The Problem With Dungeons

I ran the Iron Coral from Into the Odd: Remastered today. It was fun! At the end of the session, I asked for feedback. The main points of concern were as follows:

  • The characters were too unrelated to tie them together effectively, and no real attempt was made at doing so.
  • Aside from ‘get treasure’, the characters didn’t have a meaningful motivation that the players were excited about.
  • The dungeon didn’t have enough opportunities for social interaction between characters or NPCs.

The dungeon itself is great, don’t get me wrong. But, when I cut out some combat and the players still felt like there was too much negative interaction, it makes me think. I normally am very adverse to combat, but I tried to make a point out of running the Iron Coral by-the-book. I rolled random encounters, I didn’t tamper with any room descriptions.
After the session was done, I thought that the problem was a similar problem I’ve seen in OSR—that choosing not to interact with something is often the best choice, and that only bad can come from doing so most of the time. It’d be different if the game was intentionally doomed, like Trophy or something, but you’re supposed to be smart and outthink the GM in OSR. You’re supposed to try to win.
What I’ve now come upon is that it’s not just the OSR. Sure, that problem exists and is relevant, but beyond that, I moreso think that the fact that combat exists and will occur makes the problem expounded all on its own.

The Problem With Dungeons Is Its Structure

When I first got into FKR and minimalist games, I thought I was stripping away the rules to encourage more roleplay and interesting situations. But, I don’t think that rules are necessarily antithetical to roleplay anymore. I’ve had a Play-By-Post game going for over 3 years now. My favorite moments of the game were times when the rules didn’t apply, and that’s what originally led me to my conclusion. However, there was something else that also bound those moments together I hadn’t drawn anything from. Each of those times were nonviolent. Two PCs developing crushes on each other. The PCs falling asleep on each other after a long day and night of no sleep. In addition, those times were unsolicited and without structure.

I don’t think rules are the problem, it’s when they provide a structure that they begin to be a problem.

Structure, in this case, is when the method of play changes into something more rigid.

Commonly, this is combat. Even without rules, this structure exists, since the goal shifts from whatever the PCs wanted to do and instead “fight for survival and defeat the enemy,” which is inherently limiting. It can really be any conflict, as long as the PCs are forced to set everything aside to tackle one goal. This won’t mean that my games in the future won’t feature conflict. It just won’t be my favorite part.

This is a very longwinded way of saying that dungeons are extremely structured. They force players to act more rigidly in every sense of the word. You have to act in dungeon turns, you have to set a goal of what you’re looking for, you have to go into combat, you have to act carefully, etc. This structure can be very nice for a lot of GMs and players. However, it is no longer novel to me, nor what I truly intend for my roleplay to be, anymore.

Very interested in hearing what you all have to say about this. See you around!


Sounds like you and your players want to play a more narrative game about character interactions, emotional connections, and relationships?

That doesn’t strike me as a dungeon design issue so much as a play style issue. Your table doesn’t want a dungeon crawl, even one with heavy faction intrigue. That’s fine of course, but it’s true that running any published location based adventure I can think of is unlikely to scratch the itch.

I do think though that any game, at least to some degree, will need to include procedure and/or rules, even a game focused on storytelling about interparty relationships. Figure out your preferred locus of play and find a system that simplifies or elides everything else.


I think you’re saying you dislike scenario structures in RPGs because you find that such procedures assume the players have a certain motivation (e.g. “get treasure” or “win the fight”), but your favorite moments in RPGs arise when the players develop motivations independently.

Your first example is that your players don’t feel motivated to get treasure, so the assumptions of the dungeoncrawl procedure are falling apart for them—it makes more sense for them to just not interact with the scenerios. Your second example is combat, which assumes the players will want to fight.

Correct me if I’m wrong about any of that.

It partly seems to me like a conversation about buy-in or lack thereof. If I’m running a treasure-hunty dungeoncrawly game, I want my players to buy in to the idea that treasure-hunting is what this campaign is about and that their characters are treasure hunters who want nothing more than to hunt treasure. If they buy into that, then choosing to not engage with the dungeon is out of the question—it’s part of the conceit and the social contract of the game.

But, then there’s the broader question you’re raising: what would it look like to play a campaign without any procedures? FKR may be (typically) rules-light, but FKR types still use dungeoncrawls, even the original Free Kriegsspiel assumed you’re trying to win a battle. It sounds like you’re proposing a kind of inverse FKR: rules, but no procedures.

I’ve already been writing this post for an hour so I’ll leave it at that for now.


I totally understand where you are coming from. A lot of what you’ve said resonates with me.

BUT, I come to similar but different conclusions. I, too, am also over dungeons. They don’t interest me. I feel like I have similar issues with dungeons, but my conclusion is that I don’t really like the style and theming of dungeons. I don’t like the play that dungeons generally produce. To me, it’s not about the structure of dungeons, or the structure of scenarios, but the focus of play.

For me, For the Queen is an excellent game that is all about the characters. Answering pointed leading questions and learning about the characters (and the world) is the WHOLE game.


What system do you use for this? I’m assuming Into The Odd. Did you consider running it with Electric Bastionland?

The nice thing about EB is the shared debt mechanic and Failed Careers ties the party together in a way the original Into The Odd doesn’t.

But I agree with the others here that it sounds like you want something more relationship focused!


As one of the players in said game, I agree with the sentiments @SageDaMage expressed, although I don’t agree that the structure of a dungeon crawl is mutually exclusive with Social Intrigue or other focuses of play which I prefer to loot dungeons (another for me being TTRPG as Performance Art). In fact, social intrigue dungeon crawling is what the Court Crawls of Maximum Recursion Depth are supposed to be. Whether or not I succeeded is a separate question, but at least when I run it, it has worked out that way.

@yochaigal imo what Chris does with EB, while elegant, is more like a post-hoc way to justify the dungeon crawl, which still ultimately exists to be a loot dungeon, vs. social intrigue or acting as a vehicle towards some other end like the aforementioned performance art idea. There’s nothing wrong with loot dungeons necessarily, it’s consistent with OSR sensibilities, that just doesn’t really interest me much anymore, but that’s a personal preference.

@flyrefi I can’t speak for Sage but being one of the players in the game, I would not say so much that I’m opposed to a pre-determined shared motivation at least as a starting conceit, I just don’t have much interest in that motivation being “get treasure” or “win the fight” in itself. As I was saying above to Yochai Gal, I also think it needs to actually be the central conceit and not just a way to rationalize a loot dungeon. So I do think that rules vs. procedures question is really interesting in itself, but at least for me, it’s not core to how I feel about dungeon crawls, although we may need to operationalize those terms as it’s possible you mean something different than what I’m thinking.

@GusL I haven’t read Iron Corral, only played in this game that Sage just ran, so maybe there’s more faction intrigue than we saw, but to me it felt much more like a loot dungeon. A dungeon crawl that is very factional and based on faction or social intrigue is probably my ideal form of play. I agree that most published location based adventures are unlikely to scratch that itch, but I think that speaks more to the fact that there are few people trying to write dungeons that way, and not because it’s impossible. If that’s something you’re interested in, check out MRD and let me know what you think ;).

@yoshi I don’t know what your sensibilities are, but have you seen the show Doom Patrol? I just finished Season 3 the other day, and it does an amazing job of demonstrating how to do what are often, in effect, “dungeon crawls”, but where the emphasis is much more so on the psychology of the characters or their social interactions, with a heaping dollop of Weirdness on top. The typical loot dungeon focus doesn’t appeal to me anymore either, but don’t throw out the baby with the bathwater!

1 Like

I think you’re interpreting what I said correctly. The problem is that now we have two sets of criticisms of the dungeoncrawl in the same thread—yours and Sage’s.

Sage is saying that there’s a problem with all procedures that assume a goal. You’re saying you have a problem with dungeons that only exist to be looted, and not to serve a more complicated purpose.

These criticisms don’t strike me as all that related, despite the fact they come from the same game that Sage ran and you played.

Frankly I think this thread is in danger of spiraling in too many directions and confusing those who reply to it, unless we can do a better job of defining what this thread is about.

While I do have a problem with structure overall, I do understand its importance in creating a narrative, and think that a game without structure would be difficult at best. As mentioned in my initial post, it just wouldn’t be my favorite part of the story.
Aside from my generalizations about structure, I did make this post intending to talk about dungeons. I think @maxcan7’s reply is relatively on-topic to the discussion. As far as I interpreted it, they want a structure that isn’t as… selfish, for a lack of a better word.
However, separate from Max, I think the dungeon structure needs reform or removal before I feel like I’d want to come back to it fully.

1 Like

I think that GusL is right that this is a playstyle issue. I also agree with Flyrefi’s point: if you are running a dungeon game for your friends, it is part of being a good player to come up with characters who will be motivated by the goals of the scenario at hand. Yes, dungeon exploration, especially when combined with GP = XP, creates a certain ethos, and if the players are not on board with that, they will be unhappy with such a game. Once players have agreed to play a game with a certain premise, it’s on them to tie themselves together with motivations and relationships. Otherwise, the Referee can do that for them, if they are willing.

I do fully sympathize with the feeling that choosing not to interact with stuff in typical dungeons is often the best choice! Everybody has their different priorities in dungeon design. Mine includes not constantly punishing the players, but encouraging curiosity and bravery. It seems to me that most dungeon designers favor “whimsy” (or what appears to me as unpredictable, punishing nonsense). That’s not my style, but I’m not a dungeon addict, either.

Although the gurus of the OSR have often considered the dungeon to be the appropriate basis of a sandbox, old-time designer Ken Rolston (for one) saw it as a linear scenario–what we call a railroad today. There are only so many options for where to go in a dungeon. The players’ complaints about the dungeon that prompted this discussion make it sound as if they too found the scenario to be like a railroad–not giving them the kinds of options they wanted. Still, it’s on them to communicate that up front.

A lot of folks have heard me say that in the heyday of the early '80s, the time everybody imagines as the classic era of dungeons, I knew no player groups (out of of a large number) who used dungeons as the basis for adventures. We had plenty of fun without dungeons, and you should play as you wish–of course!


To me, I look at some of these points and when I think of the opposite (or at least not that) I get:

you have to set a goal of what you’re looking for

So maybe you want some thing where the point of exploring the dungeon is the exploration itself? Where you are trying to learn something and about the place and you are expecting to be surprised?

You have to act in dungeon turns

So something less time based and something maybe more narrative beat based? Where you can skip to the next interesting bit, instead of going through a time (rounds/seconds/minutes/hours) approach?

you have to act carefully

So, your character can make risky actions and the consequences won’t be such that the character is no longer playable. Either you know the potential outcomes, so you know when you are going to do something that will remove your character from play (and the other options are ALSO fun/interesting). Or you have mechanisms to mitigate those consequences so you can still do risky things and have a playable character.

you have to go into combat
So you want non combat mechanisms/rules/procedures to be just as if not more supported than combat mechanisms.

Is this more or less what you are thinking? Is there something I don’t understand? What would an ideal dungeon adventure look like to you?

1 Like

Thank you for beginning to clarify what you meant by your initial post, but what kind of “reform or removal” do you have in mind?

I also haven’t read Iron Coral, but I was giving Sage’s judgment on good and bad adventures the benefit of the doubt. In general I don’t use ITO/EB (or ultralights more generally) as they don’t seem to add much to systems I know already and often fail (I hear ITO is an exception here) to emphisize, make rational, or even include (Mork Borg I mean you) turnkeeping, supply, and risk mechanics that make dungeon adventures possible. However I suspect I know faction intrigue based dungeon crawls fairly well…

My point isn’t that faction intrigue based adventure are hard to design for Classic or even ultralight POSR systems, but that what Sage was describing was more then that, a certain disinterest in the focus of the system, adventure and play style that was being viewed as a failure of adventure design rather then a desire to play a different kind of game.

Faction based dungeons are simple enough to design and run if one utilizes minimal social mechanics like morale and reaction rolls, and applies design principles like asymmertical encounters, minimal skill systems, and XP = GP that discourage combat in favor of problem solving. The issue I’m reading from Sage though seems broader then simply playing with an adventure that focuses on combat resolution (though this can often be a player or referee problem, especially for folks unfamiliar with the ethos of play in Classic games) or puzzles – a seige or trap maze location as I like to think of them.

I hope I’m not reading into Sage’s point or missing it, but it seems like they are feeling burnt out on the whole structure of location based exploration because it doesn’t offer clear ways to build a narrative that focuses on character’s social interactions and doesn’t flow convincingly from the backstories players have created or directly engage them? The question I’d find interesting there is if this is more then simply a disconnect between playstyles – specifically I’m alway interested in hearing how folks from outside the play style of Classic (or perhaps OSR) play concieve of dungeons and dungeoncrawls.

1 Like

I would love that, exploration for exploration, if it worked. Sure, I’ve had it work for some groups, mostly beer & pretzel types, where I could literally plop a dungeon wherever and they’d do it, but it’s not universal. I mostly play with strangers, so I’ll never know if a group will vibe with that, as most don’t know themselves ahead of time.

Yeah, I think that beats would be helpful for exploration, like a point crawl, where the characters just go from bit to bit.

Well, not really. I think that danger being present immediately triggers the “you have to act carefully.” Sometimes you don’t have to act carefully, however, as the game is maybe more cinematic or heroic. In that situation, the structure doesn’t shift to acting carefully, and continues with just playing to find out what happens next. That’s what I’d prefer, it doesn’t really have to do with character death.

Yeah, although alternatively, no combat mechanisms at all, as I’m more than comfortable using my own procedure for such a thing. Gives a very important statement, too, when there aren’t combat mechanics.

That’s tough, and I don’t have an immediate answer, because really, my answer atm is, “not a dungeon.”

1 Like

This! I love the mechanics of Into the Odd. I think they do perfectly what they set out to achieve. I think the adventure is perfect. Again, for what it sets out to achieve. I think what I want to achieve in a game has shifted, fairly drastically, is all.

Yes, this is accurate. I want more player involvement, to put it simply.

1 Like

I have an example that answers this question and @yoshi’s question of “What would an ideal dungeon adventure look like to you?” which is: Halls of the Blood King. Removing the hostile guards, the entire “dungeon” is an extremely social and weird space that you can’t fight your way through. Indeed, even the “boss” can be defeated with the revelation of discovering his mother.


So (and this is me arguing for my own favored design terminology) I think it’s helpful to look at dungeon as a rpg design term of art. A location based adventure for a specific style of play. Specifically A DUNGEON crawl is a playstle focused on the PROCEDURAL EXPLORATION of a fantasy SPACE.

This becomes critique of Halls of the Blood King, which is a good adventure, but I think stumbles precisely where it fails to grasp that it’s a social adventure not a dungeon crawl. All vampire castles almost have to be with old style vampires – 8HD, multi-level draining killing machines with plenty of defensive powers. This is of course Ravenloft and the Hickman’s original observation, and one ties in directly to the problems of trad design. Like Ravenloft, Blood King attempts to use the design tricks of Classic play: measured movement on a complex map and keys built around location gor an adventure that might better be considered as a web of NPC relationships. It does it well. But I think a few more pages with timelines an faction web and similar social aides could have made it a stronger adventure. Also an understanding that 72 Turn may not be enough to complete a complex social adventure and exploration if one is actually using OSE RAW.

Point being some intentional design and choice of system can go along wayvwhen one knows what one either wants to play or design.

I run a lot of one-shot dungeons and I agree that as written, not many provide RP opportunities but for the filling two points I usually solve that with a player setup chat.

When we go over the characters they want to play, usually I then ask the following:
“So what brings a [character classes or backgrounds] together? How do you know each other and what do y’all usually do together?”

And then individually or as a group of they have a group goal:
“Now your character is DESPARATE for money. And not a small amount of money, but like, BIG money; chests of gold and treasures. What do they want this for?”

And as far as those two points your players put forward, that little setup ties them in immediately, and should they finish the dungeon, we get to have a narrated ending of what they choose to do with their rewards (usually me narrating with prompts to each player for specifics). In that sense, they feel like they “won” and the character and backstory they created gets closure in a creative way. Also ties them together when they create that party history.

For me, dungeons are a tool of technical play that (if you are lucky) provides in-built narrative flavour. So with that in mind, it’s less that dungeons have a problem and more that they’re to be used for a purpose, and that purpose is the party doing shit. The dungeon can’t create the party or drives, but with some prompts and player creativity, the dungeon can be a good playground for any team the players create (again, with a few extra prompts).

1 Like

One thing I’ve started doing about 1-2 years ago, inspired by PBTA and Ironsworn, is what I call “vignettes”.
You can plug it anywhere you want, right in the middle of dungeon exploration or during a travel journey or whatnot.

It pauses the action for a bit and zoom-in on a specific character (or pair of characters). The GM asks a specific question about the feelings of a character or what he’s thinking at the moment, or portray a personal interraction between two characters. You play out a quick scene to answer the question, and then move on.

For example, while the party is exploring a deep dark cave, you ask one of them: “Your character gets lost in his thoughts while delving in the silent darkness. Who does he remember about? How does it make him/her feel?” or “While you’re cooking at camp, one other character is really starting to get on your nerves. Which is it and what are they doing that is really annoying?”

It’s not really a counterargument to “dungeons are not fun to me”, but at least it’s a way I’ve been using to injects back a good dose of relationship/emotions/character developpement in a session!


So (and no beef here - just noting a play style difference) these sort of vignettes, not the concept as a whole - especially if looking inward at the PC’s motivations/past/goals etc, would likely be very much against the spirit of classic style play.

There’s a sort of strong presumption in classic design that referee control covers everything but the inner world and emotions of the characters. The referee has the right to say “rocks fall you all die” (of course they shouldn’t [see below] - it’d likely be unfair in other ways), but none to say how a character feels. So anything that includes deciding who the PC finds annoying is unfair antagonism from the GM (barring magically induced annoyance - which is obviously externally caused because the referee says it happens). Likewise the getting lost bit of the first scenario is also against the presumption of referee and player roles, without some mechanical modelling for sure, the referee may theoretically have the narrative authority to declare the party lost, or drop the death rocks, but is trusted only to do so when it flows logically from the course of play. In the lost in the dark dungeon (a usually very very lethal situation in a classic dungeon crawl) situation the question would be fine, but getting there would have to be the result of player mistakes (e.g. not bringing enough light sources).

To me this is an interesting difference between player expectations and system design between play styles. It also highlights that when we talk about “dungeons” or other game tools and scenario design questions, desired play style (and likely intended system, but let’s not discuss how and if that matters) is the antecedent question.


I guess a more OSR approach would then be just ask “How does your character feel?”.

The problem with that is it doesn’t really spark the imagination. Most player will go: “Err… fine, I guess?”
I find a more focussed question helps to have an actually interesting scene, but like you said, different playstyle!