The World Is Your Ashtray: meta-dungeon strategic play in D&Dalikes

[Marshall Burns]

Ok, so this didn't take as long as expected.

First, a definition
"D&Dalike" here refers to any game (by which I mean what goes down at the table , not what text is used) in which the PCs are a party of free agents (in the sense of not having any meaningful social roles/ties/obligations like they would in e.g. Hero Wars or Burning Wheel) in a fantasy world who go on perilous journeys and negotiate hostile environments (esp. "dungeons") in the pursuit of treasure and/or MacGuffins, with special attention paid to armed conflict with the enemies to be found in those environments. This is a really broad definition that encompasses all things named D&D (as far as I can tell) and plenty of things not named D&D. There are plenty of variations on this form regarding the emphasis of certain portions (frex dungeons over wilderness, or vice-versa), morality (PCs may be expected to be heroic, or roguish, or maybe there are no expectations at all), agency (Illusionist or bald-faced railroading, Participationism, certain meanings of "sandbox," etc.), and probably plenty of other factors that aren't occurring to me right now. The thing I'm here to talk about is one such variation that I have arrived at.

Another definition
"The meta-dungeon" here refers to the portions of a D&Dalike game that inform and lead to the perilous journeys, dungeons, and other direct conflicts (such as fleeing the constabulary, mass warfare, what have you). As far as I can tell, nearly all* D&Dalikes have a meta-dungeon. When the farmer asks you to kill the ankhegs that have invaded his fields, when the mayor asks you to defend the village from lizardman raiders, when the shadowy stranger in the tavern entrusts you with a treasure map shortly before being murdered, when you exchange your treasure for cash/goods/services, when the king tasks you with defeating the Dragonlord and recovering the Ball of Light -- that is meta-dungeon. The amount of depth and detail the meta-dungeon has in play varies widely. For some, it's simply a matter of "click here to convert treasure to cash" and "click here to receive next quest" -- and that's all well and good, if that's what you're into! At the other end of the spectrum, you have things like what I'm about to discuss below.

The World Is Your Ashtray
TWIYA is the name I've given to a specific method of "sandbox" (I won't try to define that one, but hopefully you can see what I mean from the details) Step on Up D&Dalike play that attempts to make the meta-dungeon as interesting and consequential as the dungeon and ground-level challenges.

The structure of TWIYA is as follows:
1. Play follows the players. If they want to go north, the game goes north.
2. The GM's first task is to prepare a world in general, and to prepare the interactive details of any specific locale the players journey to.
3. The GM's second task is to communicate and portray their prepped material in play as accurately as they can, according to their vision.
4.    During play, the players decide which details they will interact with and how, with the understanding that their goals are to achieve the goals of their characters (which they decide and may change at will).
5.    The GM judges the consequences of the players' actions based on any mechanically-mandated results (e.g. characters being killed) and his own vision of the prepped materials.

No. 1 requires honest-to-god actual options. For "let's go north" to be a meaningful decision, there has to be something in the north that isn't in the east, west, or south. Magician's choices, palette swaps, and Roads to Rome are cheating. Furthermore, there has to be some method of gaining at least a little information about what's in the north, if the players wish to find out before they go there. The best way to square this with no. 2 seems to be extra-session communication between GM and players ("Where on this map do y'all want to start? Do y'all want to travel next session? Etc") so that the GM can prep the local details he will need for each session.

No. 2 requires at least one grabby (as in, makes for an interesting board to play on) situation that the PCs can learn about and engage with when they roll into town. It needs to be a turbulent/unstable/barely stable situation that, given PC intervention, might topple like a house of cards. Multiple power figures, factions, and maybe even sympathetic commoners need to have a stake in it turning out in a certain way. It's up to the players to figure out how to profit it from it and to decide how or whether to engage with it. The players may not be interested in a given situation; sometimes that happens, and if it's something the GM is worried about, he can engage in extra-session communication to find out before he goes sinking hours into a situation only to have the players say, "fuck this, where's the nearest dungeon?"

For a concrete example, here's a situation that I cooked up for the (currently defunct) Madlands (which is designed to be a TWIYA game) alpha tests:

The Former U.S. is now hundreds of mostly-tiny nation states. One of these is the Kansas Free State (actually parts of Kansas, Nebraska, and Oklahoma), a peaceful country beset on all sides by various warlords and penny-ante daimyos. Despite having no military of its own, the KFS survives all attempts to conquer it. The reasons all boil down to the Great Beast of Winter, a giant monster that likes to blizzard all over things, which is one thing that keeps the invaders at bay.

The KFS land was a major front during the Great Wars that occurred as the world started going weird. As a result, there are untold tons of military equipment, scrap, and unexploded materiel buried deep in the ground there (hell, maybe they're actually growing like potatoes; weirder shit has happened). When the Beast goes a-blizzarding, the freezing soil pushes these deposits up to the surface. The resulting "Iron Harvests" are a major trade commodity. The Universal Trade Affiliation (multinational corporations that control the vast majority of international trade) makes enough money by buying Iron Harvests from the KFS that they've agreed to make it their business to protect the borders, which is the other thing keeping the invaders at bay.

The linchpin in this situation is that the Elders of Omaha (currently the KFS's primary governing body) control the Beast through the Sacred Child, a young girl who can communicate with the Beast, whose location they keep secret. So the Elders control the Iron Harvests by controlling the blizzards and thaws, and thus have a chokehold not only on the invaders and UTA, but also on their own country.

The Elders want this situation to hold. The warlords want the blizzards and UTA mercs out of their way. The UTA want to trade the Iron Harvests on their own (more profitable) terms. Various KFS political groups want to be free of the Elders' tyranny. The family of the Sacred Child just wants their daughter back. What do you do?

No. 3 is the part where the GM plays NPCs, plays combat opponents, makes rulings, applies rules, etc. The GM's agenda here is to thus portray all this material as accurately as possible, with special attention to whether or not it makes an interesting playing board. He's not allowed to be invested in any particular outcomes (although he's allowed to cackle if the players get their asses kicked).

No. 4 is where the players decide what to do about (for instance) this whole KFS debacle. Will they aid any particular faction in return for special prizes? Will they return the child to her penniless family because It's The Right Thing To Do? Will they ransom the child against one (or some, or all) of the factions that would profit from control of the Beast? Will they just slay the Beast and let everyone scrabble violently to a new equilibrium? Will they say, "fuck these fuckers, where's the nearest dungeon?" Or something else?

No. 5 is not only ground-level resolution, but also meta-dungeon resolution. Whenever possible, the consequences of the players' meta-dungeon decisions should be expressed in the form of concrete impact on the Setting. Heads may roll, wars may be decided, cities may burn, governments may fall, nations may crumble. The players should be concerned about causing changes that serve their own purposes when possible, whatever those purposes may be. The GM should be primarily concerned with what outcomes he thinks make sense (and should give serious and perhaps even charitable consideration to whether the players' desired outcomes make sense). And also, choosing not to engage the situation is also a choice, so the GM has to think about what will happen if the PCs don't get involved (frex, radical groups exploit the child's bereaved family to foment open rebellion against the Elders, who turn the Beast's power against their own people, and all hell breaks loose in Kansas. NOW what do you do?).

Such consequences feed into the situation(s) that the GM preps for the next session(s). In the long term, the players may profit from cultivating relationships with various powers-that-be, or perhaps they won't profit (enough) and may decide to betray them. Or maybe things just get too hot or too boring here, and it's time to head west. Or anywhere in between; the point is, they have to strategize in the meta-dungeon for maximum continual gain. The players' success in the meta-dungeon long-game can be assessed by long-term prosperity and the achievement of various large-scale, Setting-changing goals.

This method, of course, requires rather a great deal of effort and creativity from the GM, which is why a major design-goal of Madlands is powerful and easy-to-use tools for meta-dungeon creation and management. But all that's still a work in progress, and naturally I'd appreciate thoughts on such things.

And of course, I'm not claiming that this is "better" than any other form of play (although, for me, it's the most exciting and interesting right now). Or that I've invented this: I see very similar things going on, from the D&D With Pornstars blog to Ron's proto-game Black Fire (both of which have been major influences on the idea), and I'm sure that someone, somewhere, has been paying like this since '79 or something.

Thoughts, questions, requests for clarification, etc?

  -   Marshall

*I don't doubt that some folks just plop the party at the entrance of a dungeon without any need for context; I've come close to doing it myself. Maybe tournament play back in the day was like that, I dunno.

[Eero Tuovinen]

Regarding "D&Dlikes", I've been flirting with the term "adventure game" myself. (Ron might remember me being curious about this term in Berlin in -06 or something like that.) There are good arguments both for (transparent and has a long history) and against (already used in many different contexts) it. It's clear to me that some sort of word is useful to have, anyway.

Anyway, this is how I've been structuring our recent D&D campaigns. My original motivation for it was theoretically very clear-cut: I wanted a procedure that would entail encounter balance and player buy-in. I'd found the traditional methods contemptible or hilariously bad - and what's worse, disinteresting. As it happens, the solution is simple when you create a "negotiation layer" (what Marshall calls a "meta-dungeon") for the campaign, a space where the players can talk about the next challenge in an interesting and fiction-relevant manner with the GM. The outcome is that instead of the GM having to wheedle the players into going into the next dungeon, the players themselves set up the boundary conditions (this is what we demand before we will agree to buy into this hazardous adventure) and approach vector (this is why we go, and what we're looking for in there) and then go into it voluntarily and higly motivated, and with interesting fictional positioning, too.

(In case there's still somebody reading these discussions who doesn't understand why I call this type of play "negotiation": imagine that you have a troll guarding the only ford over the river, and you tell about this to the players in advance. Your negotiation position is that you want the players to tackle the troll, because you want to play a troll and it's interesting; their position is that they only want to engage challenges that are interesting and have a fine balance of risk vs. reward for their particular player character. At this point you can sweeten the deal by introducing adventure hooks - reasons to cross the river - or edges - perhaps there are other people who also want to cross, and want to help the PCs defeat the troll - to get the players to agree to deal with this challenge. It is essentially a negotiation of the boundary conditions of the forthcoming adventure, despite the fact that we can largely veil the negotiation in amusing fictional detail. Only very rarely do we need to talk it out explicitly, but even when we do, it's not a problem as long as the negotiation layer of the game is clear and vivacious.)

In general I've found that we don't really have much to talk about with Marshal, simply because we agree so much about how to set up a rocking D&D campaign. He clearly knows what he's doing :D

[Ron Edwards]

Marshall, do employment, obligations, enforced service, and other things that result in assigned goals and destinations, count in your concept for this category? Is that a subset or a deal-breaker?

Note that I do not necessarily speak of railroading in those things. That would depend on the details.

[Eero Tuovinen]

Hopefully you don't mind if I answer for myself, too - I find this stuff interesting, and my solutions are of course brilliant.

In my campaign enforced obligations are very much possible. They may occur because the player wants a handicap ("Hey, it'll be cool if my new character starts play Charmed by the snake cult!"), or because of prior events in play ("Damn, I guess I'm lucky that the snake cult didn't kill me when they captured us.").

However, the players have an inalienable card to play: they cannot be forced to prosecute play under conditions they do not find interesting/gameable. When the fictional bargaining position becomes so inflexible that we cannot come to an agreement over the conditions under which a player would like to play ball, the negotiation has to be dropped and we have to go to a higher level of freedom to continue play. For instance, we might find that there is no viable way for a character charmed by a snake cult to go on an adventure that the player actually wants to tackle, in which case he has to make a new character so we can restart the negotiations from a freed-up position.

(It goes without saying that if we'd ever end up in a position where starting a completely new character without particular character-specific constraints still wouldn't help us negotiate an acceptable adventure, it's time to end the campaign. I could imagine something like this happening if e.g. a zombie apocalypse got seriously out of hand and the players really disliked playing in a post-apocalyptic milieu.)

When a player decides that something won't be played, the matter will either be left ambigously unresolved ("OK, your character disappears in the snake cult pit and is never seen again, unless you change your mind."), or it will be resolved with instant methodology - usually a dice roll. For example, if a character ended their last adventure stuck deep underground, but we didn't feel like playing the expedition back to the surface (maybe the GM doesn't have any good material for it, or we have something else we want to be playing right now), I might offer the player a 50/50 roll, either die an obscure and unmourned death or get back to the surface in time to join this adventure. Up to the player to take or leave this, of course, but if he doesn't take it, that particular character won't be joining this adventure.

In summation, I find that although individual characters may easily and often end up in situations where their supposed "freedom" in the sandbox is limited in various ways, that does not and cannot actually limit our freedom as players: we can talk to each other about what we want to actually spend our time doing, and anything that is not interesting can be resolved so quickly and arbitrarily (if we're not interested in it, that entails that we can resolve it quickly and arbitrarily) so we can get to the good stuff. Thus character positioning that limits player options can be used to the maximum extent as a fun strategic-level challenge, but put aside when it no longer serves the core goal of discovering and facing interesting challenges.

[Marshall Burns]

I might not be clear on what you mean, Ron, but if I were running the game, any such service would be entered either voluntarily by the player or as a negative consequence of player decisions (e.g. you fail to fight off some soldiers, get knocked out, and wake up impressed into military service), and the player could leave any time he thought he could get away with it. They can renege on deals, breach contracts, abandon posts, and break oaths all they want. There would probably be consequences, but that's the player's risk to take.

Which isn't to say that, for instance, "okay, you're all vassals of the Duke of Standcastle, that's the game we're playing" is wrong or non-functional; it's just not particularly what I'm getting at. It could certainly be considered a subset.

And now I see what Eero just wrote, and I think I'm pretty much down with that angle on it too. With the added detail that I'm all about multiple characters per player, and the possibility of retiring (permanently or temporarily) them to focus on other characters who represent different vectors and possibilities of play that the player is interested in today.

[Mike Holmes]

Yeah, I think that there's clearly a knee-jerk reaction to any use of GM force that is a reaction to railroading and/or illusionism. But, in fact, some of it has always been mandated. If the players go down the left corridor in a pre-set dungeon, and they find a room full of orcs, that's the consequence for the decision in question, not the GM trying to railroad an encounter with orcs.

I got the solution to the problem of "when is force railroading?" from Ron's definition of "Bangs" from Sorcerer years ago. GM use of force is abhorrent to players if/when it reduces player agency. If a particular use of force increases agency, it's seen as a good thing.

So in the case of the orcs, what the GM has done is given the players a situation in which they have to decide to retreat, fight, sneak, or otherwise deal with the orcs. If they do fight, then the decision-tree in D&D flowers large. Good use of force.

If late in the fight with the orcs, the PCs are losing, and a mighty NPC ally shows up just in the nick of time to save their bacon... then all of that decision-making during the fight has been nullified. Bas use of force.

If you force the PCs to take jobs in a caravan, be sure that this immediately provides them with some decisions to be made. Interesting decisions, note, not just the "window dressing" of what to say in accepting the job. And also be ready for them to say no to the caravan job, even if the decision has dire circumstances that you threaten to impose. Meaning that this decision should be an interesting outcome with more new interesting decisions.

A seeming decision that has only one possible outcome is not a decision that's interesting for players to make. Use GM force to introduce new interesting decisions.

This is a simple concept, but one that's often difficult to execute upon. Because what you might think is an interesting decision may not seem as interesting to your players. So there's a skill you have to develop. But what you are doing is at least setting up a baseline where there's a CHANCE that what you're throwing at the players is interesting.


I think what you have here sounds like a relatively interesting version of D&D, Marshall. Which means it's not something that I would enjoy a lot, but folks who like D&D should love this.

[Marshall Burns]

Regarding the issue of Force, I think Apocalypse World provides the best applicable methodology I can think of: if you Announce Future Badness and the players don't do anything to evade or prevent it, or if the players botch one of their own moves, you make a Hard Move, which if I understand you, Mike, is the same thing you mean by using Force to arrive at an interesting decision point, yeah? And of course in most D&Dalikes, being in a dungeon IS an announcement of future badness.

[Mike Holmes]

No, I think that's possibly a different thing, actually, Marshall. What it sounds to me like you're talking about is giving teeth to consequences of actions. This is ALSO very important. Basically it's like the example I gave with the orc fight. If you save them last second, then was there really any threat at all? Will there be any tension the next time they decide to dive into a fight with orcs, knowing that if it turns out to have been a bad decision that you'll just pull them out of the frying pan?

Hard moves CAN lead to interesting decisions, and should where possible. But in and of themselves, they represent actions being consequential. Which means that the decisions that lead to them are meaningful. Note that this is why the GM is prohibited from making a hard move until a player makes a decision that would lead to it.

A problem with D&D is that it makes every combat infinitely meaningful by having permanent character "deprotagonization" (death in this case*) be a real possibility. You can get by with less meaning than that as potential outcomes. Hence why 4E's design that makes death unlikely is an improvement over earlier editions, IMO.

BTW, a good way I've found to make D&D combats consequential, without having to have them be particularly lethal, is to have some permanent injury rules. Better yet is to have something really interesting at stake for every fight other than life or death. But when it's just another combat in the encounter chain and we can't have the damsel be at risk in each one of them, risking permanent injury or scarring etc, is a great way to make each fight seem more meaningful. "I remember that orc that took my left little toe! Surly bastard he was. Ten feet tall, I swear..."


GM Flowchart
1. Use force to set up a (real) decision.
2. Player makes decision.
3. GM rolls with decision, uses force to make it appropriately consequential.
4. Go to 1, using the outcome of 3 to determine in what way to use force.
(Note that sometimes people prefer the use of the term "authority" to "force," since the ability the GM has to create situation, etc. is given to him by social contract with the rest of the group.)


*I've personally given up on having characters get "rezzed" in TTRPGs. It feels like I'm playing a ghost when that happens, and I immediately start looking for a way to finish the character's story when it happens. Better to just leave them dead, like some very existential commentary on the meaninglessness of life, than to bring them back. Might just be me.

[Mike Holmes]

Whoops, I focused on the Hard Move part, when you meant the Announce Future Badness part. To which I say, done right, Announcing Future Badness can create a decision. Meaning you have to actually present it in a way that the players have a choice other than just the "Avoid The Badness" move. If you will. But I think that's just one limited methodology and that situation will dictate which is the best method.

One thing I don't like about AW is how it's channeling people's understanding of play into narrow troughs. Yeah, it's easy, and yeah, it's great for newbs to our style of play. But I like my description better. :)

[Marshall Burns]

Right, I was talking about how Announce Future Badness (and also rolling poorly) provides an excuse (for the GM) and expectation (for the player) for the Hard Move. It's a technique to help ensure that the GM's Hard Moves are accepted by the players, thus passing smoothly through the Lumpley gauntlet so that play proceeds without dysfunction. For counter-example of another method of Hard Moves that _isn't_ usually accepted (but may be at some table, somewhere) by players is the Gotcha Game infamously played by some GMs, where the GM subtly coaxes the players into position so he can spring his Hard Move on them, or he suddenly uses situations that up to now have always been safe and SOP to spring Hard Moves (player: "I listen at the door." GM: "Gotcha! An oakwood brainworm was lurking in the door. It slithers into your ear and eats your brain. You die.").

When the Hard Moves aren't properly accepted, you end up with dysfunction such as disruptionism and extreme turtling. When they are, you get a vibrant arena for Creative Agenda to be pursued. In Gamism in particular, there's another piece to this besides how Hard Moves are introduced; it's all about how losing and consequences work. Somewhere I have a Philosophy of Losing that I wrote up for Madlands; I'll try to find it and post it.

Oh, also! The terminology of "Hard Move" and "Announce Future Badness" are of course from Apocalypse World, but those techniques are positively ancient. AW definitely provides the best language for them, though, which is laudable and appreciated.

And another also: on G+, I posted a snippet of the OP here with a link. Gregor Vuga, who isn't registered here, replied and voiced his difficulty with no. 3 from the list in the OP. I want to repost what he wrote here, along with my response. I'm waiting for his permission to do so.

[Ron Edwards]

I can see where some terminology is causing some trouble, just like it did for Josh Roby years ago. Force isn't synonymous with "GM now affects fiction drastically." People always read it that way and I think it's because that particular act is so fraught that they think I must be talking about it. Mike, some of your uses of it in the above posts grade into that territory, and without looking at it yet, I'll bet a dollar to a donut* that he's thinking about Force in this way.

Best, Ron

* That phrase isn't what it used to be, is it?

[Mike Holmes]

Right Ron, that's why I put in the parenthetical about Force possibly being better called authority. Or at least hoping that it would be instructive in this way.

But OK, yeah. GM Force is the GM doing things using the mandate they have been given, which is traditionally to "play the world and NPCs." Which is no different than a player dictating what their character intends to do. Call that "Player Force" if one wants a parallel. Or, again Player Authority.

Yes, "Gotcha Games" are an example of a GM using their authority to take away decisions from the players by not giving them enough information on which to make a decision. This is a very fine line, in a lot of cases, and becomes problematic when the GM assumes that the player will not take the interesting BAD angle. And so are forcing players to take the bad road accidentally, instead of in an informed way. There are much better ways to do this. One is just give the player both options and make the bad one attractive to the character. The guy playing Walter White in the (theoretical) Breaking Bad RPG has Walter take the bad choice just a about every time. Another method is just to provide two bad decisions, and let the player choose from both.

But yes, if the "choice" you're offering is between something that the character would not do if well-informed, and something they would do, and you can only get them to do it by making them ill-informed, that's pretty bad technique, IMO.

So, that all said, if you want a character to make a bad move because they're uninformed, then just frame to them having made that bad move. Just be sure that the player is OK with it (ask if necessary). But as long as such a GM move leads to a cool decision to make, most players are OK with it.

For instance, if you know that a character has a gambling problem, instead of asking the player if the character will play against another NPC - and NPC that turns out to be a real shark who can't lose - simply frame to, "Your character has, once again, lost his life's savings, this time to a shark named Frankie, who may have been cheating. The game is over, and Frankie is laughing at your character. What does your character do, keeping in mind he does have his gun on him?" A stilted example that would flesh out in play better, but you get the idea. Don't ask the player to fall into your trap with Frankie, and feel dumb for having done so because you didn't give the player all the info. Just frame to the interesting choice of how to deal with Frankie after the fact.

By using techniques like this well, you can see how nobody is opposed to GM's being heavy-handed. What they're opposed to is GM's failing to give players interesting decisions by making them all for them.

(And, of course, opposed to accidental "deprotagonization." You have to watch making any decision for a character as it's very easy to accidentally ruin the character for the player. But that's another subject.)