It was the discussion about whether or not NPCs exist only to frame and oppose PC actions, or if they exist and PCs can pit themselves against them if they choose. I get that you're missing the distinction here, and, to be fair, it's somewhat subtle. In the former, which appears to be how you play, NPCs have no point except to act as foils to PC actions -- they only have enough form and substance to provide suitable obstacles (or perhaps allies) to PC intent. They do nothing except act in reactions to the PCs. An NPC in this model will never have it's own agenda that it pursues absent PC involvement -- any such agenda will only exist in the event that it's needed to oppose PC intent in a challenge. You've indicated as much with statements about keeping NPCs vague so that future changes to them due to player declarations and need to challenge them are coherent.
The latter concept, though, involves NPCs that are created as if they have PC level interests, motivations, and agendas. In this version, the NPCs are acting on the world independent of the PCs, and this may be the source of conflict. This is the proposed version Max is using, the NPC as alt-PC, not merely as foil to PCs.
To bring this analysis to bear on your play example, in your version the advisor only has merit as a foil to the PCs. He was framed as a challenge, and then the challenge was enacted, but the advisor is entirely bound to the results of the challenge. He only has an agenda in so much as it exists as a challenge to the players. In this model, it's right and proper that the advisor cannot engage in mitigation, because the advisor was only a toll to challenge PC intent, and when the PCs succeeded in implementing their intent through the challenge, the advisor was defeated. The advisor cannot initiate a new challenge that alters this success, only the players can enact a new challenge that might alter this success. The advisor only ever reacts to the players.
In the other method, the advisor still has independant agendas, so the player success at the challenge is now a setback, but the advisor can now plan steps to overcome the setback and act upon them, even without the players engaging in a new contest that stakes their previous victory.
l don't agree with this. At the risk of repetition, I think it is presenting a difference in GMing technique as if it were a difference in the fiction.
The advisor in my main 4e game had his own plan and (within the fiction) his own agency. Eg at one point the PCs discovered the cavern where, many years before, the advisor had almost succeeded in seizing the tapestry before being driven off by gelatinous cubes. The even found a piece of fabric torn from the hem of his robe. (Which then formed the subject matter of the final taunt during the skill challenge.)
When you say
the advisor still has independent agendas, if that is taken literally then it is as true in my game as in [MENTION=23751]Maxperson[/MENTION]'s. The advisor has agendas indpendent of the PCs. It's just that they are all in tatters. But I don't think you mean it literally. What I think you mean is that
the GM has a power, independent of the outcome of action resolution, to narrate the advisor achieving certain things adverse to the interests of the PCs (and thus of the players).
And I'm sure that's true of [MENTION=23751]Maxperson[/MENTION]'s game. My point is that it is not a difference about the fiction. It's not a difference about the point of the advisor. It's a difference about the power of participants to establish truths in the shared fiction.
EDIT: I reread the quote and was struck by
The advisor cannot initiate a new challenge that alters this success, only the players can enact a new challenge that might alter this success. The advisor only ever reacts to the players.
The advisor doesn't react to the players. The advisor does variouos things. Some of those (eg dealing with the PCs at the dinner) are reacting to the PCs. Some of those (eg forming a goblin army to help him recover the tapestry) aren't reactions to the PCs - they take place before he or the PCs have ever crossed paths or even heard of one another.
The advisor can also initiate whatever he wants. He can try this, or that. But the players' victory at the table ensures that, whatever the advisor might be trying as far as his relationship with the baron is concerned, I as GM am obliged to narrate it as failing. This is similar to how, in AD&D, a player can narrate his PC attempting to pick the lock. But if it failed once, and the PC hasn't gained a level, then the GM is obliged to narrate the attempt as failing.
The shorthand that you favour - which, upthread, I characterised as expressing a category error - seems to me to run together
stuff in the fiction (eg the advisor tries to win back the baron's trust) with
stuff at the table (it is open, at the table, to establish as true in the fiction that the advisor has won back the baron's trust). But as soon as the game has some sort of "no retries" or "let it ride" or similar rule for finality, any such running together is just going to mislead.
For instance, [MENTION=23751]Maxperson[/MENTION] keeps saying that the advisor is "retarded" because he can't try to mitigate. Which is a product of the same sort of running together. The advisor can try whatever he wants; it's just that the fiction isn't going to change in a direction where the advisor has achieved what he wants.