Dungeon Design Note: Defining Interactivity

A lot of dungeon design advice focuses on “interactivity”, and to a degree this word can be a meaningless substitute for “stuff I personally like”. Yet one should strive to write locations with “interactive elements”. In the context of location based (or dungeon crawl) adventure design interactivity in an adventure means the degree to which there are obstacles and dangers beyond inimically hostile monsters and simple traps. An interactive adventure describes things in the area or room “key” that the players can ask questions about and do stuff with…

STEP INTO THE DUNGEON DRESSING ROOM

Interactive design offers elements of the adventure that aren’t a superstructure for simple die rolls - something beyond simple traps to “save” against and monsters ready to “fight until killed”. The more reliant one’s adventure is on these sorts of purely mechanical challenges the more the RPG becomes a sort of dice and board game. It’s likely that to some degree this was how Dungeons & Dragons was originally envisioned. That the dungeon adventure was originally intended to be played from something like the “zoomed out”, impersonal perspective of a wargame. OD&D reads like this at times, and one can imagine a referee running it who rolls each turn and declares new mechanical challenges or events … “You encounter three Ogres” or “One of your hired footmen falls into a pit trap as you go down the hallway.” Players then roll dice to fight the monster and remove resources such as their missing footman from their sheets, then they collect any treasure, move a few squares on the map and the next turn begins.

To run such an adventure you would need basically a map and set of random encounter tables, or a location with keys like this:

AREA 1 5 Orcs, 50 GP

There’s not a lot for designers to do with this kind of adventure, the dungeon generation tables at the back of the AD&D Dungeon Master’s Guide can generate a decent version of this sort of adventure. However, this isn’t how RPGs have ended up being played by most, not even in the early days of the hobby. Some outliers like Dave Megary’s DUNGEON! and the amazingly charming Rules for the Game of DUNGEON by then teenage Craig VanGrasstek seem to have sprouted from Arneson’s Twin Cities game, but for most, and certainly in 2025, this style of dungeoneering board game is not considered an RPG.

Gygax also recognized this over time and didn’t want to play a board game. The few pages of “Solo Play” dungeon generation tables he originally published in the first issue (Spring 1975) of Strategic Review Magazine were included in the Dungeon Masters Guide as Appendix A, but had added random generation tools in Appendices C, G, H, I, J, K, L, and M - mostly adding description and specifics to the bare halls of the earlier version. Appendix H even includes “Dungeon Dressing” - the sounds, smells, air movements and mundane objects that might be found in a dungeon. This, along with Gygax’s dungeon design, show an appreciation for “interactivity” and a step away from the board game possibilities of early D&D.  Gygax may have written more military or wargame style scenarios then some other designers, but he wasn’t playing a board game like adventure limited to mechanical challenges.
FROM PROCEDURAL GENERATION TO BESPOKE
Of course these type of random tables suffer a lot of issues - the most important being that the more complex randomly generated spaces become the less thematic and interactive they can be… that is the octagonal room filled with ... a feculent smell, mushroom patches, a chasm, 5 orcs and a portcullis trap doesn’t do much more than the ultra-minimalist room. The smells, sounds, trap, and room contents are only incidentally related to the room’s inhabitants and the contents of surrounding rooms. They are disparate descriptive elements that players can interact with by asking questions about, but they don’t “do” or “mean” anything that will help the players. All these sorts of dungeon dressing accomplish is to give a vague sense of the real to the dungeon - it is full of things and fixtures, just like real spaces.

The example given above is easier to fit together then many randomly created sets of features, but even when the dice line up like this in one room they rarely line up in the next or the one after that. Improvising coherence, theme, can create an evocatively dressed dungeon, but doing it again becomes increasingly difficult. The struggle to produce coherence from randomized table results quickly turns into another version of spending the time to actually key a location while thinking about theme and history. That is one has to go through the same steps as making a properly interactive dungeon in the first place. At this point procedural generation becomes an extra step. Of course if rolling on tables to create a dungeon and then fitting the puzzle pieces together is helpful for one’s process there’s nothing wrong with it of course … but at some point the dungeon designer will have to figure out how the dressing of the dungeon works and how to make some of it something more interactive - how to create a readable coherence that connects the descriptions of the space to its inhabitants, treasures and risks. To me it always seemed best to start with the themes and then design the dungeon around them. Procedural generation may be more useful at the highest levels of dungeon design - setting the specifics of a location, its purpose, overall themes and such - so solo game like Tony Dowler’s How to Host a Dungeon might be more useful then the DMG’s appendices.

PASSIVE INTERACTIVE DRESSING
Dungeon dressing, especially randomly generated, is almost always mere description, and because of this it has the bad reputation of adding pointless information to keys. Typically this becomes a sort of design maxim against listing mundane items in mundane spaces -- a “bed” as the content of a “bedroom” or “4 chairs” in a “sitting room”. This kind of description is called out as bland, pointless, or something any referee can improvise with ease. To the extent this complaint is about a lack of descriptive interest it’s true much of the time. Dungeon design can get bogged down trying to make dungeons too realistic or exhaustively listing the contents of rooms.  However, a greater issue with prosaic descriptions is that they almost always lack interactivity. Even simple dungeon dressing can have uses - sturdy wooden tables are a player favorite for making bridges, barricades and rafts for example - but this only matters if there are challenges that make such endeavours worthwhile and if the specifics the players are likely to seek out are described without too much focus on useless dressings. 

Assuming however that there are challenges and obstacles where the furnishings, supplies, natural features and other dressing in the dungeon often become tools for the players, the complaint about boring dungeon dressing is still valid. First, it’s easy to go overboard with dungeon dressing, and it’s easy to omit things that a referee might note.

Worse, it’s fun for many designers to go to excesses with dungeon dressing … fully imagining the space, seeing the dungeon in one’s mind as it changes over time, filling with clutter. Old junk is put to new uses and the trappings of the inhabitants pile up. At its core this is one way to create interactivity, making sure that the dressing has connections to the rest of the dungeon, and junk becomes something that players can use. A pile of owlbear pellets full of gnome bones might offer the party knowledge that the dungeon once had a population of gnomes and may still contain their typical traps and treasure, even if it’s now a lair for voracious growly hoots. As interesting as this sort of informational dressing is, it is passive and it doesn’t work alone.  Passive interactivity has to have a pay off - maybe not one every group will benefit from, but it needs to be connected to something that matters.  In the simplest case, like monster droppings, the information is a clue to a potential challenge.  In many other situations though, especially with clues to a location’s history, the knowledge offered by interactions with the dressing are harder to learn and harder to apply. Not every fresco will reveal a secret door, and not every statute provides the secret name to the long dead king terrorizing the dungeon. These sorts of powerful secrets should also require a bit of thought to use, and for them to work best need to be concealed among other less useful details. 

THE OVERDRESSED KEY
As much as one needs evocative description, dungeon dressing and interactive elements to work … the dungeon is still a fantasy space that acts as the arena for a game.  It’s great to include interactivity that lets players learn things about the space, make educated decisions about what they’ve learned, and combine those decisions with the things described within the dungeon (or on their character sheets) to overcome its obstacles. This design process though, leads to complexity, especially if one wants to make the solutions or helpful bits of dungeon dressing less obvious. It’s very easy to go to the opposite extreme from minimalism and to cram too much information and description into a space. It may even be that this maximalism is the natural evolution of designing dungeons with a focus on naturalism and thematic consistency … the dungeon keys of the 90’s and 00’s definitely tend toward unusable verbosity.

How to walk the line between too little key and too much? The answer is to consider the interactivity of the space one is describing… but not just the passive aspects. Include dressing in a way that gamifies it.

To think about this issue it’s best to think a bit about how RPG play works…

RPGs always involve several imaginations, the space a designer creates will be interpreted by both the referee and players. Visualized, described, visualized again, and then the story that takes place there will be negotiated by the players, referee and rules.  At each point here the fictional elements are being recontextualized, the possibilities considered and by the end the events are extremely unlikely to proceed as the designer expected. Designer ideas and concepts need to be durable and sticky - both easy to comprehend and something memorable.

The task of designing such spaces, and both using and writing such keys becomes harder the more complex they are. The more detail the referee needs to read, interpret, visualize and describe, the more the players will need to do the same and the more likely important elements that make up the room’s obstacles or that the designer has connected to other parts of the adventure will be forgotten. 

Consider the dragon’s lair above as described here.

AREA 123. Lair of the Dragon Waterfalls from the ceiling of this natural cenote, countless gallons of water that pour through a wide opening that leads to the jungle far above. An endless crashing roar of water on rock is deafening, and it is impossible to hear all but the loudest sounds, while the falling water creates clouds of swirling mist that block sight beyond a few feets and then suddenly whirl away to reveal the contents of the vast cavern. The cascades scatter light and paint dancing patterns of shadow across the slick stone walls and reflect from both yellowed bone and glittering gold. Rushing streams wind from the dozen pools beneath the waterfalls, through outcroppings of blue grey rock, clumps of pale subterranean fern, the thousands of yellowed skulls and tangled bones lay in moss covered drifts. Scattered among them are rusted and rotten arms - enough armor and weapons to outfit a legion, but one from no single era or culture. The half molten wreckage of the most modern articulated plate armor is tumbled atop the verdigris shrouded bronze panoply of the ancient legions. Scattered among bones, clear water, and bleached plants are pieces of treasure, gold temple vessels, blackened silver ingots, caskets of jewels, carved jade, gilded armor, and crates of rotten wood marked with the royal mint’s seal that spill coins across the uneven floor. In the center of the chamber the streams collect, and a copse of taller albino tree ferns reaching upwards toward the dim light above and home to foot long beetles that grow green with bioluminescence. Even with the array of trees the pool is visible, 12’ of clear water teeming with pale blind fish.  Its bottom is covered in more gold, bones and armor, as well as  a fallen idol depicting a frog faced rain god.  The great pool is the nest of Itzcotal - the “Sky Knife Serpent” an azure wyrm of great age and cunning. He rests with his sinuous bulk curled around the fallen idol, nostrils and watchful eyes above the surface. The wyrm can smell intruders and will rear up with a huge splash as soon as they move to within 50’ of its pool. If they have plundered anything on their approach it will gain +1 surprise as it quickly lashes out with its breath - a line of crackling blue fire. For those who have been more circumspect Itzcotal seeks to awe and intimidate and will negotiate from a position of strength, demanding sacrifices (of trivial things, gold, or party members depending on the reaction roll) to refrain from attack. Like all of blue wyrm kind Itzcotal is unpredictable, a creature of sorcery and madness whose response changes each time it is encountered. It is also a beast of incalculable vanity - and susceptible to flattery to the point where the Sky Knife Serpent can sometimes be cajoled into trading gold and flattery for spells or knowledge. Omitted: Dragon Stats and treasure hoard details.

I think it’s a pretty cool space, and a dragon lair should be a bit majestic … but here I’ve spent my first few paragraphs describing a dragon’s lair of cascading waterfalls, mounded skulls, and broken rock pillars … there’s even a good chance referees will miss the dragon, puffed up with a gout of apocalyptic flame and ready to incinerate guests. The referee won’t even miss it because they are lazy, but rather because they’re already answering questions about, or trying to figure out mechanics for waterfalls long before the fourth paragraph where the dragon is noted.

Itzcotal’s waterfall cave lair is worse, because while they are described many of the details aren’t described in ways that make them interactive - and once you’ve found a lair full of gold and angry dragon passive informational interactivity (also really here) won’t help much. This encounter could be fairly complex (beyond the complexity dragons always bring) based on the room key - the waterfalls provide not only mist and noise but may protect against the dragon’s breath.  Likewise the rocky outcrops, pools, ferns and such. These are all potential actively interactive details - things that good players will try to use for advantage in dealing with the ancient wyrm. They should be able to use them as well.

A good referee should think out some common possibilities before the encounter and a good designer should make this easy by offering mechanics or at least useful description for the dungeon dressing that players are likely to interact with… will the fern clumps hide a character? (yes of course.) Can they dive into pools to escape the fire (no - it’s magic! Diving behind a waterfall though might give a big bonus to saves as the dragon can’t see through them) Can the party sneak up on the dragon in the mist and noise (unlikely, say a 2 in 6 chance of surprise.)  Without this sort of information the referee is expected to make up rulings for everything in the room, and the more complex the room the more likely they’ll be dealing with stranger questions like “If my tree speaking druid talks to the ferns will they fall on the dragon?”

Evocative dressing isn’t enough, because a big part of the designers job is to provide answers for the basic mechanical questions a dungeon or dungeon room presents, so the referee can save their mental energy and focus on the weird questions. Many of you will have also noticed that I have pulled a bit of a trick with this dragon lair - it also has bad information design, and this is a big concern for published adventures. To make this room work, not only would you need some details about the interactive features, but also the dragon should be first.  Of course that doesn’t fully solve the problem - especially if the room dressing has mechanical effects - the referee and party will be so focused on the dragon that they might miss the waterfalls and trees that they can use to their advantage…

THE UNDERDRESSED KEY RETURNS
Keying and providing interactive dungeon dressing isn’t just about information order and adding interactivity. A part of the designer’s job, especially when writing works for others to use, is thinking through obvious possibilities that one’s keys present. Privileging the active interactive feature over the passive ones and spreading out interactivity and detail over the whole dungeon so it’s not overwhelming in any one place.

This means performing a tricky balancing act - providing passive interactive elements and general logic that lead to, highlight and sometime soffer potential solutions to the obstacles of the dungeon, but also keeping things from being too complex or worse confusing and contradictory in any individual room.  Once players start thinking about the meaning of a dungeon, its history and how to use it they often get a bit obsessed - they read into unexpected things and this means a good designer has to keep them from wandering down too many mental dead ends. On the scale of the whole adventure this means avoiding incoherent location design -- things like a guard barracks whose only access to the throne room they guard is through the “hall of traps”. At the individual key level It’s more complex, but largely accomplished by making sure one’s useless information and purely descriptive dressing aren’t too tempting - either to the referee or the players.

Consider again how players and referees both need to interpret and visualize the dungeon, and that this visualization will likely focus on the main things in a keyed space: enemies, treasures, big obvious obstacles/puzzles, a few general details, interactive elements, details about unique dungeon dressing, and finally mundane detail. This forms a natural hierarchy of design with the most important elements always required in a key and the rest being increasingly optional or open to description in a more cursory way. Even with strange environments this hierarchy holds - perhaps more so, because the designer needs to give more detail about the strange and wonderful. Of course even the strangest spaces have mundane aspects - and when describing such a space these mundane details can usually be omitted. A crystal space elf kitchen will still have pots and pans, and a wizard’s bedroom a bed for example. One doesn’t really need to describe them in a dungeon key because the referee can manage that - use the space you have instead to describe what’s different and abnormal about these types of spaces.

Using these techniques one can design fitting challenges for specific levels of characters by making interactive spaces - combine “evocative detail” with passively interactive elements useful for solving obstacles and add active interactive elements to promote players overcoming obstacles. This is essential in games where there’s no rolls (skill checks etc) to overcome such obstacles, or where such checks can be avoided through clever plans. Even where such rolls are common, interactivity has the advantage of allowing obstacles that are difficult to overcome through skill or other dice checks alone - because they can be unpuzzled with clever thinking. Interactive key design frees the designer from some of the difficulties of “balance” because rather than assuming that challenges must be gauged to the skill level of the characters challenges can be included that have interactive solutions found within the dungeon. Variety and possibilities in a fictional environment are the heart of interactivity - players making connections between the elements of the fictional space and using them to understand and likely solve the problems it presents. Sure the players can always trivialize or be utterly stumped by a seemingly simple obstacle, but this is why one includes multiple paths with a variety of obstacles (or multiple dungeons). 

EVOLVING ONE INTERACTIVE KEYS

When designing a space it helps to visualize what the basic approaches might be and help the referee by including the information they need to manage them. 

For example…Consider this somewhat minimal key...

AREA 1 The Gate of the Swinehold stands across a muddy chasm, it’s bottom covered in effluent and broken crockery.

Now assuming the players want to enter the “Swinehold” and get to plundering, or even if they are bringing a bunch of blueberry pies to invite their swinish pals out for a picnic, there’s a few obstacles here. Specifically a chasm and a gate … and there’s a lot of missing information. What might a referee want to know to better run this area?

1. How wide is the chasm? How deep? How muddy … is the bottom quicksand or something?
This matters because the party will likely want to cross it.

2. Is there anyone home in the Swinehold? Are they watching? How do they react to visitors?
The party may not have to cross the muddy chasm if they can interact with a creature on the other side and buy or bluff them into offering a way across (and presumably they have one).

3. How sturdy is the gate? Is it locked? What’s needed to force it open? Will there be anyone trying to prevent the party from opening it?
The players are very likely to care about these things and all are very likely to come up.

So what are the players likely to do in this room, and what’s likely to happen - assuming they will have reasons for crossing the chasm and opening the gate?

They may try to jump the chasm, maybe pole vault it, or climb down and then back up on the other side. They might try to make a bridge - with ropes and a grappling hook most traditionally.  They may use magic or special abilities to fly across or otherwise circumvent the chasm. Finally, they may hail the gate and see if something will let them in…

A good key will offer the referee what they need to answer these questions. This doesn’t necessarily mean that these are the only ways the players will try to cross the chasm (players are inventive), so what the referee needs is more than just answers for each obvious possibility. Instead the referee needs a description that provides sufficient information about the space to make judgments about the consequences of the various methods that the players try to overcome these obstacles.

Something like this…

AREA 1 - SWINEHOLD GATES A chasm splits the 30’ tall natural cavern strewn with boulders, stalactites, stalagmites, and patches of knee high white mushrooms. These natural features make it difficult to see the iron and timber gate is built into the wall only a dozen feet beyond the chasm’s edge. The chasm bisects the cave cleanly, always roughly 15’ wide, and drops precipitously into the darkness. The chasm floor is a jagged gulley coated in filth and broken crockery, 80’s below. A foul, feculent odor rises from the depths and the walls of the chasm - friable, coated in filth and clinging fungus - make climbing difficult (requires 3 climbing checks at -1/-20%). The wide gate itself is made of four wooden timbers, still sound despite their age with rusting iron rivets and wrist thick crossbars. A set of heavy 20’ long timbers and boards lean near the gate and can be used to create a makeshift bridge in two Turns. Visible in the shadows beyond the gate, a cow sized stone counterweight, pulleys, wheels and a set of ropes too thick for an arrow to slice keep it closed. The gate can be lifted or broken only by a feat of enormous strength or with the help of engineering or mining tools. The Swinefolk keep a poor watch on their gate, a trio of sentries that spend the majority of their time racing centipedes and dicing in their Guard Hole (AREA 2).  Their is a 1 in 6 chance every other Turn that one of the Swinefolk Sentries will walk to the gate and peer put, usually in the process of relieving themselves or tossing a pail of filth through the bars. Even then they are not especially observant and will only notice careful strangers who are near the chasm or gate. The swinefolk will be drawn to investigate especially loud noises or magical light.  They are typical of their kind, boorish, suspicious and greedy - they are especially fond of “cave truffles” and beer.

 
This key should provide all the information needed, with some evocative detail to help referees and players visualize the space and hopefully make some of the key ideas stick. Unfortunately it’s a bit overdressed … too long… Now a good editor might pare it down a fair bit, but it has at least three interactive parts, or more like 5… The chasm, the gate, the gate/chasm crossing mechanisms, the guards (and the cavern formations that hide anyone far enough from the chasm). Since all of these have to be managed, it’s tricky to cut things down and retain useful information.

What can be done is to organize it better.

Bullet points are popular thanks to OSE and its adventure format, but bullet points are also limiting, they don’t work well for things that have more than a few bits of information attached and they tend to limit detail. They are another form of minimalism, though more expressive then some. Plus I personally don’t enjoy bullet points - they remind me of one of my first jobs and writing emails to seniors who insisted on “executive summaries” in bullet points. The goal of a location key isn’t just to summarize a space, but to provide imagery and even phrases that help referees visualize it so they can best describe it to their players. As with all forms of minimalism bullet points aren’t the best tool for this, as much as they can offer clarity, save space, and help people with poorer reading skills and shorter attention spans.

Instead my personal choice for simplification is rigorous editing to cut away extraneous information, confusing phrases, and generalities while retaining as much description as possible. To aid with comprehension I also use both minimalist descriptors by the area name. These cover the lighting conditions, dangers, and secrets in the room.  Second, I always start with a general area description, and bold the important room elements. These include risks such as traps, monsters, obstacles and secrets, as well as things that are likely to be carefully examined - that is the major interactive aspects of the key.  Bolded words then get their own sub paragraph.  In the end I aim for keys that look like this:

AREA 1 - SWINEHOLD GATES - DARK - MONSTER AMBUSH A trio of swine folk maintain a poor watch of this natural cavern, peering occasionally through the iron and timber gate at the cavern’s Southern end. A chasm splits the floor 20’ before the gate, wafting a feculent odor from its depths. Poor Watch (MONSTER AMBUSH - 3 Swinefolk Guards): The swine folk in the Pork Hole (AREA 2), just beyond the gate rarely glance into the cavern, preferring to race centipedes and bicker. Every other Turn there 1 in 6 chance that one will emerge to glance out or throw a bucket of slops through the bars of the gate. The swinefolk are not observant, and while they will be attracted by loud noises or significant light, the cavern’s natural features will hide anyone who is more than 10’ from the chasm edge. If they spot dangerous looking intruders they rush off to ring the Swinehold’s alarm.  Otherwise the guards are typical swine: boorish, talkative, suspicious, and greedy - they are especially fond of “cave truffles” and beer.  Natural Cavern: strewn with boulders, stalactites, stalagmites, and patches of knee high white mushrooms, with flowstone walls that rise to a 30’ ceiling. These features provide concealment to anyone moving through the cavern with care.   Iron and Timber Gate: 20’ beyond the chasm, the gate is set in the cave wall and built of split tree trunks, still sound despite their age, with rusting iron rivets, and wrist thick iron crossbars. A set of heavy 20’ long beams lean near the gate and can be used to create a makeshift bridge in two Turns. Visible in the shadows beyond the gate, a cow sized stone counterweight, pulleys, wheels and a set of ropes too thick for an arrow to slice keep it closed. Only enormous strength, siege equipment or mining tools can noisily force the gate. Chasm: 15’ wide and 80’ deep along its entire course between the walls, it drops precipitously into the darkness. The chasm floor is a jagged gulley coated in filth and broken crockery while the walls - friable, coated in filth and clinging fungus - are difficult to climb (requires 3 climbing checks at -1/-20%).

You may notice that this key takes up more space then the one I said is too long, but the key is likely easier to understand, and the organization places risks first and follows a more useful information hierarchy.  It should still be easier to play though.  Plus this is a complex room - the guarded entrance to an organized faction fortress.  Not only does it include a potential encounter with the guards - and the subsequent activation of the entire faction against the party, but it has a number of actively interactive features that both allow the party infiltrate the Swinehold and multiple obstacles (guards, chasm & gate) that they will need to overcome to do so.  It might be possible to reduce the length (beyond the removal of a couple of lines by a good editor), but the best way to manage this now would be by moving some aspects (the swine folk interactions) to a separate section -- a swine folk order of battle with ideas about how they guard and defend their fortress.

Of course these kinds of regional (the region here is the Swinehold) meta-systems are beyond including interactivity in an individual key and instead operate at the level of the entire dungeon.  

SOME MAXIMS ON DUNGEON DRESSING

1) A good dungeon is interactive
I’ve spent the whole post here describing how to flesh out an adventure with interactive bits … but to move beyond adventures that are a matter of calculating movement between fights and tallying up treasure, one needs to include other obstacles that players overcome by unpuzzling them. To run such an obstacle requires the referee to have a solid idea of what it is and what happens when players start messing around with it and with the various things in the dungeon that they might consider tools to overcome it.

2) Detail and Dungeon Dressing isn’t always interactive

It’s helpful to fill a dungeon with description and things that are not interactive.  This helps offer a bit of concealment (and gives the players the fun of making decisions about what matters), but it also builds theme and coherence. However, detail, no matter how wondrous or fun does not on its own offer interactivity - obstacles need to exist for players to overcome by using dungeon dressing.

3) The Dungeon can Never be Real

It’s impossible to design an RPG location with sufficient detail to cover every question and define every possibility within. This leads to over long keys, and perhaps madness. Detail takes up time, space and referee focus, meaning it’s easy to overdo. Instead stick to the most exciting, most unique and most wondrous aspects of dungeon dressing and leave the theme to help referees describe both mundane items and answer less obvious player questions. All art by the late great Russ Nicholson