter, and not just for physical space, so if you use a line on information that can easily and better be somewhere else, that’s a line you don’t have for cool stuff in that room.

This lack of spatial information is one thing I don’t like about those spreadsheet pseudo-maps with the connected names and boxes. The other being that a good map also helps communicate all sorts of things about the feel and look of a dungeon—starting with room shape. I don’t mean to be overly harsh, and this is a collection of my personal hobby shibboleths, but if you aren’t drawing actual maps you are missing out. Likewise, if you aren’t or aren’t including dimensions on them (that’s what the squares are), and you want to run an actual dungeon crawl you will need to waste a line or two in each key listing these things! Now player mapping is less popular these days, as it’s especially hard to manage online (and it’s easy to use a VTT or otherwise display a map), but it’s still an aspiration for many, and having a solid referee map to refer to when describing things to player mappers is a great help.

Please, save space in your keys, but even if you don’t believe me and have been seduced by the exactitude and pleasing repetition of listing dimensions and exits it’s worth the trouble of drawing a map. This is even more important if you want others to use your adventure.

However, listing exits can be helpful if things are confusing, the entrance is itself important, or it holds some evocative detail like a taxidermy crow nailed to a door. I mentioned before that there’s no absolutes in these rules. I also suspect there’s some odd specifics with Dungeon23 that make including dimensions and exits more helpful - a lot of people may not have a map when they write the room and megadungeons are huge connected spaces and Dungeon23 is a stretched out process, so it may help you to have notes about what room connects to what other rooms. They are best as notes however, something to set aside in a final version or maybe include in a referee aid.

3) Start with the Most Important Thing!
This one shouldn’t be hard, but it’s the one I see done badly most often. Because the key is a utilitarian piece of writing it has work to do, and the most important bit of work is to make sure that the dangerous and obvious are noticed by the referee. The players should also have a chance to spot the obviously dangerous first, followed by the merely obvious.

I start with the room's inhabitants. They’re going to need to be addressed in some way before the adventurers can get on with poking around looking for treasure. Second, I tend to include obvious environmental dangers (chasms, lava, etc) and visible valuables. After that, or if these things are lacking, I note the rooms general characteristics. Hidden things such as traps tend to go in the same order with clear notes that they are hidden. Phrases like “concealed” or “waiting in ambush” are great, but because of how important these hidden dangers are, I tend to write “(trap)”or “(hidden)” after them to make them jump out at the referee.

4) Plain Text is Better than Boxed Text!
Controversial, I know, always has been. I dislike box text as it makes various other keying sins easier and increases length and size of keys.

Boxed text evolved as an attempt to control how an adventure is run for tournament adventures. You aren’t writing a tournament adventure … or writing a screenplay. It’s a trite insult but too often boxed text really does become a space intensive excuse for bad fiction writing. More than risking useless content though, boxed text is paradoxically contrary to usability.

Book text assumes the designer can speak directly to the players, and that removes the referee from the process of describing the spaces in the adventure.

Given how central description, questions and answers are to the basic play loop of RPGs, and how important referee decision making and adjustment of the setting is to player action, by sidelining the referee boxed text risks depreciating one big advantage RPGs have over computer RPGs. More practically, it discourages the referee from reading the adventure and thus understanding what’s in the room before playing it. Boxed text makes it seem like the designer has done all the work, and the referee can just jump from text box to text box, referring to the details and mechanics underneath only as they become necessary. But, since box text is player facing, the referee doesn’t learn about the important hidden aspects of the room until they are already in play. They aren’t prepared to run those ambushing goblins or the gas trap, and might omit them entirely by accident.

Finally box text can be wordy and confusing, it’s far better if the referee uses their own words and understanding of the fantasy space because referees know their table and what they will care about or understand. Plus, if the referee gets the location description wrong or is confused at least the whole table will be on the same ground, and it’s far easier to move forward from there.

5) Don’t Let Things Get Muddled!
When writing rooms that don’t have any complex interactive elements like monsters, traps, or interesting dungeon dressing, it should be easy to keep the key to a few descriptive lines. Obviously the more notable and interactive elements in a key the longer and more complex it will be, and it should be equally obvious that it will be harder to use.

In order to make those keys work you need to keep things organized.

It can be done simply by structuring paragraphs, but that can be fairly tricky, especially if the content involved requires more description, such as when one isn’t describing the expected fantasy dungeon environment. Additional structure can make things easier for the referee to use the key as well as potentially helping longer descriptions retain functionality.

The currently popular way to add structure is through bullet points with a list of descriptors or brief summary of a specific tags that are bolded in the initial paragraph. The bullet points below are usually just a list of descriptive words. Like pure paragraphs, bullet points also have their shortfalls for novel environments. I find that their pretense of minimalism is often insufficient for complex dungeon elements, and it tends not to provide very evocative details.

An additional risk of bullet points is that the short lists of adjectives or other descriptors they most often contain are quick to write and read, so it doesn’t seem like much of an issue to add a lot of them to a key. This means it's easy to go too far with bullet points, because each bullet point is a separate idea that the referee needs to keep in their mind, and too much bolding in a key makes it hard to know what’s important about the bold words.

My personal style involves using a small number of bolded words; but only 3 or 4 bolded words maximum for a complex room that takes up a page or so - and rooms shouldn’t take up multiple pages unless there's tables and sidebars involved.  Each bolded word refers to a follow up paragraph that explains it and offers more detailed description. This form both provides some structure and still allows full sentences that produce more involved description.

I also use tags (Bold Italic usually, but italic is hard to read and should be avoided generally - an excessive variety of tags and references are also hard to keep track of.) that indicate either sidebars or a reference to an appendix. I use this system for monster stats (and you really should include your stats with the key if you are writing a system that doesn’t require full page monster blocks), complex faction relationships, orders of battle for more organized groups, and magic items that have longer descriptions.

Key structure matters, and while formalism or a matrix of room elements is often counterproductive, your keys should be written in a way that allows use and presents important information so that it’s easy to find. If you having something too large for a key ask yourself if it's necessary, and if it is figure out some place else to put it. This aspect of key writing is where one has to stifle a poetic soul and produce something functional, concise, clear, and quick to read.

6) Expand With Important and Evocative Details!
Now some more lyrical, poetic and creative writing.

The sorts of designer advice usually offered by people who think a dungeon crawl should be endlessly artless hikes through rooms rolled up from Gygax’s solo play lists often insist that adventure writing should be purely functional. This is bad advice.

There are already far too many adventures about orcs in holes and dragon’s lairs cribbed from the Hobbit, plus it would be utter drudgery to write 300 keys of abandoned Dwarfhold using random encounter tables from AD&D. Unless someone is paying you well, there's never a reason for these things. There’s not even anything authentic or properly old school about it, unless that means design dungeons like a very serious ten year old in 1981. Keying advice has long pointed out that descriptions can and should include detail, and evocative elements beyond the names of gems rolled up randomly.  For example in at least one early edition there is advice to remember that senses other than sight exist: smells, sounds, and sensation etc.

This old advice to include other senses is still good, but following it too far leads towards the equally bad habit trying to include everything about a space, and that’s both impossible and ends in the sort of bloated and unusable keys that fill the pages of 90’s Dungeon Magazine. A 350 key dungeon could easily be 700 pages written this way, and the world only needs one Ruins of Undermountain (which is much shorter.)

Instead the question becomes what to include, how to pick and choose where to expand and add creative or poetic flourishes? It’s a pretty subjective question, but focusing on the goal of giving a few intriguing details that help the referee (or the players they are related to) visualize the space has worked for me.

here are some areas that I think don’t help:

a. Don’t generalize: Describe specifics. An idol shouldn’t be described as “evil” or “forbidding”, rather it has a “toothy glare” or is “sharp edged obsidian”. If you find yourself generalizing, consider that maybe you haven’t visualized the space or object sufficiently yourself. Details work when they offer something for the mind to stick to. This is basically the definition of “evocative” - that the words and phrases used have associations and creates images in the mind. Everyone says to write evocatively, so do it.

b. Don’t describe the mundane: Rooms are full of stuff. Look around the room you are in right now and imagine you wanted to describe it to someone. You don’t minutely catalog everything, you say “it’s a home office” or “it’s a coffee shop”. People can fill in most of the items in these sorts of mundane rooms and if they get it wrong, so what - as a referee they will still have an image of a coffee shop or office they can pull from to fill in details and anything else significant is in the key.

As a designer though you often want a bit more than the mundane you want to indicate things that might be useful, valuable, or give a bit of extra description to set the office or coffee shop apart. Say you want to describe a bland corporate coffee shop rather than a good one? The walls are covered in worthless art: abstract color blocks in tones from the last decade printed on canvas. The counters are worn black veneer, peeling away at the corners from the chipboard beneath, and every surface has a little piece of stained promotional signage for some kind of reward card or sugary drink.

This isn’t a perfect description, it's only two sentences but it can be shortened, and maybe there’s something else that better captures the bland corporate coffee shop - perhaps the detail that “the profane icon of a bifurcated mermaid leers from behind the counter”.

Whatever the flaws, a few details the get the reader started imagining a location is more efficient and useful then taking up a page how the shop serves coffee, contains 278 white porcelain mugs worth three cents each, how the handle to the bathroom is loose but the latch sticks, how students come here at night and read chemistry texts, or that it was once a tire store in the 1950’s owned by a kindly man with a big beard. It's also a lot quicker and more fun to write..

7) Limit Your Description To The Room As It Is!
When some folks start writing adventure keys they get confused about what to put in them. There’s so many things one could say about a fantasy space: history, detailed appearance of everything in it, what various factions and creatures use it for, or how they feel about it. All theoretically could have a place in an adventure, and writing location based adventures is a bit like writing a story about a place. Except it shouldn’t be.

The note above discusses this when it comes to providing details, but there are also entire categories of content that just aren’t helpful to writing a location. They can be summed up as information about what the location isn’t.

a. Don’t describe the room’s past. No one will ever know what a room was for unless this is somehow reflected in its current state. Yes, the barbarians have stabled their horses in the burnt church, but players can learn that from the soot stained religious fresco on the walls, the shattered mosaics of saints beneath the straw, and the ornate buttresses outside. Describing how and when the desecration happened, or the fate of the old priest are entirely unnecessary and unhelpful.

b. Don’t describe the room’s future. Yes, some NPC might have plans for the room or something may happen to it if specific events occur, but the key is very rarely the place to describe that. Again it’s not useful information for the players, and it’s not something the referee needs to know unless the room actually changes during play (and then you effectively need a second key for the area) or redecorating plans are essential to interacting with an NPC, and even then most details aren’t necessary and those that are should be included with the NPC, not the room key.

c. Don’t describe what might be happening. Sometimes when visualizing a location designers imagine how the adventurers will interact with it. This is not worth including in the room description. Yes, monster tactics or a few solutions to a trap (e.g. “the guillotine blade can be jammed in place with a pole or spear”) are useful, but elaborate descriptions of what might be aren’t for the designer to write, they are for the referee and the players to discover in game.

Anytime you find yourself writing something in a way that would belong in a story or screenplay you are likely writing a dysfunctional key. Describe the space as the players find it. Briefly note plans its inhabitants have there, or how its fixtures work. Everything else is someone else’s job, and that’s good because you need to save space.

8) Include Motivations and Personality for NPCs and Monsters!
One of the aspects of the classic Dungeon Crawl is that monsters aren’t only for fighting. Combat will happen, but it’s not usually or exclusively a great solution to the problem of recovering treasure from the underworld. Instead players often want to talk to foes or otherwise interact with them without fighting. The reaction and morale systems support this, and so should your keys.

This does not mean every goblin needs a name, let alone a backstory, but given the party is fairly likely to ally with the sneaky bastards it’s worth knowing who their leader is, what they are like, what goals they have and such. Obviously you can easily go too deep, and it’s a lot of fun to think up goblin schemes or history, but if that's need the referee can figure it out (you get one scheme as a treat if it's good and not entirely obvious). For the designer a monster needs some basic goals, description, faction relationship, and personality. Keep the rest to yourself for playtesting the adventure.

Note: As a designer you likely have a lot of extra information in you head or notes about a location, but because you know how everything works even without the written adventure, it’s essential that for a published project of any scale you get some other folks to playtest. It really helps find what’s confusing, what plays well and what doesn’t.

9) Don’t Assume How the Party Got There!
You don’t know how the party entered the space or when. Keying like you do can lead to confusion.

You can’t assume that the party enters through the trapped door (even if the trapped door is the only logical entrance--maybe they teleported, maybe they pick axed down a wall because they do things Die Hard style). Likewise, the designer doesn’t get to assume what precautions the party has taken--they might send someone in to scout, or use all sorts of magic and shenanigans to avoid potential risks. In dungeon keying there’s no “The door creaks open on the scream of rusty hinges to reveal the Garderobe of the Damned!” only “The rusted hinges of the door will scream if they aren’t oiled, alerting the Feculent Horrors within the Garderobe of the Damned and granting them +1 Initiative.”. Even more so, don’t assume that an ambush or other situation happens in a certain way when players enter the room.

Describing a room from a specific point of view is a habit I’m sympathetic to. There’s a natural tendency to write locations as we visualize them that I’m sympathetic to. It's even good practice for designing dungeon spaces to imagine a person entering the space and then pick out the details they might focus on, the whole experience feels natural. At least it happens to me when I’m imagining locations … “whoa trip out, I’m actually in a cave of chaos right now…” Don’t. It's great for imagining, but not for writing a key because having a fixed viewpoint can lead to complexity and trouble writing a usable space. It's even more important when rooms have multiple entrances, and the majority should.

10) Don’t Tell the Characters How to Feel!
You don’t get to tell the players how their characters feel emotionally (barring creepy magic) or even physically. You are not their supervisior.

It’s not hard to do this, just avoid first person writing and description of what happens once the party enters the space. That’s a literary device, a way to make a story more accessible, but again as an adventure designer you aren’t telling a story to a passive audience, especially not in a location key. In key writing it’s a waste of time, instead describe the location not the adventure that will happen there.

It’s a waste first because there’s so many varieties of adventurer. Do robots experience heat and cold the same as ice barbarians or lizardfolk? Describing emotional impact is similar, the referee, let alone the designer, doesn’t get to control the internal mental state of the characters. That’s a job for the players.

Intruding on the players’ part of the game is risky, because a part of establishing the trust in the referee’s ability to be fair that is necessary to run the high lethality, open ended problem solving style of classic dungeon crawls is limiting referee control to the setting. The referee and the designer get to run the whole world and make ad hoc decisions about how events unfold--but in exchange the players are the ones who get to act in that world and to tell their characters’ stories.

So that’s ten keying tips. I should have written them as a listicle and made them all things you should do, rather then things not to do, but I will never be in advertising. I also know there are more, but ten is such a nice round number.

The last thing I like to add to an adventure is information about what might happen after it’s done. Perhaps this is unnecessary, but in my own games players actions in the dungeon have consequences in the larger world and it seems worthwhile addressing them. What happens if the characters rescue someone important? What if they free an ancient evil? I think of the last story in Beowulf where not only does Beowulf have to battle a horrific dragon to mutual destruction because some random fellow stole a golden cup from it, but the poem ends with a brief note on how Beowulf’s people are doomed now with him gone, because their neighbors will jump them. I aim for bits like this, if nothing else they remind referees that thier can be changes to the setting based on the characters actions in the dungeon.

Something to keep in mind when writing up an adventure is if there are additions that will do the job better than keying. Art can often describe a monster, complex space, or NPC better than description, and maps, diagrams, tables, handouts or tracking sheets all can make an adventure easier to run and play if used right. The question is of course how to use them right.

As an example, random tables are great, part of long tradition to use a special table to fill in what a party might find searching through a large number of similar rooms or containers for treasure: sixty poor huts in an abandoned village, the cocooned victims of a cavern full of giant spiders, or the coffins in a necropolis. Having 10 results for sixty empty hovels (as long as you structure the table well so the results don’t feel too repetitive--most will be things like “wrecked stools, scattered straw bedding and shattered crockery”) is a great way to save space, referee energy, and writing time.

As useful as random tables are for sketching in repeated character experiences, the word is "repeated". Random tables are fetishized (a lot of things in post-OSR spaces have an element of fetishization - beware taking maxims about how to play literally) and this leads to some weird uses. Even in generally well regarded adventures I’ve started to see tables used to determine the result of a single experience or a very small set of them. You shouldn’t use a table with ten entries (or three) to determine the contents of three sarcophagi, let alone one. This is creating a random table for the sake of random table, a sort of cultish implementation for the sake of dogma, and a profound waste of space as well as of referee time and energy (they have to roll on the table for starters).

So referee tools are great, and often can save the designer space and the referee running the adventure time. They can explain things visually or fill in large repetitive spaces, give overviews of how conflict with a faction or a timeline in the dungeon will work, but…they aren’t always necessary or helpful. Again formalistic applications of design ideas or structures doesn't work that well.

That’s largely the “rules” and process I use for designing location based adventures. Embarking on something the size of Dungeon23 will probably end with changes and compromises. I suspect that the need for structure, referee tool and supporting additions to the keys grows dramatically, exponentially as dungeon size grows.

As I try to finish the project I’ll likely have more thoughts. Contradictory and horrible thoughts--the Void Wyrm will do that if nothing else. I’ve put thought into the ones here for quite some time and stand by them as advice, but this is a hobby and a creative one. Please take these ideas as they are intended, an outline of my process and techniques, not something that demands loyalty--maybe you’ll figure out better ways for you, especially as you get some practice in. Hopefully though they can help avoid some of the more obvious mistakes in writing location based exploration adventures.


  1. Very helpful, very informative, and a great support for those doing the Dungeon 23 thing. Especially for those of us who are either out of practice, or are pretty new to this whole thing.


Old Games

Let’s talk about old tabletop roleplaying games - specifically the kind of games played in the 1980’s and recently depicted in the nostalgia...