Another possible insight. I will here look at 3 semi-fictious extreme styles of play I am going to shorthand railroady (RRY), sandboxy(SBY) and nary(NY). In all of these playstyles GM has the ultimate responsibility for framing scenes.
In RRY play, the GM don't care about player input at all, but are free to frame scenes according to their own grand vision.
In SBY play the players are responsible to explicitely indicate what their characters are after here and now trough action and intent descriptions. The GM is obligated to take this into consideration when framing the scene.
In NY play the players are responsible to communicate clearly what is important for and about their characters trough character descriptions and other "flags". The GM is obligated to take this into account when framing the scene.
Let us look at a few of the concepts we have been looking at in this thread. Be a fan of the characters are an essential tenet of NY, as that neatly summarises the last GM obligation. However it serve no scene framing purpose in RRY play. For SBY it seem like innefficient/unneccessary advice, as the players are supposed to be explicit about their desires for the new scene trough their actions. The nature of the characters that the fandom is fixated on is not supposed to be taken into account.
For the players to make the character's life not boring also looks different in terms of scene framing. For NY, this involve making sure that the implicit signals about who the character is are inspiering the GM to frame not-boring scenes for that character. In RRY play this advice hardly make sense, as the players are in no position to affect the scene framing at all. In SBY play however, for players to make their life not boring in terms of scene framing would require them to indicate not boring actions and intent. This can make perfect sense, but will get expressed very differently in SBY play than in NY play.
Finally, all styles of play want to prevent a failed roll from stalling the game - that is after a failure there should (still) be an interesting scene. A technique to assure this is that the GM bakes a reframing of the scene into the failure narration, making sure that reframed scene is interesting. This technique works fine for NY and RRY play. However it do not work for SBY play in general, as the GM do not have enough information to frame the scene according to their obligation. The players need to be given a chance to explixitely state an action and intent as input to the new framing for the GM to follow their SBY obligation.
I hope this framing can give a new perspective into what I think have been a core difficulty in this conversation - seeing how various techniques interact wildely differently with different types of play. Even if the types of play described above is extremes, I think there are relevant things to learn also for more realistic play situations.
It's an interesting analysis, but it feels like there really isn't much difference between NY and SBY (except for one I'll address below). That is, the only difference I can see is that, in SBY, the players have a more
overt responsibility to be as proactive as possible, while NY they should still be proactive, but not being all-proactive all the time is fine. In pretty much every other sense, though, your analysis seems to put the two very close to one another. Basically, their only difference is the degree to which the players
need to be "taking the lead" in terms of what matters and where the action "goes".
Conversely, RRY seems to be distinctly different from both of the other two at almost every turn, except that very last one...which I'm not entirely sold on?
That is, there is no need to "reframe" anything at all in the RRY approach. The GM just declares what they wish to declare; the players are simply along for the ride regardless of what the rolls say. In NY, it's necessary, and in SBY, it's simply a greater challenge because, as before, the players are required to furnish more than a minimum amount of information. They need to be pretty proactive, as before; it is
possible for the GM to struggle if the players simply aren't giving them enough to work with, but far from guaranteed. Hence, it's not that the technique
doesn't work, it's that the technique doesn't work for a GM
acting autonomously.
And I think that's a pretty big kicker here. In both NY and SBY, the GM cannot act entirely autonomously. The only meaningful difference between NY and SBY is that
sometimes the GM can act autonomously in NY, but she is nearly incapable of acting autonomously in the most extreme versions of SBY. In RRY, the GM not only
can act entirely autonomously, they
always act autonomously.
That's a pretty big difference, and shows how NY is much closer to SBY than to RRY. In NY, it
might be the case that the GM can act autonomously, but it's not guaranteed to be possible--some of the time, they won't be able to. In SBY, it's rare to nonexistent that the GM can act autonomously--so both have a presumption, unless proven otherwise, that autonomous GM action is not guaranteed. RRY has autonomous GM action not only guaranteed, but outright enforced, all the time.
Can we come up with two additional methods that would fit into the missing spots? That is, it seems to me this reveals a pretty major gap: there should be a #Y where autonomous GM action is always possible but never required (the exact midpoint between SBY and RRY), and a separate $Y where the natural state is the GM acting autonomously but it isn't
guaranteed that they will.
Or, in more tabular form:
RRY : GM functionally always acts autonomously, barring extremely rare exceptions
$Y : GM is presumed to act autonomously, but might decline to at times
#Y : GM always
can act autonomously, but never
has to; no presumed bias either way
NY: GM might be able to act autonomously, but is never guaranteed to be able to
SBY: GM functionally can't act autonomously, barring extremely rare exceptions
For me, I find that the vast majority of D&D games that bill themselves as "sandbox" games are closer to NY, or (occasionally) #Y, whever we choose to call it. Games like pure no-myth PbtA or
Ironsworn are full-throated SBY, especially the latter, since it doesn't actually
need to have a GM at all. Further, most games that are railroad-y are $Y, not RRY; it's actually pretty rare to get something as utterly extreme as RRY (hence my use of "Dragonlance-style" etc. as my term for this sort of thing, as that's one of the rare places where D&D goes so far as to assign individual characters and even
spoken lines to specific players.)
