Then it would have happened on success or failure. The fact that it only happens on a failure, that the room is empty on a success is the issue.
Why would it happen on success or failure?
I talked about a "diegetic reason" - that is, a reason that comes from the shared fiction. As I said, it's a kitchen, there is a burglar, the burglar has bungled their burgling attempt, and so they have startled a cook in the kitchen. There is nothing there that departs from in-fiction, "diegetic" reasoning.
If you insist that
the shared fiction concerning the kitchen, and who is in it must be the same
whether or not the player's roll succeeds or fails, you are not talking about things being "diegetic". You are asking for a far more particular sort of process of determining the contents of rooms - at least roughly, map-and-key resolution, or some approximation to it (eg upthread you have said that you as GM would work out, in advance of the burglary attempt, who is where in the house: that is a version of map-and-key).
There's nothing wrong with playing map-and-key style. But it is not the only way to pay RPGs, and it is not the only way to generate "diegetic" consequences.
At 2am having a cook in the kitchen wouldn’t be diegetic.
That's a matter of opinion. I think I know as much about pre-modern households as anyone else posting in this thread, and I don't find the idea that a cook (i) sleeps in the kitchen and/or (ii) might be working in the kitchen in the middle of the night, problematic at all. I also note that "2 am" is one the random accretions to the situation not found in the blog.
I am right that cook in the kitchen at 2am is viewed as a flawed example by you because it’s not a complication that would typically happen in narrativist games implementing fail forward or success with complication?
I don't regard the cook in the kitchen at 2 am as a flawed example. Nor as a strong example. What is the context? What threats had been telegraphed?
Here's an actual play example that is in the same general neighbourhood:
Aedhros re-entered the scene: with a successful Stealthy check I entered the kitchen unnoticed, and found Alicia. I proposed that we relieve the innkeeper of his cash-box (repay hurt for hurt) and Alicia agreed. Then we would take on the master of the ship. Alicia used her Weathersense to determine if a mist would be rolling in; her check succeeded, and so her prediction of mist was correct! (We'd agreed that a failed check mean clear skies and a bright moon.) She also rested (for about 6 hours) to regain one point of Tax, taking her Forte up to 2.
With the morning mist rolling in, it was time to clean out the innkeeper's cash box. We agreed that the day's takings would be 2D of cash. With successful checks, Alicia cast Cat's Eye so she could see in the dark; I succeeded at a straightforward Scavenging check so that Aedhros could find a burning brand (he can see in dim light or by starlight, but not in dark when the starts are obscured by mist). Alicia went first, in the dark but able to see, but failed an untrained Stealthy check despite a penalty to the innkeeper's Perception check for being asleep. So as she opened the door to the room where was sleeping on his feather-and-wool-stuffed mattress, he woke and stood up, moving his strongbox behind him.
Suppose that the Stealthy check for Aedhros to enter the kitchen unnoticed had failed; or the Scavenging check to grab a burning brand for light. What sort of consequence might have followed? I have no memory now, some years later, of what we at the table were thinking at the time. But one obvious possibility, implicit in the situation, is that the innkeeper comes downstairs to his kitchen, for some-or-other reason (he is hungry, he can't sleep, he has a big feast to prepare for, he want to check that his new hire Alicia is not doing anything untoward, etc, etc).
The fact that the cook example, in the blog, is being presented through the lens of D&D play rather than Burning Wheel play doesn't change any of these considerations, as far as I can see. If a D&D GM wants to use "fail forward" resolution, then they are going to need to adopt some of the other practices that support it, and help make it work: situations with an implicit trajectory of threat and promise; action declarations with express or implicit intents, so that there is some
desired outcome of the action which can then be used as a touchstone or measure to aid in determining what will count as a
failure; etc.
I think that depends a lot more on the amount of detail given for each scenario. There’s some variations of dragon in this hex because of a low roll that likely corresponds to the cook in the kitchen at 2am. I would expect though haven’t confirmed that such scenarios would cause similar dislike.
And theres also the additional difference around what the roll is described as representing, which I think is the bigger concern, or at least it is for me.
Well, in "fail forward" resolution the
roll of the dice doesn't represent anything at all. The character attributes represent things; the obstacle/difficulty rating might represent something (depending on the details of the RPG in question; in D&D 5e I think the DC is generally understood to represent something). But the roll doesn't. It's a decision-making device: everyone at the table has agreed to abide by the outcome that the roll determines.
I don't see a random encounter roll representing anything either. The
chance of a random encounter often represents something (like how dangerous, or heavily populated, etc) an area is. The distribution of the entries on the table often represents something (eg frequency/rarity of particular sorts of creatures). But the roll itself is just a decision-making device, I think.
And if you don't like the dragon example, because dragons are rare and special, make it an owlbear. Or a nest of giant ants. Or an angry bear. Whatever it is, the point remains that
the reason everyone agrees that the shared fiction contains that thing in it here-and-now is because of some dice rolls.
There is in fact a difference. I came up with that town waaaaaaay before that session, so that's when the farrier was there, not during the session the player reminded me that I had forgotten to write it down.
The difference is timing. Making it up when the player asks is creating it on the spot. Confirming the existence of something that was in the town prior is not making it up on the spot.
Huh? When you invented the town, you did not expressly invent a farrier. Perhaps a farrier was implicit in its invention; but as I said, you could just as easily have stuck with the absence of a farrier and turned that into a plot point.
I mean, suppose that you decided that the village had 9 buildings in it. Then it is true that, at that same time, you decided that the building had 3^2 building in it, even if no one has yet done the maths. Mathematics works by entailment.
But the farrier, even if perhaps implicit, is not entailed. It wouldn't contradict anything to have no farrier. "Confirming" the farrier is a creative, authorial decision; just as it would have been an authorial decision to affirm the absence of a farrier.
It has no real world existence, but it does objectively exist within the world that I use. I put it there. It was written down.
The farrier wasn't written down, until you wrote it down. Like the cook. Like the randomly encountered owlbear.
@pemerton just for clarification. Are you arguing that even those that eschew formally making things up on the spot are doing so informally anyway? I'm trying to understand what you are saying. It seems a bit inflammatory if you are saying what I described.
@Maxperson posted an example of play, way upthread, where:
- He as GM narrated the PCs arriving at a village;
- A player, as their PC, asked about the presence of a farrier in the village;'
- Maxperson had not made a note about a farrier in the village, but regarded this as an oversight, because logically there would be a farrier in the village;
- Maxperson therefore told the player that yes, their PC can find a farrier in the village.
My point is that Maxperson made up, at that moment of play, the existence of that farrier in that village.
Likewise, when a GM rolls a random encounter die, and it comes up 6 (or whatever, that triggers an encounter), and then rolls on the wandering monster tables, and thus decides that there is (say) an angry owlbear whom the PCs here-and-now encounter, the GM is making up, in those moments of play, the existence of that owlbear.
As a random encounter? No problem. Added as a result of a failed attempt to open a lock when it takes the same amount of time, makes the same amount of noise? That's why I wouldn't do it.
@Emerikol: this is the topic of discussion.
Some posters assert that when the cook, or owlbear, or whatever is narrated as being there because
the GM extrapolated it from their notes (as Maxperson did with the farrier) or because a random encounter is rolled, that is not "quantum"; but when the cook is narrated as being their
as a consequence of a player's failed check that is "quantum". And I am saying that I do not see how one is more or less "quantum" than the other: they are all examples of something being written into the fiction here-and-now, by the GM, in response to some real-world prompt: be that a question from a player, a roll of the wandering monster dice, or a player's failed skill roll.