when did you decide this yellow-clad skulker was a Vecna cultist?
I don't know. As in, I can't remember.
But I do want to reiterate that there is a significant difference between
the thought that the skulker might be a Vecna cultist - which I think I probably had at the time of first mentioning him - and
it being the case, in the shard fiction, that the skulker is a Vecna cultists. It is when the latter occurred that I don't know.
This difference - between
ideas for what might become part of the fiction and
the shared fiction per se - is what gives significance to
Paul Czege's description of the following technique:
I frame the character into the middle of conflicts I think will push and pull in ways that are interesting to me and to the player. I keep NPC personalities somewhat unfixed in my mind, allowing me to retroactively justify their behaviors in support of this.
"Unfixed" doesn't mean "no ideas". It means
establishing the fiction as part of framing and resolution, not prior to it.
I think most RPGers recognise that,
on the player side, there's a very big difference between having an idea, and establishing something in the shared fiction. Thus, when a player declares an attack by his/her PC, the player has something in mind for the fiction, along the lines of
his/her PC has struck and killed the opponent. But that idea doesn't become part of the shared fiction until mediated via the appropriate procedures, which could be anything ranging from initiative rules and action economy, to system combat mechanics, to table practices about how action declarations are handled, to another player saying "No, don't attack - we want to talk to this NPC!" in which case the player might take back the action declaration and abandons the idea.
What I am trying to convey is that the same thing can be true on the GM side. So, to relate back to an example I mentioned upthread, just because I have notes that say that such-and-such might happen at the baron's funeral (or celebration) that doesn't mean that the shared fiction does, or will, include any such thing. As it happened there was no funeral (the PCs saved the baron from the catoblepas come to kill him) nor any celebration (instead the baron collapsed upon learning of his niece's death at the hands of the PCs).
Notice how very different this is from (say) the map and GM's key that Gygax and Moldvay talk about in their DMG/ch 8 respectively: in these cases, the key and map aren't a list of ideas about elements to be introduced into the fiction if appropriate in some future context; they actually establish the content of the fiction, and provide a basis on which to adjudicate action declarations by reference to fictional positioning that is secret from the players (and deliberately so - the whole point of those games includes the players learning the GM's secrets). Notice how different it is, also, from a standard event-based adventure path (I regard Dead Gods as a paradigm of this): the notes on the sequence of events are not presented as ideas about what might happen (eg they're not the same as, say, tactical advice that might be given to a GM about possibilities for handling a particular encounter). They're presented as things that will unfold in the shared fiction (it's something like a four-dimensionalist application of the Gygax/Moldvay dungeon-mapping idea).
your use of the term "secret backstory" tends to include the idea that it is being used to thwart the players, but for many of us that may not be the case. It's more a case of campaign or world information that the players do not know, which is something that every game has. So how that information is put to use by the GM is the thing in question, more than simply the existence of such information.
<snip>
I can understand your aversion to this....but I don't think I share the opinion that it is always bad.
Your use of the word "thwart" is itself tendentious, though. It's not a word I've used - I've talked about
determining that an action declaration fails by reference to elements of the fiction (ie fictional positioning) that the player is not aware of, rather than via the action resolution mechanics. Nor have I used the word "bad". And that's deliberate. I don't think it's
bad. Rather, it's not something I really care for in RPGing.
As I said above, though, you can't play classic dungeon crawling D&D without doing it. Eg a player declares "I search the southern wall for secret doors". If the GM's map indicates that there are no secret doors there, then that's that - whether or not the GM fakes a die roll, the answer is going to be "You don't find any secret doors". That's an instance of
determining that an action declaration fails by reference to fictional positioning of which the player is unaware. Whether or not you would call that "thwarting the player" I will leave up to you, but it is an instance of application of GM's secret backstor. Hence it's not something I'm really into (and, as I posted upthread, it's not something I'm particularly good at either).
But contrast, say, the Cortex+ Hacker's Guide, which gives as an example of an asset in Fantasy gaming, "Hey, there’s a Secret Door Here!" (p 220), and gives as an example of a Scene Distinction (ie something the GM establishes at the start of the scene as a highly-salient and mechanically-exploitable aspect of the fictional situation) "Secret Doors" (p 52).
An actual play example that illustrates the same thing: when a player in my Cortex Fantasy/MHRP game wanted his ranger-ish PC to look for an ox in the barn of the giant steading, the procedure was not (i) roll a Perception-type check, then (ii) I check my notes for an account of what is in the barn. Rather, he made a check (opposed by the Doom Pool - all checks in MHRP are opposed) and, when it succeeded, established a Giant Ox asset. This is the same procedure the book contemplates being used to establish a Secret Door asset. (When he wanted wolves to help with a later action, though, he didn't need to establish a Wolves in the Great Hall asset, as I had already stipulated a Scene Distinction along those very lines.)
This also seems an appropriate context to respond to this:
I think as a player or DM I would have an issue caring enough to enjoy a game where the setting is ill-defined
The setting is not ill-defined. In fact, I would say that a setting which is generated via the methods I prefer tends to be quite rich. This is in part because it is more likely to contain content contributed by multiple authors (eg I don't think it would ever have occurred to me to narrate the giant ox; or - to allude back to another episode of play that I posted about upthread - an ancient battle between angels and demons in the Bright Desert); and it is in part, I think, because the connection between the setting and the dramatic trajectory of play is normally very transparent.
it's possible that the players don't have any desire for their characters other than to play the adventure presented to them
In that case, I would say that the idea of the GM
being constrained by player concerns/interests as expressed by the build and play of their PCs has no work to do. Likewise in these circumstances it would make no sense for the GM to "go where the action is", as there is no action in the relevant sense.
even if they do throw in some basic motivations beyond the traditional ones...."I want to find my brother's killer" and stuff like that....the DM can easily incorporate these into the AP. "Turns out your brother was killed by the Wearers of Purple" or what have you.
That doesn't sound, though, like a GM being constrained, at every moment of framing and narration, by the concerns and interests of the player as expressed by build and play of the PC. It sounds like a nod to the PC backstory as a passing event in some other trajectory of play.
I think that the players can hook the GM in any game. It really boils down to the game in question, and how the group has decided to handle what goes into it, more so than the rules system being used.
I don't fully agree with this.
There are some systems (eg Classic Traveller, at least some versions of RuneQuest, Moldvay Basic) where PC generation is almost totally random, and so building a PC gives the player almost no chance to "hook" the GM. And in some games (eg Moldvay Basic again, Tunnels and Trolls, many 1st level AD&D PCs), PCs - especially at the start of a campaign - are so
thin that they don't contain any hooks.
Although RM and RQ are both ultra-simulationist games, they have important differences, and it's not a coincidence that I fell in love with RM whereas - while I have long admired the austere beauty of RQ - I have never fallen in love with it. RM allows the player to make choices at PC build that send signals - eg choosing to give your PC skill in Cooking and Lie Detection, or in Etiquette and Seduction, tells me as GM something about what you want to do with your PC. RM also, in action resolution for melee and spell casting (not so much archery, which is a bit of a weakness0, allows choices to be made - roughly, trade offs of risk vs potential reward - which (again) allow a player to express an attitude towards the ingame situation and set stakes in a fashion; whereas in RM everything is just percentage skill checks without the same scope for player stake-setting.
It's not a coincidence that I discovered the way I like to run games in two D&D campaigns in 1987 or thereabouts: one using Oriental Adventures, where PC build - despite a strong random element - is able to send much richer signals than default AD&D (because of the way PCs end up embedded in a social environment), and where the game (via its encounter rules, its Honour rules, etc) provides much richer resources for framing situations that speak to those signals; and one involving an all-thief game, thieves being the standouts of the traditional AD&D classes for bringing built-in thematic heft and providing players (via their thief abilities) with a rich capacity to send signals in action declaration and mode of engaging the gameworld. (Only paladins, monks and druids are comparable, I think, but are (i) harder to build and (ii) perhaps more distorted by the demands of conformity to the baseline mechanics.)