Or another example that came up on the boards some time ago when discussing exploration. The party rescues a child in the wilderness and sets up camp with Leomund's Hut. The DM decides that the child wanders out of the hut during the night and attracts wandering monsters. Pure MMI. The party could not do anything to prevent this, short of tying the child up. All to create "challenges" for the party. And the response was, "Well, it's the DM's job to create challenges" with wide eyed innocence.
I haven't heard of this one before. Just bizarre!
I’m often amazed at how many GMs… myself included… will not want to allow what might be considered an “easy win”.
As if any given challenge is going to be the final challenge.
I think we need to decide as GMs whether we are being neutral adjudicators, curious explorers of the fiction, and playing the world with integrity or we are actively seeking to frame scenes to challenge the player characters. We cannot have it both ways. If we are doing the latter than I think we need a way to maintain competitive integrity because we are no longer actively maintaining it ourselves. We cannot both be referees and not referees at the same time.
@Campbell, I don't want to glom together different techniques and approaches without recognising their differences (and I think my posting history backs this up). But I think that there are some principles that can apply across quite a variety of techniques/approaches, that are relevant in the context of this thread and some of the examples being discussed (Rustic Hospitality and hiding from the Duke's soldiers in the barn; the druid and spirit magic in places of power; the child in the Tiny Hut).
The key principle I have in mind is along these lines:
honour the players' contributions to the fiction.
In classic dungeoncrawling, this means - for instance - respecting the players' decisions about which doors they open and shut, where they do and don't use their light, etc. If the GM makes decisions about a "living, breathing world" or "for fun" that make those player decisions irrelevant, then the integrity of play is lost. An example of what I have in mind here would be the GM deciding - not via some game process/check but just via their imagination - that some wandering monster has removed the iron spike the PCs used to wedge open the escape route behind them.
In scene-framing play, the need to honour the fiction is less about that sort of attention to operational decision-making, and more likely to be about theme and character trajectories. But respecting previous consequences also matters. If the players have won a victory in scene A, it's just bad GMing to frame scene B in a way that puts that win in jeopardy again - unless the players have themselves done something to put it at stake. If the GM is not sure about whether or not the players have intended to do this, they can ask! This covers the child example, for instance - why are the players supposed to accept that, simply by declaring their rest in their Tiny Hut, they're staking both their own safety and that of the child.
Relating this to
@hawkeyefan's point: if the GM doesn't have enough material to frame challenges that honour what the players have done - if the only adversity the GM can think of is
negating what the PCs have just won - then to be perfectly frank, they need to step up their game. Obviously there are many RPGs (including 4e D&D) that have excellent advice on this, that is followed through with related elements of PC build, situation creation, etc; but even vanilla D&D shouldn't make this too hard to do.
And connected to this: the system, even 5e D&D with its rather squishy, "rulings not rules" structure, will provide support here. The basic motto could be
Forward on a success, backward on a failure. Use the checks.
In the child case, the child running out makes sense as
failure narration for an attempt to coax information out of the child (failed CHA (Persuasion)).
In
@hawkeyefan's case, the Duke's soldier turning up makes sense as
failure narration for an attempt to shore up the resolve of the villagers (failed CHA (Persuasion)) or to lead them away (failed CHA (Deception) or WIS (Survival) or similar).
In the druid-spirit-place-of-power case, PC ignorance makes sense as
failure narration for an attempt to recall ancient druidic mysteries (failed INT (Religion) or INT (Nature) or whatever applies at that table in that situation).
Again, 5e D&D isn't perfect for this because it has a large number of fiat abilities that don't, in themselves, trigger checks. But advice around how to use soft moves in response to those, for instance, will help here. Or in the case of a dungeon, it should be designed so that if the players go from room 1 to room 2 via fiat, there is something interesting in room 2 to engage them.
A final problem is the tendency to treat the "uncertainty" test for calling for a check in 5e as turning on
whether or not the GM thinks it's uncertain relative to their private imagination about the fiction rather than
whether or not it's uncertain relative to the shared fiction (which in some approaches to play includes a certain sort of prepared notes put together in accordance with appropriate dungeon-building principles). But that's easily changed!