Designing GM-directed PbtA

Hi everyone! I just had my very first playtest session for a hard scifi PbtA game I am putting together. One of the features of the game is that it is a more classically GM’ed game where the GM comes has a scenario in mind and the players interact with it (the “GM-directed” experience – for the lack of a better word). The players may be invited to collaborate and contribute to the fiction “writer-room style” (“Looks like you flubbed the Assess roll. What do you think jumps out at you from the darkness?”), but the GM may very well have something in mind about what is happening behind the scenes (“You could have sworn you saw a shadow moving out of the corner of your eye, but when you pan your flashlight over, there’s nothing there.”).

Currently, I’m using the “choose from a list of questions” approach for moves like Assess the Situation. One of the comments I received from the playtesters is that sometimes questions led to uninteresting answers (ex: “Who is in control here?” in an Alien-style scenario led to “You are in control here.” because I knew that the alien had already moved on.). One possibility we discussed was that the PbtA “list of questions” is better suited to improv / semi-collaborative play, since you can always choose a question that had an interesting answer (because the GM could always turn it back to you for the answer), and perhaps it was not as well suited for GM-directed games.

I am curious what others think about this. Have you had run more GM-directed PbtA games before? Or know of PbtA games that were more focused on a GM-directed experience? I ran The Sprawl as a pretty GM-directed game before, and also recall instances of “QA let-down” there too. I’m now wondering if I’m bucking the trend or plainly swimming upstream with this idea.


For moves like this I’ve opted for as general an approach as possible that does away with the list of questions and replaces them with ‘ask a question about the situation, the scene, or someone/something in it’.

1 Like

Thanks @Alun_R. The playtesters also suggested variations of this. Ironically, this actually makes the moves simpler to write!


Does this approach actually solve the problem of “Some questions have uninteresting answers” though? I’ve played lots of ‘trad’ games using the “roll a knowledge skill and the GM will tell you something you found out” sort of approach, and a lot of the time, the answer is “Nothing interesting.”

I feel like this is a flaw in GM-led scenarios that you just need to accept – if the scenario is ‘predetermined’ some questions that you might want answers to will have boring answers. Because the GM CAN’T just “turn it back on you” in the name of making things interesting. This isn’t solved by broadening the list of questions. If anything, a bigger list makes it harder, because a GM with a known list of questions can try to prepare interesting answers for them, while a GM who has to answer whatever the players come up with cannot.

I’d say that rather than use an ‘ask some question about what’s here’ methodology and hope the players ask for something interesting, I’d instead hone your GM processes/prep so that there is more likely to be an interesting answer to the short list.


For what it’s worth, I think the default in most Gmed PbtA games is a pretty traditional GM/player division of power, and works basically as you describe it in your game. There are exceptions (Fellowship comes to mind), that explicitly give players power over the world. I’d say in particular games encourage world building in session 0 to be very collaborative but in general, once play starts, the line between GM and players is quite clear. Players only get to step over it when invited by the GM. There are places where the play culture is for the GM to invite them quite frequently (the Gauntlet is one of those places!), but it’s not inherent to the games in any way. See, for example, John Harper’s essay on “The Line” between GM and players in Apocalypse world: Obviously, not everybody would agree with him about “the line,” and when the GM can cross it, but importantly, the game never suggests that players cross it either. I’d argue that most of the time, PbtA games are “GM directed” in the sense that you mean.

As @Airk says, this doesn’t seem like a PbtA problem, it’s just an issue with GMed RPGs. The answer to a question can be boring whether the GM has a fully prepped location or not, and whether or not there are a fixed set of questions that PCs can ask. Different games solve this problem in different ways (or fail to.) Trad and OSR games often solve it through prep, random tables or pre-written modules that give GMs something to turn to in order to find potentially interesting answers when questions come up.

The classic PbtA solution is GM moves. Most times the GM opens their mouth in a PbtA game they should be making a GM move that follows their principles and agenda. A PC asks a question? “Reveal an unwecome truth” or maybe “present an opportunity, with or without a cost.” Well written Assess type moves in PbtA games generally have questions that point GMs back at their moves, so that they are easy to answer in the context of the game.

All of that is to say, I don’t think you’re swimming upstream, I think you’re designing a pretty standard PbtA game in terms of GM authority. But I also wonder if you mean something different when you say “GM directed.” It sounds more like you’re talking about the kind of prep the GM is expected to do, rather than the kind of authority they have. It’s hard to say without more detail about the game, but if you’re having this problem, I’d suggest you take a look at how you prep the game (or how the game tells GMs to prep it, if you’re that far along in writing), and what the GM moves are, and how those pieces interact. If GMs are coming up empty on questions, the issue is probably somewhere in there.


I do get the sense that @andante is probably talking about GM prep vs GM “authority”, but I don’t think there’s a “PbtA standard” for what kind of prep a GM should do for a PbtA game. (Not that there are lots of “PbtA standards” for other things) The phrase “GM comes has a scenario in mind and the players interact with it” leads me that way.

That said, I don’t think that’s… super nonstandard? Though most PbtA games would use the term “Situation” rather than “Scenario” because I think most gamers think of “scenario” as a thing you “run through” whereas a “situation” is something you… interact with? Anyway, while PbtA has this… reputation for not being “GM lead” or whatever, even freaking Apocalypse World tells you that after your first session (where you are basically just supposed to be poking the characters where it hurts to see what happens) you’re supposed to be creating a bunch of threats. Maybe this isn’t a “GM lead scenario” where there’s a “right solution” but it’s not exactly a “prep nothing, there is no GM setup” environment either.

So yeah. I’m starting to wonder if maybe there’s a misunderstanding here somewhere.


If we’re talking specifically about “reading moves” (moves that allow you to ask questions), there’s also entirely nothing wrong with saying that they’re just not fun or not appropriate to the game you’re putting together.

There are lots of great PbtA games that don’t have any “reading moves” - even early on in the development of PbtA games (e.g. Monsterhearts). Don’t be afraid to ditch bits of the system that aren’t giving you what you want, or rewrite them in a way which has interesting stakes, so that it doesn’t feel like an inconsequential move.


Thanks everyone. That’s a lot of great feedback. I’m glad to know that making a more GM-centric / prep-heavy game isn’t unusual.

To answer the question of authority / prep, the game I am designing is meant to have more GM authority (i.e. players generally do not introduce their own setting facts beyond a specific scope). But the session I prepped is more highly prepped (I haven’t gotten so far to decide how prep should be done in the game). I tend to prep a history of events and then allow the players to find bits and pieces of it through play, and they can piece together the original timeline.

The issue of “uninteresting answers” may also be confounded with the feedback that the playtesters didn’t like having to sort through questions. That when they have to choose from the list, they feel they are taken out of the moment in their desire to find a question that would reveal the most information. I recall feeling that way when I played other games, and seeing that “puzzle solving mode” switch on when I run other games too. I always viewed it as a desirable bits of strategic play. and I let the players “bend” the questions slightly. How closely to other GMs “stick” to the questions as written? Do you see the “puzzle mode” to be desirable?

I don’t know that I “bend the questions” in any meaningful way. In well designed games that have these question lists, I generally find that the questions to ask in any given situation is obvious, and the bigger issue is that on a 10+, if you get to ask a bunch of questions, players have more than the want/need. If questions need to be bent a bunch, it’s a sign that there’s a mismatch between the questions and the game, or that the game shouldn’t have a “list of questions” move. I once played in a game of Pasion de las Pasiones that was hacked to play a children’s adventure story (it worked great, which is a real testament to the flexibility of what seems like a very specific game), but the questions were just so ill suited to what we were doing that our GM rewrote them. Usually you can see what questions the players actually want to ask, and just make them the questions for the move. But I think this also leads into your other point…

The other question here is what is your trigger for this move, and how are you, as the GM, interpreting it? Players can ask any questions they want, whenever they want, and they shouldn’t always need to roll to get the answers. At what point are the answers hard enough to spot (or the PC’s actions sufficiently involved) to warrant a roll? What circumstances require a roll? Apocalypse World needs to you be in a “charged situation” before you roll. In Crossroads Carnival you’re “keeping an eye out for trouble.” Even Monster of the Week (which has what I think is a weak version of this move) requires you specifically to “Investigate a mystery.” None of these are just about checking out a space, they aren’t a D&D investigation check, there’s always a specific goal or context which informs the set of questions, and that context will usually push players to a specific question or set of questions. Coupled with on point GM moves to answer the questions that drive play forward, you’ve got a machine for keeping things interesting. (There is a GM skill element in running the machine, but the game gives you the tools and practice makes perfect). If the “question asking” move is only triggered when one or more of the questions are relevant, and you’re putting players in positions where the trigger is hit, players won’t usually go into “puzzle solving mode” because they’ll have obvious, immediate need for the answer to one or more of the questions.

Edited to add:. All of this assumes you want a game with the kind of snowballing moves and forward momentum of classic PbtA. For a game that is actually about player skill in investigation and puzzle solving, for example, these are very bad tools. So far as I know, there is not yet a good set of tools for playing that kind of game in a classic PbtA chassis.


I don’t really have much to add to Jesseabe’s excellent answer, except to note that I don’t “bend the questions” in my games either – in general, if the game is well made, the questions will be spot on as-is, though the 10+ problem can be real. (This makes me think of maybe writing a move where investigating lets someone hold 3 and spend that hold to EITHER ask a question OR do… something else that I’ll need to think of.)

I do think that a lot of people have experience with Dungeon World though, and that its “Discern Realities” move is… not great, though better minds than mine have gone over Dungeon World and not changed it, so maybe it’s okay and just needs a better trigger? I dunno.


ask a question AND take +1 forward on following an hypothesis / the next interaction with the same prey / … ?

1 Like

I was thinking more like “Ask a question, take +1 forward OR Take +1 Forward again on the answer to a question you’ve already asked” or something.

1 Like

For what it’s worth, I’ve been playing for the past few months in a very GM-directed game of Homebrew World, and the revised Discern Realities and Parley moves have been suiting us quite well–even with very investigation-heavy scenarios.

It’s a pirmarily city-based based game (far-flung cosmopolitan city on the edge of a Roman-esque empire, lots of different factions and politics and bureaucracies, ships and jungles and ancient ruins, all that). Two of the bigger “adventures” we’ve done were straight-up mysteries.

In both, the GM figured out in advanced what the story was behind the mystery: what was actually going on, what had happened and when, who was involved, who was affected, who knew what, etc. He dropped some hooks on us, relevant to our characters, and then… off we went.

And frankly, it’s been… great. Discern Realities has regularly opened up new leads for us to investigate, and just as often it’s confirmed that, no, that’s a dead end. The answers we got never solved the mystery itself, but rather gave us a better sense of what was going on. Likewise, the revised version of Parley really helps move things along when interviewing a witness or a suspect.

But we’ve still had to be the ones who put the big picture together. And we had to be the one’s to take the initiative to talk to X, Y, or Z. The GM has said multiple times “I always make sure that I’ve got at least one way you could learn this info, but pretty much every time, you end up learning it a different way than I expected.” So it’s been very emergent and fun, and the “null response” answers on Discern Realities are almost always useful even if they are exciting.

Anyhow, I for one am a big fan of a good “list of questions” move, because they ensure that the questions being asked by the move are actually generating insight rather than just data. But if you prefer to make it open ended, maybe check out this move from Freebooters on the Frontier 2e:


Greetings, my first post here. I am one of the playtesters in Andante’s game, and I signed up to give my input on the subject.

First, I would like to say Andante is a great GM and I have enjoyed his games.

I think the misunderstanding may be due to the term ‘prep-heavy’. You can prep heavily and still improvise by changing the environment, characters, and obstacles as required by the story, prep doesn’t stand in the way of that. What to prep is another topic which I won’t touch on here.

And of course, GM-led is not the issue here either. No player was asking for authorial/directorial control over the story.

I interpret most PC Moves in PbtA games as let’s find out what happens next, and the answer should be interesting and/or add to the plot. Otherwise, why roll?

To give a specific example, we were in a space habitat and there was a module at the end that contained the emergency escape vehicle. My character decided to explore that, going there, climbing up the vehicle then exploring the interior, which yielded nothing pertinent to the story (there was a PC Move where I rolled). In the end, this was 10 mins of wasted screen time as far as the story is concerned.

My philosophy of play says if you are going to allow the players to do actions/Moves it should count for something/ It should be an interesting encounter or reveals some significant detail. it should advance the story in some way.

With regards to asking questions, the same philosophy applies, are you testing the player’s ability to choose the right question to ask? Or do none of the questions yield new info anyway? Or are the questions there for you to reveal something interesting that the players can work on? “Yup, sorry no new info here.” on 10+ success made it feel like the Move was a waste of time.

I think Andante’s style is he prefers a relatively more simulationist, exploratory environment, and he subconsciously prefers a more tactical game that tests the player’s skill (rather than character). To me, a Move in PbtA shouldn’t be analogous to a skill check like in more conventional RPGs. You are not rolling because the skill in question comes up, you roll because you want to find out what happens based on the result of your roll.

Of course, all of these are just preferences, rather than a correct way to play.


I wonder then if the most helpful thing here is to think about the role random encounters and wandering monster tables play in classic dungeon crawls as a way to square this circle. They are meant to represent the ecology of the space being explored, and have a chance of generating an encounter even if there isn’t anything explicitly planned in a given location on a dungeon map.

When I’ve run classic modules in Dungeon World I’ve often used the wandering monster table from the module to inform the answers to a relevant Discern Realities question even if the map has nothing keyed in that location. The idea is that there’s always a chance these things might be around, so indicating a sign of their passage makes sense even from a simulationist perspective. You can create these lists as part of prep to model the location ecology, and they help the space feel alive and dynamic, rather than static and waiting for the players to arrive. Then, instead of rolling for random encounters at certain intervals, you use GM moves to surface them at fictionally appropriate times, sometimes in answer to questions.


Jesseabe, I think games like Ironsworn embrace this concept and it works well. With this method you could automate many of the mechanical and adjudicative roles of the GM, leaving only the actual narration.

There is a space where OSR type and pbta types games intersect and IMO it is worth exploring.

1 Like

Hello, I don’t know much about the topic, but this looks that it could be of use: .
It’s a methodology for determining what the GM can or can’t say after preparation. It’s for D&D, but the guidelines are system-free.