I like the way you've laid this out. In some PbtA game systems (DW is an example), I see that the complications or trouble of some moves is up to DM (e.g. Volley, in which "you have to move to get the shot placing you
in danger as described by the GM"). I agree that this is not intended to be unconstrained, yet within that I think there are still essentially infinite possibilities.
When I run the thought experiment of collecting ten DMs and giving them the same situation and result, I feel justified in predicting that what they describe will
not all be the same. Perhaps it is that leeway, that speaks to the point
@Pedantic is making?
This doesn't seem to me like a very meaningful or insightful conclusion, TBH. That is, sure there are an 'infinite' number of incredibly minute variations on the following fiction. ALL of them must honor the agenda and, for the most part, live by the principles of play which the game (DW in this case I presume) espouses. Not only that, but the fiction has to follow from whatever came before, and that is a VERY strict constraint!
So, basically, yes, the player rolls an 8 while conducting the Volley move. Lets assume they choose "move into danger" as the complication. There is a tactical situation here. That situation is going to be fairly well-specified in most cases, so there's really a limited repertoire of threats. Most likely the GM is going to have another enemy deal damage to your character (fiction will be something on the order of a bad guy moving in on you, or the PC getting too close and getting hit). Conceivably there could also be environmental hazards that could come into play, but those are again going to be dictated by the FICTION, not so much the GM! Finally, maybe the situation might call for an entirely new and thus-far unanticipated danger (IE because nothing much else is available). AGAIN this is a constraint of FICTION, not a GM whim. Here we all agree, there are probably a lot of options, but consider there are hinted-at dangers, fronts, factions, and dungeon moves. Most of the time the GM should be drawing from these, as opposed to spinning off into lala land to invent something completely new.
In other words, I can tell you with some fairly high reliability things like "what is
@Manbearcat going to say happens here?" I won't likely be exactly right, but I know it will be one of a small menu of possibilities that meet certain well-established criteria. This is really no different from the sort of choices that GM's make in, say, classic D&D, when a player does something aside from 'make an attack' or some other established type of 'move'. The GM, constrained by various factors picks a response. In a few cases that might be directly based on pre-existing notes, but 99% of the time it won't be, certainly not entirely.
The real question though is, what ultimately is the worth of this kind of statement in the first place? Clearly "stuff happens which nobody anticipated" should be pretty routine in RPGs, right?