Perhaps an example is the one of high-ground. Vincent says it is F > S, but you here seem to say that you would analyse it as F < S if the player described a motivation that they sought high-ground to gain the mechanical outcome.
Motivation is not relevant to Baker's clouds and boxes analysis. The player has their PC take the higher ground in order to be better able to fight their foe. The player has their PC use a trait in a certain way to get some or other advantage. But one is clouds-to-boxes, the other is boxes-to-clouds.
One way to settle things can be to scrutinise where we land. That produces results consistent with Vincent's assessment in many cases, such as that for high ground. The end result in the case at hand is marking the box on the character sheet to store up a check for the next camp phase. Or maybe scrutiny doesn't belong on where we land... I'd be curious about the reasoning for that, if so?
The object of analysis is the process of play: how the shared fiction is established, and what role (if any) cues/mechanics play in establishing it.
In the high ground example, it is already established in the fiction that that there is high ground. Then the player declares
I stand on the high ground - and that fictional change yields a mechanical consequence. Hence it is clouds-to-boxes.
That mechanical consequence (+2 to hit) interacts with another mechanical state of affairs (the dice roll to hit) to produce an arrow from boxes back to clouds -
Your character hits mine.
When a TB trait is used, it's in the context of resolving a test, in order to settle the content of the fiction. It is established that the PC is doing such-and-such a thing: that is the action declaration. But it's not established that they are doing it hurriedly, or carefully, or whatever. And the process is that the player establishes a mechanical state of affairs - eg suffering a debuff and thereby earning a check; or spending a limited resource (uses per session) and getting an advantage die - and as part of the rules for doing that, also establishes some fiction (eg
I was quick-witted and so go the drop on them or
I jumped the gun and misjudged the situation). The fiction has no "life" to it other than as mere colour that is an accompaniment to doing the mechanical thing. It is boxes-to-clouds.
Coming back to the case at hand, the written description feels a little unnatural to me. If as I think the razor is how it is played at the table, then it could be
Player - "I work quickly, not worrying about being careful."
Assuming it's a context where working quickly isn't needed and being careful would be beneficial, then...
GM - "Sounds like Quick Witted is working against you, lose one die and mark a check."
I feel like that is fiction-first. Player said what they were doing without invoking mechanics. GM translated that into system. F > S.
I don't think this is an accurate account of how Torchbearer plays. Torchbearer is
not "if you do it, you do it"; and as a special case of that general feature, the GM does not impose trait-based mechanical consequences that follow from how players declare their actions. Players establish trait-based mechanical consequences when they want them, and as part of the rules for doing that must also narrate some appropriate fiction.
For me, this all suggests a very great divide between PbtA and TB.
As best I can judge, PbtA and Torchbearer do not have a great deal in common as far as the process of action declaration and action resolution is concerned, until we get to
consequence narration where a Torchbearer GM who is familiar with the PbtA "soft move, hard move" approach to narrating consequences will benefit from that, I think, in narrating twists.
I don't recall anything in the TB2e text urging a fiction-first approach.
I think it is labelled
Describe to live. The GM describes the situation or obstacle - which is fiction - and the players describe what their PCs do to overcome it - which is fiction. But the process of then determining the full scope of the action declaration - including who is helping or aiding, what gear is being used, etc - and the process of resolving that - what skill is being tested, what fate or persona is being spent, what traits activated, etc - is not fiction first at all. AW and DW have no real analogue to this. And obviously rolling the dice in those RPGs is not much like building and then rolling and resolving a dice pool in Torchbearer.
I think, after starting another adventure in TB2 I definitely find that the way I approach it at least is from a mechanics standpoint. I find I'm deciding how to deploy my character's mechanical attributes in a way that produces effective results, and then working out how that can be extracted from the current fictional position and my conception of the character's personality (which is already rather heavily embodied in attributes like goal and instinct as well).
That makes it pretty distinct from, say, Dungeon World, where IME the use of mechanics was much more an outgrowth of the GM and players generating the story. Like, you would decide what to do in DW, and then someone (the GM in theory at least) "Oh, that's an XYZ move."
At least to me, you seem to be describing here exactly the difference between DW's "If you do it, you do it" approach, and the lack of that in TB. In TB there are
lots of ways to do it, in mechanical terms - skills, buffed in various ways, with or without help, gear, etc, all which is brought in by the player, or not, depending on available resources - and a big part of the player skill required is to decide how to do it on this occasion.
I do share
@AbdulAlhazred's sense that TB is separated from DW.
I don't think this is contentious at all. From the OP:
At a high level of description, Torchbearer can be compared Dungeon World: a modern system dedicated to capturing the feel of classic D&D. At a more detailed level I think there are significant differences; I'll get back to these below.
I see fiction-first and story-now as sympathetic but separate modes of play
Burning Wheel is "story now", but I don't think is "fiction first" if by that we mean DW-style "If you do it, you do it." Conversely, Classic Traveller is fiction first in that sense, but needs a bit of tweaking to play as "story now", and Moldvay Basic can be played fiction first in that sense and will need a
lot of tweaking to play as "story now".
In general: "fiction first" is a description of (some features of) the process of action declaration and resolution. Whereas "story now" is a description of the "creative agenda" - ie
what are we all hoping to get out of creating this shared fiction together.
I gather you have a very large amount of experience with Burning Wheel. Have you found a point where system gets out of your way, and you can uphold a fiction-first approach?
I'm not sure what you mean by fiction first. First in respect of what process, or what unit of analysis?
Like Torchbearer, in Burning Wheel descriptions of situations, and descriptions of what is being done to overcome them, begin with the fiction. But when a player is building their dice pool -
OK, I test Inconspicuous and I'm FoRKing in Acting and Cultists-wise - it's not wildly different from Torchbearer: it's all about building up a mechanically-framed conception of what the PC is doing.
In an extended resolution process like Fight! or Duel of Wits there is an action economy, and a suite of moves to choose from, much like a conflict in Torchbearer. In TB, the reason that PC X is acting this round, rather than PC Y, is not flowing from the fiction; it's driven by the rule that if there are at least two characters involved then no one can take consecutive actions. In Burning Wheel, the reason that a player blind declares 3 actions in a Duel of Wits is not because anyone thinks that's the fiction of argument: it's a mechanical device, adapted into social conflict resolution from the melee combat resolution framework.
Burning Wheel is not "if you do it, you do it" any more than Torchbearer is. In this respect both resemble 4e D&D, and differ from (say) Classic Traveller and (I think) some approaches to classic D&D.
Where I feel TB might depart pretty substantially from BW is that TB seems to lean into GM-prep, while I interpret BW as desiring more on-the-fly invention responsive to player goals.
BW has systems for prep - for "burning" monsters, magic items, NPCs, etc. But it is actively hostile to GM prep of situations, whereas Torchbearer is (as far as I can tell) reliant on GM prep of situations.
TB doesn't forestall dramatic character development. Rather it specifies the scope of that development (this will be a gritty story about doomed adventurers.)
Rather than Story Now, Torchbearer strikes me as more Right to Dream. In our sessions at least, the mechanics have really enforced the game’s themes. Being an adventurer is hard. The Grind will wear you down even if you play well, but if you do, you might get to return home and enjoy some of the treasure you found.
I don't think it is Right to Dream, because as a player you can't just turn up and play with no metagame agenda other than exploring your character and the situation. I think if you do that, you'll get hosed. You have to actively think about how you can "win" - collecting and spending your resources, optimising the distribution of tests across the party, etc.
But I agree with
@kenada that it isn't "story now" just because play might, over time, tend to produce a narrative character arc.