My Framework for GMing Dungeon World

I've been working on the GMing chapters for Stonetop, and it's made me think about how the conversation really flows. I've also been thinking about GM moves, and Principles and Agenda, and how they all work together. I thought I'd talk about them a little here. 

I don't think that what follows is fundamentally different than what the game text tells you to do. Like, if you read the DW text and the DW Guide, and GM the game enough, I think you end up doing what I describe below. This is just how I conceptualize it, with 8+ years of experience running, playing, and talking about DW and similar games. (This is also pretty similar to what I describe here.)

Maybe you'll find it useful? Maybe a new GM will find this and something will click for them. Regardless, I'm going to be posting some excerpts from the Stonetop GMing chapters over the next couple weeks, and I think this will help set the stage. 

As always, feedback and questions are appreciated!

The Game is a Conversation

You say something. The players say something. You say something in response. You ask questions of each other, clarify, interrupt, talk over each other. To quote Vincent Baker: "you take turns, but it’s not like taking turns, right?"

The whole point of this conversation is to create the fiction, the shared imaginary space that we're all talking about, where the PCs and NPCs and monsters all exist and act. The game's rules mediate the conversation, and help us figure out what happens when there's uncertainty, and help introduce unexpected and challenging elements into the game.   

Dungeon World is (despite what some folks will tell you) a rather traditional RPG. It structures the conversation and assigns responsibilities and authority in a very familiar way: 

• The players are responsible for portraying their characters (who they are, their actions, their thoughts and opinions, their experiences and backstory). 
• The GM is responsible for portraying everything else: the world, the NPCs, the monsters, etc. 

Dungeon World is different than a lot of RPGs because it explicitly encourages (and arguably requires) the GM to ask the players for input on the world, particularly during the first session and about areas where their characters have experience or expertise. But that's not that different than how lots of folks play D&D. (Ever done a session 0 where you make characters, talk about backstory and the kind of setting you want to play in? It's like that, but it's done during play.)

Different groups take this collaborative spirit to different degrees. Some DW GMs are very cognizant of The Line (I know I am) and avoid asking the players to make up details about what their characters are experiencing on the spot. Other GMs will actively ask the players to make up details about the room they've just entered, or the NPC they've just met, or what happens next. Some groups collaborate on where the story should go, and what kinds of scenes they want to have. None of these are wrong. They're just a matter of taste.

The Structure of the Conversation

Here's how the conversation generally goes. Stuff with black outline is stuff you say (or at least facilitate), as the GM. Stuff with a purple outline is stuff the players are saying/doing.

click and zoom for details!

Side note: this is similar to Adam Koebel's "how to play DW" flowchart, but I think this represents more of the conversation that happens during play. 

First: frame a scene. Say (or ask) who's present. Say (or ask) where are they. Say (or ask) when is the scene happening.  Say (or ask) what they'e doing.  Give some impressions. If you're not sure what scene to frame, or how to frame it, then ask questions until you are. 

Once the scene is framed, you establish the situation. Give (or ask for) details and specifics, enough to visualize what's going on but not so much that player's eyes gloss over. In an action scene, establish momentum and relative positions. Tell the characters what they see/hear/sense. Ask questions. Answer questions from the players and clarify what's going on.

Now, make a "soft" move. In my opinion, a soft GM move is when you say something in order to:

• Provoke action/reaction from the PCs
  and/or
• Raise the stakes/tension in the scene

That often means saying that something bad is about to happen, or is in the process of happening, but the PCs have a chance to do something about it. But it also could mean that you're presenting them with the obvious choices and prompting them to pick. Or that things were previously calm and peaceful and now there's trouble. Or that you're giving them an opportunity to act and seize the initiative. Or or or.

It could also just be the slow turning of screws: their torch is getting lower, their food is getting used up, the storm is getting closer, the fire is spreading. 

SIDE NOTE:  Dungeon World's text defines a soft move as "one without immediate, irrevocable consequences.... [Either] something not all that bad... [or] something bad, but they have time to avoid it." And it defines a "hard" move as having "immediate consequences." 

I've never found those definitions satisfactory. The difference between them ends up being differences of degrees. You can almost always find an "immediate consequence" in whatever soft move is presented, and supposedly "hard" moves like deal damage aren't all that bad if the damage die is low and the PC's HP are high.  

I think it's much more useful to think of a soft move as "provoke action or raise the stakes" and a hard move as "establish badness."  There's still overlap between soft and hard moves with these definitions, but they're more actionable. It's much easier to look at the scene and think "how can I provoke action or crank up tension?" than it is to look at the scene at think "how I can say something bad-but-not-too-bad or something that threatens badness but gives them a chance to escape it? 

I'm sure that my definition isn't perfect, either. But it makes more sense to me!

My list of GM moves are similar to those in Dungeon World but slightly different. They are:

• Announce trouble (future or off-screen)
• Reveal an unwelcome truth
• Ask a provocative question
• Put someone in a spot
• Use up their resources
• Hurt someone
• Separate them
• Capture someone
• Turn their move back on them
• Demonstrate a downside
• Offer an opportunity (with or without a cost)
• Tell them the consequences/requirements (then ask)
• Advance a countdown or grim portent

And the "Exploration" GM moves (which replace the "Dungeon" GM moves) are:

• Change the environment
• Provide a choice of paths
• Bar the way; make them backtrack
• Hint at more than meets the eye
• Present a discovery
• Point to a looming danger
• Introduce a danger, person, or faction
• Offer riches at a price

I do NOT think it's important to intentionally pick a move off these lists. I think the point of having a list of GM move is to give you inspiration when you're stuck, or to inspire you to say something different than what you'd say naturally. When you make any given GM move in an actual game, you can often retroactively match it to more than one of these moves. That's fine. It doesn't really matter which of these moves you're making, as long as you're provoking action/reaction and/or raising the stakes/tension.

Your moves should follow your principles. They should begin and end with the fiction, and you should address the characters, not the players and you shouldn't say the name's move. Etc. etc. (More on that below.)

After you make a soft GM move, ask the player(s) "What do you do?" 

"What do you do?" is a ritual phrase. Like, you know how when you're playing traditional D&D and the GM says "roll for initiative," and everyone sits up and starts paying attention? It's like that, but way more frequent. It's the GM's way of saying "your turn." This question drives the back-and-forth between GM and player, and it's a touchstone of good PbtA GMing.

Okay, the conversational ball is in their court. Now what?

If they ask questions--about the fiction, about the rules, about what you just said--then answer them, honestly, generously, and enthusiastically. Then put the ball back in their court: "So, what do you do?"  If they ask something that wouldn't be immediately obvious, tell them so and what they'd need to do in order to get the answer. Then: "What do you do?" (or "Do you?")

If they do something that triggers a player move, like Spout Lore or Discern Realities or Hack and Slash or Defy Danger or whatever, then resolve the move. Begin and end with the fiction.

On a 7+, do what the move says and establish how the situation has changed. If the move prompts you to add some detail or action to the fictional situation, you can use the list of GM moves for inspiration. For example, if they trigger Hack and Slash and get a 7-9 and "suffer the enemy's attack," then their attack succeeds but they also suffer the enemy's attack. That attack can be any GM move, from using up their resources to hurting them to capturing someone to whatever.

On a 6-, then they mark XP and you make a hard move.

If they do something that does NOT trigger a player move, then they're looking to you to see what happens.

• If they ignored a threat, did something stupid, or did something with obvious bad consequences, then you make a hard move.| 
• Otherwise, you just say what happens as a result. 

Making a hard GM move means: establish badness. Say that something bad happens, or make us aware of something bad that happened in the past.  Use the list of GM moves for guidance and inspiration, but, again, you don't need to intentionally pick from the list.

Regardless of what they did and how it resolved: ask yourself, is the scene still going?  If so, say how the situation has changed (and recap the situation if appropriate), then go back to making a soft GM, asking "what do you do?" and resolving their action.  Keep doing this until the scene ends.

When the scene is over: wrap it up. Take care of any bookkeeping (using bandages/poultices or the Recover move in Homebrew World/Stonetop; erasing "hold," figuring out how much time has passed, etc.). Have any meta-discussions you need to have as a group, like:

• Do we want to keep playing? Or wrap up for the night?
• What should we do next? 
• What are we trying to accomplish, again?
• How far is it back to __?
• Etc. etc.

Figure that stuff out, then frame the next scene. If you're not sure where or how to frame the next scene, ask questions until you are.

To summarize:

1. Frame a scene, if you haven't already
2. Describe the situation
3. Make a soft GM move (provoke action, raise tension/stakes)
4. "What do you do?"
1. If they ask questions: Answer, clarify the situation, back to "What do you do?"
2. If they trigger a move: 
1. On a 6-, make a hard GM move (establish badness)
2. On a 7+, do what the move says to do!
3. If they don't trigger a move:
1. Did they ignore a threat? make a hard GM move (establish badness)
2. Otherwise: say what happens
5. Scene still going?  Return to #2.
6. Scene over? Wrap it up, take care of bookkeeping & meta-talk. Return to #1.

The Spotlight

As you have this conversation, you'll change who you're addressing, moment to moment. This is often called "moving the spotlight" or "pointing the spotlight." Whoever you're talking to right now is in the spotlight.  

Sometimes you'll keep the spotlight unfocused and address the group as a whole. Sometimes you'll focus it on just one character.  When lots of action is happening simultaneously, you swing it back and forth between individuals and groups.

There's no formal process for managing the spotlight, just like there aren't formal rules for most conversations. The most obvious time to move the spotlight is after you resolve a character's action. Before you describe/summarize the situation, address a different character and describe the situation to them. Make a move at them. Ask them, "What do you do?"

But you can technically move the spotlight at almost any point in the conversation. For example:

• You ask "What do you do?" and they freeze or stall; shift to someone else, give them time to think, then come back to them. 
• You ask "What do you do?" and they say that they're going to do something that takes time; shift to someone else and come back to them as they finish their action (or as something interrupts it).
• You ask "What do you do?" and they describe an action, and it triggers a player move, and they roll. You're not sure how to resolve the results, so you stall for time by shifting to someone else for a bit, then come back and resolve the first player's move.

Regardless of how and when you move the spotlight, be a good facilitator. Try to keep everyone involved. Make sure everyone gets some good screen time and has a chance to contribute. It's generally okay if players jump and interrupt each other, or kibitz, or have their characters act "out of order."  Heck, sometimes a move (like Defend) explicitly gives them permission to do so.  

With that said: don't be afraid to shut down an overly eager or aggressive player with a polite-yet-firm reprimand. “Andrew, you’re being rude. I’m talking to Jamie right now.” 

Principles and Agenda

The Dungeon World text (and almost every PbtA text that I've read) actually leads off with a GM's agenda and principles, before talking about GM moves and how to make them.

I get it. The principles and agenda are important. They inform what you're saying and doing during the conversation. But without understanding the structure of the conversation, they're just like a bunch of Zen-koans that can make the whole processes of GMing really intimidating. 

Here's the thing: your agenda is just what you're working towards. Your principles are "best practices" for getting there.

For Dungeon World (and Homebrew World), your agenda items are:

• Portray a fantastic world
• Fill the characters’ lives with adventure
• Play to find out what happens

Basically: it's your job and responsibility to portray the world, and it should be a fantastic world, filled with magic and monsters and all that jazz.  It's also your job to keep things exciting, and give the players interesting, dangerous, exciting stuff to do. And finally, it's not your job to decide what happens in advance.  

That last agenda item--play to find out what happens--is, I think, the most important one. It's the one that shows up unchanged in almost every PbtA game that I can think of. It covers a huge range of approaches. Some GMs take it to mean: "do almost no prep, ask a bunch of questions, and improvise everything with the players."  Other GMs take it to mean "prepare interesting situations--tenuous and unstable, dynamic and fraught--and see what happens when the PCs interact with them, following the dice and the PC's decisions, respecting your prep and the integrity of the fiction."  Both approaches are valid and great. So is just about everything between them. What's truly important is that you avoid forcing the game into your pre-established storyline or your expectations of how things are "supposed" to go. 

Side note: the agenda items for Stonetop are...

• Portray a rich and mysterious world
• Punctuate the PC's lives with adventure
• Play to find out what happens  

The differences are subtle, but important for Stonetop. It's much more grounded setting, and while there are fantastic elements, the mundane elements are equally important. And while the bulk of play focuses on the PCs' adventures, those adventures are interruptions to their day-to-day lives. 

"Play to find out what happens" is, notably, unchanged.  

Now, what about the principles? The principles are just your guidelines, your best practices, the things you should strive for. 

Some of them are really quite easy to do. Like, these principles are just establishing protocols for your part of the conversation. They're pretty easy to do:

• Address the characters, not the players
• Make a move that follows
• Never speak the name of your move

Basically: talk to the characters in second person, rather than talking about the characters in the third person.  Don't make wacky shit happen just because you can. Don't announce the name of your move when you make it, because that's dumb and sounds weird and doesn't add any value. Once you've internalized them, these principles are easy easy to follow. 

Another set of principles is basically just "things that will help you portray a fantastic world:"

• Draw maps, leave blanks
• Embrace the fantastic
• Give every monster life
• Name every person
• Think offscreen, too

These take a little effort, and represent a mindset, but they're basically just good advice. Draw maps, and prep a bit, but don't go nuts--leave yourself space to improvise and be surprised. Be cool with fantasy tropes, make your monsters more than just numbers, and try to make your PCs's interesting and memorable and actual people. Think about what's going on offscreen, and how that might come into play onscreen, because this is supposed to be a whole world that we're playing around in and not just a little bubble around the PCs. 

If you fail at any one of these principles, the game won't, like crumble. It'll just be a little flat.

Then there's the last set, the ones that I think are critical for running an excellent game of DW. They're also the hardest to do consistently and well. They sometimes conflict with each other.

• Ask questions and use the answers
• Be a fan of the characters
• Think dangerous
• Begin and end with the fiction

Like, thinking back over the years of DW-related conversations, these four points are the core of most GM's struggles. 

Ask questions and use the answers is one of the most radical things about DW (and PbtA games in general), because it pushes you to collaborate with the players in a way that D&D and other more-traditional RPGs don't. There's a real art to this: knowing where to ask questions and what questions to ask; how to phrase the questions to get interesting results without letting the players stomp all over your prep; recognizing which players enjoy this sort of thing and which ones don't; changing the nature of the questions based on the player or the amount of prep you've done or how much the world is established.  It's not easy.

Be a fan of the characters is fucking crucial, but it also bumps up against all sorts of GM instincts. Like, here's my favorite recent example (from reddit/r/dungeonworld):

My player's bard has a dinner date with the bad guy. How do I not reveal who he's secretly working for if the bard decides to be Charming and Open? 

The scenario: This is my first time running a game and I'm running an intrigue-based campaign. One of the party's contacts/quest-givers is a demon, disguised as an affable & handsome gentleman, whose goal is to further the designs of the city's overlords. He's trying to get the party involved in a scheme he's running. The rest of the party is interested in the money/items he's offering to get them involved, but the bard said he was only interested in dinner with quest-giver/demon. Since the demon is a bit of a flirt and is happy not to part with his items, he, of course, agreed to this. 

As the DM, now I'm realizing that my player may be thinking of using their dinner as a time to be Charming and Open and ask the demon who he's working for. I should have seen this coming, but I am not a smart man, so I didn't. Is there any way I can keep the intrigue going if the bard decides to ask who he's serving after being charming and open on a lovely dinner date?

The answer (which pretty much everyone gave) is "be a fan of the character" and let them enjoy the benefits of their move.  Don't be precious with your secrets and your storylines, and play to find out what happens.  This GM took the community's advice to heart and ran with it and was delighted with the results, but the whole episode speaks to how difficult it can be let go and let the PCs be the badasses that the game wants them to be.

At the same time, "be a fan of the characters" can also be used to justify just giving the PCs everything they want, or letting them walk over the opposition, or act without consequences. It is, I think, the principle that most often conflicts with other principles--especially the next two. 

Think dangerous is the DW version of my favorite principle from Apocalypse World: "Look through crosshairs." Basically: don't protect your NPCs, your monsters, your institutions, or the status quo.  In Stonetop, I call this "let things burn." 

Regardless of how you word it, I think this concept is crucial to making the world wonderous (or rich and mysterious), filling/punctuating their lives with adventure, and playing to find out. It's a mindset that requires effort and intention to develop. It's basically a discipline of non-attachment, applied to the fictional world you're creating. 

It's hard. Pretty much everyone has an instinct, a very natural instinct, to preserve their darlings, and preserve the player's darlings (because you want to be a fan of them, right?).  But you have to threaten the things the PCs care about, and then be willing to follow through on your threats. You have to let the PC's wreck your shit and upend the status quo. You have to let them one-shot the dragon, if the fiction and their moves and the dice all say that should happen.

Anyone who's done serious fiction writing will tell you that you have to "kill your darlings."  But if they're honest, they'll also tell you that this requires an act of will.  

Okay, finally, we've got this guy: begin and end with the fiction. There's a ton say about this, but this principle is basically what turns this...

“You attack the ogre? Cool, roll Hack and Slash. A 10? Do you evade its attack or deal extra damage? Okay, roll your damage +1d6. 7 damage? It’s still up, and it hits you back for 1d10+3 forceful, knocking you down. It's going to attack you again. What do you do?”    

...into this:

“You attack the ogre? Cool, what’s that look like? Stabbing upward into its gut? Yeah, sure, roll Hack and Slash. A 10+? Do you evade its attack or deal extra damage?  Okay, roll your damage +1d6. 7 damage? Okay, so like you said, you like stab up into its gut and it goes in but not all the way, and the ogre like doubles over, howling in pain, but before you can get away it just uncoils and backhands you across the face. Take 1d10+3 damage as you go flying and land in a heap, head spinning. You hear it lumbering towards you, grunting in pain and anger. What do you do?” 

This isn't just an issue of style or fancy language. Without establishing specific fictional details, the game starts to break down. A player says they do something that doesn't make sense based on what you're picturing. Or they do something that they thought was simple and safe but you think is Defying Danger, and now they're salty about it. Or you invoke a player move, and they roll, and you aren't sure how to resolve it because the details aren't there.  

Beginning and ending with the fiction is how you decide what to do when the rules aren't clear. It's how you determine which PC actions are possible, and what moves they trigger (if any). It's how you decide on a GM move to make. It's how you keep everyone on the same page. It's arguably the single most important thing to do as the GM.

And it is hard. You have to juggle dozens of inputs, from the players, the dice, the moves, your prep. You have be able to visualize a fictional world and think through how it would react to different inputs. You have to communicate that in a way that is clear and accessible and evocative to your players, without overwhelming them or boring them, usually with only your words and your gestures and maybe some crude drawings or props.  You have to be able to inhabit NPCs and monsters and portray them, making them do things that make sense based on their instincts, wants, needs, knowledge, perceptions.  

Oh, and you have to decide when to ignore (or change) previously established fiction in order to support other principles (like be a fan of the characters or ask questions and build on the answers). You have to learn what sorts of things to prep and what to improvise, based on your own skills and abilities and weaknesses.   

Some GMs are naturally gifted at working with the fiction. Others really struggle with it. Some are gifted in one area but weak in other (I myself can run action scenes pretty effortlessly, but struggle with compelling NPCs).  I maintain that this is a skill that can be developed, worked on, improved. 

Which brings me to...

GMing is a Practice

I'm quoting myself, but:

After each game, think back on the decisions you made, the things you decided to say. Run those things against the game's proscribed agenda. Did you say or do anything that violated the agenda? Try to avoid that next time.

Look at the principles.  Did you say or do anything that violated them?  Think about what you could have done instead. Think about what adhering to that principle might have looked like.

Look at the GM moves. Think about your major decisions, the things you said to prompt action from the PCs or to raise stakes/tension. Can you match each of those things to one or more of the GM moves?  Were there any decisions you made, where you could have done one of these other GM moves instead? Keep that all in mind for next time.

The GM's agenda, principles, and moves are just ways to codify and describe good GMing.  Some GMs adhere to them closely and intentionally make their moves from the lists. Some GMs keep the principles constantly in mind.

But the core loop of the game is this:  Describe the situation. Give the players something to respond to. "What do you do?"  Resolve a player move or say what happens. Repeat.

And then look back on your work and see how you could have done better.

GMing is a practice, like yoga or martial arts or meditation or painting or whatever. You get better at it by doing it, by reflecting on it, by constantly trying to do better.  No one starts off as a maestro. Don't be afraid of being bad or mediocre or less than excellent.  Do the work. Show up. Get better. Get good. Get great.