D&D General [rant]The conservatism of D&D fans is exhausting.

[Enrahim]

That would rather blow up the whole "what's around the next corner" exploration piece, which is IMO and IME a key element of dungeon crawling.

You already have given an exelent example of the point I am about to make in the next part of your post, but I wanted to amplify that. I am a map nerd like few and absolutely love mapping myself as a player. Still I have found myself a but tired of listening to yet another "20ft hallway with a right turn at the end" not to mention clumsy attempts at describing an irregular room. So while I agree giving away the full detailed map is completely removing the joy of mapping I think it might point in the right direction, just taking it a bit far for my (and probably your) tastes.

I have been playing with the idea of using obviously inaccurate or simplified maps to try to bridge this gap. Like reconstruction of tricky well known rooms prepared by a professional chartographer based on historical notes. Or hastingly scribbled treasure maps. Or the map found on a group of dead adventurers perishing before having mapped out more than the basic structure of the dungeon. Or the map scribled down by the goblin captive quite possibly seeking to deceive them.

The idea is to have always a sense of uncertanty what might be around the next corner, but having it mostly be "as expected" making those moments when something really interesting and/or unexpected is found much more impactfull.

 

[SableWyvern]

The nearest thing to a megadungeon I've run in the last very long time is Dark Tower, which while fairly big still doesn't meet my "mega" threshold. It ain't ToEE or Castle Zagyg or Ruins of Undermountain, that's for sure.

My average is usually 40-50 rooms or areas plus surrounding wilderness, much the same as a typical TSR-era module.

40 or 50 is still pretty big. Do what interests you, of course, but if you can fit it in somewhere, I do recommend trying a dungeon of that size, where the PCs are able to acquire a map in advance. Have some spots on it that look like maybe they have secrets. Leave a couple things out. Potentially have the most obvious paths the most dangerous.

In any case, in my experience, players with a map (but who have learned the skills of no-map exploration in complex, dangerous, dungeon environments) are still cautious and still spend similar amounts of time plotting and planning where they want to check out next, it's just that the focus moves from more abstract plans where they have limited information and filling the map is one key objective, to ones where they have more information, but it's still hard to judge the value of the information. "Let's see if we can work our way to this part of the complex; that series of rooms looks interesting and that passage looks a bit weird, like it might lead somewhere."

I have run a few megadungeons, and exploring areas that are pre-mapped has made for an interesting change of pace at times.

Or, if you are completely uninterested, don't, of course. My sales pitch ends here.

 

[clearstream]

In the discussion about "bypassed encounters", I've seen several mentions of GMs planning out encounters in advance, only for the PCs to "bypass" them.

I don't really understand how that squares with the statements folks have made regarding non-planning from the GM, and I think this is at least adjacent to pemerton's questions. Specifically, this implies that there is, in fact, a planned sequence of events that will happen, and the PCs have found a shortcut which skips some of that planned sequence.

It's probably right to label Griffin Mountain (Kraft, Jaquays and Stafford, 1981) a classic sandbox (or at least, highly suitable for running one.) Chapter VII - Encounters - contains prepared encounters tied to regions. Additionally, there are encounters tied to Citadels and Points of Interest (location rather than chance-based.) I wouldn't use the language "bypassed" to describe players not encountering any particular one of these, even though they are prepared.

One practice I associate with sandbox play is that the camera follows the players so that we narrate whatever they meet. Player characters could conceivably become aware of an "encounter" (in the Griffin Mountain sense of prepared NPCs and potentially locations for interaction) just before reaching it, and choose to draw the camera away. That would be in a sense "bypassing" that encounter.

Another practice I associate with sandbox are NPC motives which are to be given effect. A motivated NPC could well send an encounter the players way (even in cases where that NPC is oblivious to them), and the player characters might well "bypass" that encounter. Such "bypassing" could become a lively and repeated element of play if player characters and NPC are actively engaged.

In short, the "planned sequence of events" you envision might not describe sandbox play very well. An NPC could "plan a sequence of events" (meaning of course, that the GM plans a sequence of events) which is that their guards routinely patrol their castle walls. The NPC (of course, the GM) has in mind that anyone seeking to come over those walls will encounter their patrols. And then of course the player characters could "bypass" those patrol encounters (find a way over the wall that avoids them.)

Still, it does not seem right to say that the GM planned a "sequence of events" where the players encounter those guards: the players dictate the sequence of events. Or perhaps one says "GM plans" but "players decide".

I think this misses the context? You are for each point explaining how each of these options can be desireable to a player and a "good thing". However the trilemma in question was not how to maximise player satisfaction, but how to maximise player agency spesifically. I feel what you are doing here is more in line with what you did in your post above arguing how restricting player agency can be desireable and might even be necessary. But that is not the point we are talking about here. The premise of the conversation the post you replied to was claims prep do not (meaningfully) reduce player agency and this was the claim under scrutiny. In that context a wider analysis of deireability seem irrelevant?

The dilemma then seems one of -- how can mode of play A work for persons seeking mode of play B? The problem is one of mismatching criteria: critiquing apples on failures to be pearish.

What exact kinds of player agencies are at risk (not just "player agency is at risk" generically, but player agencies to do what in particular?) and why are those good criteria to judge sandbox by?

 

[clearstream]

In this case the PCs didn't end up having a combat on the battlemap as I'd prepared - they instead made use of stealth and illusions to reach the target lair without having a fight.

This is kind of why I raised this example - the resulting gameplay to me felt like it relied a lot more on my GM fiat than it would if they'd had the fight (or if I'd sufficiently detailed a large enough area to allow for the approach they chose without continual negotiation)

As a side note, I think this is also an example of the current discussion about Encounters and bypassing them. I'd certainly be comfortable describing this whole event as the PCs "bypassing" this encounter - the map was not engaged with, the combat game was avoided and so on - but on reflection it could equally as reasonably be stated that they fully engaged with it. In some games there would be no mechanical difference between the two, it's more a quirk of D&D and similar games that there is a difference.

(Emphasis mine.) How was GM fiat involved in players making use of stealth and illusions? I'm assuming not in the sense that GM decided players would make use of stealth and illusions. Was it connected with how the outcome of using stealth and illusions was decided?

 

[Gimby]

(Emphasis mine.) How was GM fiat involved in players making use of stealth and illusions? I'm assuming not in the sense that GM decided players would make use of stealth and illusions. Was it connected with how the outcome of using stealth and illusions was decided?

Yes - essentially within the described zone I had plotted out cover and so on, outside of it I was improvising. It felt to me as the GM that the primary determinator of success wasn't the characters or their capabilities, it was how willing I was to let them succeed and let them "bypass" the prep work that I had done.

 

[clearstream]

Okay, but that...doesn't seem to have taken this horn of the trilemma at all? Like this is just saying "Exposition, and player participation in its development, is important and one of the critical parts of sandbox play". I completely agree with that. That's not controversial in the slightest. But does it address the risk (again, not the guarantee, just the risk) of a GM placing excessively high expectations on the players to provide the needed prompts/questions/etc. to trigger the reveal of critical information. I gave an example above of that problem, where I (in the absence of people explicitly telling me otherwise) would 100% guaranteed never have considered to ask the one-eyed man at the tavern about stuff, especially if the GM never says anything about the one-eyed man unless I specifically ask a question about the patrons of the tavern and what they look like. Which was the point of that horn of the trilemma: if you say too little and/or demand too-specific a question/prompt in order to divulge information, the game descends into what is derisively called "pixelb!+©#ing" and what might more charitably be called "inadequate prompting".

To preface, I was thinking more about why each horn was worth entertaining, than trying to resolve them. Generally speaking -- per my general conclusion -- it's true that the features of each mode of TTRPG "risk" degenerate outcomes.

Grasping the first horn then, inadequate prompting is a problem found in some modes of play. I haven't observed it in sandbox, but I have observed it in Deathmatch Island. The game expects players to move node-to-node and engage the challenges they reach. In Phase Two they must engage with the Scout, Scramble, Battle Royale structures which have specific rules and sub-structures. If Production (GM) inadequately prompts the play can wind up eddying.

In sandbox, inadequate prompting is possibly masked by the result that players simply go in another direction. Seeing as GM had no particular direction in mind, it's never noticed that they inadequately prompted.

Sure, but folks have been highly insistent that they don't do that thing. That their job is never to "elevate the experience", but only to furnish a coherent world in which the players may participate. Putting too much emphasis on specific things A and B and not enough on the seven million other things that theoretically could permit interaction very easily becomes "the GM put her thumb on the scale" rather than "the GM merely provided inclusions and exclusions to improve the experience". Where the previous situation is a (potential) faltering by way of making it too difficult to know what things are of great importance and what things are not, this is a (potential) faltering by way of making it seem like only the "inclusions" matter, and everything else is simply set-dressing.

I don't follow.

Really? I find that...a bit hard to swallow. Going into excessive detail on too many things (especially if one is attempting to avoid, as I would term it, "privileging" some answers over others and thus inducing player choice) seems like quite an easy pitfall to fall into.

That surprises me.

It is not a philosophically skeptical view of GMing in general. It is an unconvinced view that this very specific type of GMing, where the GM is committed to populating and iterating on a world, but simultaneously consistently avoiding anything that (for lack of a better term) "incorrectly" induces player action, is workable in practice without any process or procedure beyond intuition--which is what I have been repeatedly told is the case for many, if not most, GMs of this style. That there simply, flatly ISN'T any kind of process or procedure, and thus it is flatly impossible for someone to explain any part of their process or procedure because such a thing just doesn't exist.

What do you mean by "incorrectly" induces player action? What would be an incorrect player action in a sandbox?

 

[AnotherGuy]

If the players avoid the defence of a castle, then the GM's planned disposition of forces won't be triggered. But that does not seem to me to be bypassing or avoiding an encounter. It's choosing to sneak in rather than to assault head on.

Okay by this logic.

1. Engaging with the encounters is choosing to assault head on.
2. Not engaging with the encounter is NOT choosing to sneak.
3. Avoiding or bypassing the encounters is NOT choosing to sneak.
4. Choosing to sneak is choosing to sneak.

Are you standing by this logic.

But I'm not familiar with reifying some of these possible events at the table, calling them "encounters", and then saying that the players, via their play of their PCs, bypassed them. To me the language of bypassing seems to connect to a concept of prep and planning and GM expectation that I'm not used to.

You played 1e and RM. Presumably you used to prep back then - in the Trad sense or were you using PbtA principles in the 80's before they became known?
I know you were familiar with 4e's Cairn of the Winter King and likely other modules. So, in your mind what happens (what do you call it) when a trad GM uses the module as is but the party avoids/bypasses some of the encounter due to smart play?

 

[Enrahim]

The dilemma then seems one of -- how can mode of play A work for persons seeking mode of play B? The problem is one of mismatching criteria: critiquing apples on failures to be pearish.

What exact kinds of player agencies are at risk (not just "player agency is at risk" generically, but player agencies to do what in particular?) and why are those good criteria to judge sandbox by?

No, this is not the conversation we are having here. It might be a less interesting or valuable topic. But the situation appear to be that we in the depths of a soon 10k post long tread have managed to dig us to a point were we actually are having a serious and interesting conversation about exactly the level of pearishness of certain apples. And while a polite reminder that there are more to apples than pearishness might be aproperiate, insisting that the pearishness of the apple is not a worthy conversation to be had at all is maybe a bit less welcome

That being said you are bringing in now that there might be different kind of pearishness that might be interesting to look at for the apples, and that is a path I would be happy to hear more of your thoughts on!

 

[EzekielRaiden]

To preface, I was thinking more about why each horn was worth entertaining, than trying to resolve them. Generally speaking -- per my general conclusion -- it's true that the features of each mode of TTRPG "risk" degenerate outcomes.

Grasping the first horn then, inadequate prompting is a problem found in some modes of play. I haven't observed it in sandbox, but I have observed it in Deathmatch Island. The game expects players to move node-to-node and engage the challenges they reach. In Phase Two they must engage with the Scout, Scramble, Battle Royale structures which have specific rules and sub-structures. If Production (GM) inadequately prompts the play can wind up eddying.

In sandbox, inadequate prompting is possibly masked by the result that players simply go in another direction. Seeing as GM had no particular direction in mind, it's never noticed that they inadequately prompted.

Again, I refer you to the example that I was explicitly given by others who favored an old-school sandbox experience: the "you didn't talk to the one-eyed [or one-armed, or various other maimings] man, so you never heard that the slimes in the mines are weak to lightning but divide when struck by regular weapons, which means your death at their hands/pseudopods is entirely on your head." That's not an exact quote, of course, but it covers the core points: (1) the players were just supposed to know that some NPCs in the tavern were necessary sources of information; (2) failure to interact with the one and only source of that information is construed as the players' mistake; and (3) any deaths/losses that result from failure to interact with that source are thus earned by having made that mistake.

This isn't just a hypothetical. I genuinely would have been caught by surprise by such a thing, in a game self-professed to be a sandbox, which would have gone completely beneath my notice if it weren't for these GMs (more than one!) explicitly saying that this is supposed to be an obvious prompt. Inadequate prompting, and resulting problems, were one of the earliest things I ever encountered when discussing this style with its own proponents, on this forum specifically.

I don't follow.

I have been told, over and over, that the sandbox GM must not manipulate the players' choices. But in the very act of giving detail to thing X and not giving much detail to thing Y, that can manipulate player choices. "Oh, this is the thing the GM wrote a lot about, it must be Important" is a perfectly natural thought for many players, of any style. Likewise, things that don't get any description at all are at risk of being written off as unimportant or non-interactable.

This is...kind of essential to the trilemma. The first path, giving no bias by (almost) never prompting, only waiting for players to act--but then the (from my perspective) very high risk of players never even considering something due to lack of prompting. The second, accidentally manipulating player action toward the things you prepared and away from their own choices/creativity. And then the third, overwhelming (or distracting) them with too much information by preparing lots about TONS of things so they won't feel pressured toward any given thing. Both extremes have (what seems to me) a high risk of not just undesirable but actively harmful outcomes if the goal is heightened player agency; yet the seeming middle-ground risks a different threat to player agency.

That surprises me.

Really? I can't imagine why. If it's easy to see why a dearth of information would be a possible problem, why would overcorrection, flooding the player with so much info they can't keep track of it all, or so many options they break down into analysis paralysis, not be also reasonably easy to grok?

What do you mean by "incorrectly" induces player action? What would be an incorrect player action in a sandbox?

I already clarified that in the post itself, but you have misread what I said. I wasn't saying the action was incorrect. I was saying the inducement is incorrect. The idea is that there are some ways to induce which are good, and other ways to induce which are not just bad, but extremely bad, as in, outright contradicting the purpose of sandbox play.

The player actions themselves were never even under consideration.

 

[pemerton]

That seems like an enormous thing to be decided by running out of social hit points to me

I thought the consequence of running out of hit points was supposed to be enormous - eg PC death.