Please bear with me, but this entire concept of 'stakes' is quite unfamiliar to me. I know that Burning Wheel uses it, and people are using it to discuss 4E (and 5E!), but I've never actually sat down and read a rulebook that discussed its use. As a DM, I don't sit down and plan the stakes one way or another; it's not something I've ever consciously had to acknowledge.
It sounds like a bunch of meta-game stuff that I don't like, though. It sounds like it's the DM talking to the players, about what's going on in the game, and the potential outcome to actions that the PCs might take.
Burning Wheel doesn't have any formalised notion of "stakes". Nor does 4e. It's just a term for talking about what thing
of importance to the players, as mediated through their PCs, is on the line in any given encounter.
In scene-framing, "indie"-style play of the sort the anti-illusion crowd in this thread are talking about (or, at least, me and [MENTION=6696971]Manbearcat[/MENTION]!, but I think others too), every encounter should have these sorts of stakes. Otherwise it's a waste of time!
Part of PC building in this sort of game is the players signalling to the GM what will be important to them in their play of their PCs in the game. There are a range of ways this can be done. In my Rolemaster games, my players signalled this through mechanical build choices (eg if a player choose Seduction for his character, then his character wants to be in situations where developing or manipulating emotional relationships are at the centre of things) plus backstory (when a player built a character whose backstory was of a freed slave trying to achieve a magistracy in his city, I confronted that PC with situations that put his independence and social standing under pressure).
In Burning Wheel, these things are signalled primarily through the choice of Beliefs and Instincts (and to a lesser extent Traits). Because of how BW works in both player advancement and how failed checks are resolved, there is not the same sort of reason as in Rolemaster for a tight correlation between player concerns for the game, and mechanical build choices around skills etc. As an example, one of the PCs in my BW game had a Belief that agreements must be honoured, in a context where he had made an agreement with the captain of a ship he was sailing, and where a mutiny was likely. When the mutiny came, his Belief was put under pressure: what is the character prepared to risk to honour his agreement to the captain whom the rest of the crew is planning to hang?
In 4e, these things are signalled through mechanical build choices (class, race, powers, etc) and through backstory, but unlike RM backstory is in part a function of those choices, and even moreso build choices like theme, paragon path and epic destiny. Thus, for instance, when players in my 4e game build PCs who serve the Raven Queen, I confront them with Orcus cultists. (Orcus is the Raven Queen's number-one rival, as explained in the PHB entry on the gods.)
As well as signalling via PC building (which itself has an ongoing, unfolding nature - in RM or 4e players make choices as their PCs level, and in BW a player can change a Belief or Instinct at (almost) any time), players are expected to signal via play. That is, they should be making it clear to the GM what they (as players of their PCs) care about in a given situation, and the GM should be responding to that. In terms of labels for techniques, this falls under the broad rubric of "saying yes" - in this case, what the GM is saying "yes" to is the players' evinced concerns and desires.
An example of this from my 4e game:
the PCs travelled into the past and rescued an apprentice trapped in a mirror by her crazed master. In the module as written this rescue was not possible, but my game had already involved use of a wizard teleport power (suitably manipulated via a successful Arcana check) to free people from trapping mirrors, and the same thing was done here. So this apprentice ended up playing a bigger role in the game than the module author, or me as the GM who selected and ran the module, anticipated.
I then built on this in a session around 6 months later: when the PCs (returned from the past to the present) were
invited to dine with the Baron of the town they were in, there were pictures on the Baron's wall of his grandmother (resembling the rescued apprentice, but older) and of his niece (the spitting image of the apprentice). The players (as their PCs) were interested in this; and I had prepared some backstory about the niece: that she was the fiance of the PCs' wizard nemesis; that they would be told, at dinner, that she was indisposed and unable to attend the dinner; and that in fact she was off conducting a secret necromantic ritual.
When the first bit came out, the PCs - having sympathetic memories of the apprentice they had rescued - were anxious. When they learned that she was absent, they realised (from memory, via Insight checks) that she was not really "indisposed" and was in fact missing, and they feared that their nemesis - himself a necromancer - had done something terrible too her. After the dinner they therefore set off to rescue her. I then used that rescue attempt as the opportunity to spring a surprise - that the niece was also a Vecna cultist and necromancer, trying to resurrect an ancient priest of Vecna but in fact, accidentally, bringing Kas back to life. This forced the players to reassess their attitude towards the niece (and to do the same for their PCs), and
]an interesting series of negotiations resulted.
There was nothing in the PCs build choices that signalled the relevance of the niece: the series of episodes described above resulted from the players manifesting an interest in a particular aspect of the fiction
in play (namely, the rescued apprentice and then her descendant) and me building on that by adding more complexities and developments, until it built up to a dramatic confrontation, where the stakes were clear: loyalty to the niece - who is a necromancer, trafficking in undeath; or giving Kas what he wants - who is a vampire, trafficking in undeath. In the end the players were able to strike a bargain with Kas which upheld their loyalty to the niece; but in the next session they then killed the niece when her evil wizardly ways reasserted themselves. (And it is arguable that my handling of that denouement was not as strong as it could have been - there was probably scope to do more interesting stuff with the niece, but I also didn't want a sentimental resolution to her story. The PCs had rescued her, but not really taken any steps to redeem her, and so I wasn't going to give the players
that for free.)
I can see how it makes sense when you're discussing something like a skill check, in order to better describe a situation that the PCs should be able to understand ("because of the wind blowing across this rock-face, a failure by more than 5 points will mean that you fall and could die"). What I don't see is, in your hypothetical example with Vecna stealing the souls, how would the PCs know if that plan was in place?
Which hypothetical plan?
In the actual play example of Vecna, the player knew that Vecna wanted the souls because Vecna took control of the PC's imp familiar (which had the Eye of Vecna implanted in it). So the player had to choose whether his PC would let Vecna have the souls - via the imp - or whether the PC would oppose Vecna and make sure the souls went to the Raven Queen. The player was in a position to make that choice because
he had taken temporary control of the flow of souls through Torog's Soul Abattoir after the PCs had destroyed the machinery that controlled their flow and extracted their essence.
I didn't know which way the player would go (but was silently barracking for Vecna!), but he chose the Raven Queen.
If the question is how Vecna even came into the picture at all, well Vecna had been on the radar since very early in the campaign, and his wizard servant (the PCs' aforementioned nemesis) had been trying to find the way to the Soul Abattoir so as to steal the souls. So the player was hardly shocked that Vecna made a play for them! And part of his reason for implanting the Eye of Vecna in his imp was so that he might be able to get some sort of leverage over Vecna - which he did, sort of (ie Vecna acted through the imp, and hence in the presence of the PC, who was therefore able to thwart Vecna's plan).
Is the DM beholden by social convention to include some foreshadowing, such that all important choices are made with (reasonably) complete knowledge of the outcomes?
The expectation is that what is at stake should be evident on the table.
So, in the case of the Soul Abattoir, it is not a McGuffin (ie not a mere plot device). The players (and their PCs)
know that robbing Torog of his soul energy will weaken him, and hence weaken his ability to hold the chaos of the Underdark at bay. So in choosing to destroy the Soul Abattoir, and thereby stop Torog stripping the essence out of those souls, they are choosing to expose the world to the risk of chaos, as the price for powering up the Raven Queen (which is what the paladin of the Raven Queen wanted) or ending the suffering of the souls in question (which is what the fighter/cleric of Moradin and the chaos sorcerer wanted - of course, the chaos sorcerer is also not averse to increasing the world's exposure to chaos!).
The PCs who don't fully support the Raven Queen (mostly the fighter/cleric, but to a lesser extent the sorcerer too) know that the consequence of destroying the Soul Abattoir is that those souls will flow to the Raven Queen, so they also know that in liberating those souls from Torog they are powering up a god they don't fully support.
The Vecna aspect to events is something that was mostly of interest to that one particular player - the invoker/wizard - but as I said he was ready for it, and made his choice, knowing that by doing so he would bring down Vecna's wrath upon himself (which he did!).
In this sort of play, the surprise to the players is not, generally, in finding out what is
really going on. There are two elements of surprise. One is the GM escalating things (as with the niece - the players want their PCs to rescue her, but how do they feel about it now they know she's a Vecna-ite necromance?). The other is the outcomes of play - who knew that the players would do a deal with Kas, for instance? Or that the souls would go to the Raven Queen rather than Vecna? And these outcomes feed into subsequent events and stakes (eg only because the players showed some interest in the trapped apprentice did I introduce the Baron's niece in the way I did, and that in turn led to the particular way that Kas was brought into the game).
My issue here is that I don't plan out the story very far in advance. I can't foreshadow that there might be a lieutenant left to carry out the evil plot, because I don't know what the players are going to do leading up to that point.
On planning - I don't plan out story. That's what play is for. I conceive of situations that will speak to the players and motivate them to throw their PCs into the breach!
When I first introduced the idea of Torog's Soul Abbatoir - from memory, somewhere towards the end of heroic tier - who knew that the invoker/wizard would have an imp familiar (that was a choice the player made at mid-paragon), or that it would be implanted with the Eye of Vecna (something which happened in mid-epic)? Not every bit of backstory ends up being important - part of being an effective GM in my style is to recognise what has developed traction and is worth running with (eg the apprentice and the niece) and what is not of interest to the players and so isn't worth taking any further (examples are hard to think of because they've by definition turned out to be non-memorable, but one might be the precise reason why the gnolls were trying to convert a temple of Baphomet into a temple to Yeenoghu - although there is a chance that I will have a go at reviving the interest of that in a forthcoming session, so maybe it will turn out to be interesting after all).
If, as GM, I thought a lieutenant would be interesting, then I would introduce it into play and see whether or not the players bite. And then build it up, or let it drop, as seems appropriate. (This has reminded me of something that really got no traction - the first time the PCs met the cleric of Torog I mentioned above, she had two tiefling sorcerers helping her. I knew who they were - followers of the devil Levistus - but this never came out in play, and while Levistus remains relevant in play those two tieflings are completely irrelevant, and I wouldn't try to make anything important hang on those two NPCs. Some NPCs, like the apprentice or the niece, develop traction; others don't.)
Obviously, if they take out the lieutenant before moving on to the boss, that lieutenant won't be around to pick up the pieces. Or maybe they'll meet that lieutenant in town, and convince her to betray the Big Bad? Or convince her to pursue some other agenda, and abandon the Big Bad before the final fight (possibly without even knowing who the lieutenant is, but just treating her like any other random NPC).
If the PCs learn of the lieutenant and then kill her, or bring her round to their side, then it would be illusionist GMing to have yet
another lieutenant pop out of the woodwork, behind the scenes, to keep the scheme going.
As for the players persuading the lieutenant to pursue some other agenda without knowing who she is - that doesn't sound like my sort of thing. What is the players' motivation, here? Why are they having their PCs engage with the lieutenant? Is she important to them in some other capacity (as in my case, where the players, and therefore their PCs, cared about the niece but didn't know she was a necromancer)? In that case, some sort of reveal or twist would seem in order, so the players actually know what has happened in the fiction.
What if there's an evil lair, and the left wing leads to a library full of incriminating documents, but the right wing heads to the actual goal for the PCs? If they go right first, and never investigate the left path, then they'll never find the information. If there's an important document - say that it's a list of people who are on the Big Bad's payroll, and the PCs can use that information to learn that the mysterious stranger at the inn is secretly working for the Big Bad - should I contrive to make that available to them?
Because that seems like Illusionism, to me. If I did that, then I'm saying that the player's choices don't matter, and I'm going to make sure they reach my chosen outcome either way. Or does that not count as a real choice that I'm subverting, because it wasn't a choice made with knowledge of the stakes?
I have trouble with the example - why is there something in the evil lair that is both (i) meant to be important - the secret documents, yet (ii) the players don't know about it, or have no way of knowing about it except random chance (choosing left rather than right). To me, that seems like poor design.
Another part of the design doesn't seem very clear to me, either. If the information is about a mysterious stranger at the inn, why does it matter whether the PCs get that before or after they explore the right-hand path? So what is actually at stake in choosing to go left first, than right; or right first, than left? If the answer is "nothing at all", then this reminds me of what [MENTION=82106]AbdulAlhazred[/MENTION] said a way upthread about random encounters: what is the virtue of randomness over GM decision-making?
In other words, if nothing turns on whether the PCs go left first, then right, or right first, then left - if it really is just random - why preserve that choice at all?
What would be more interesting would be if getting the information gave a reason
not to go right, but also created a cost - instead of stopping the sacrifice now, say, the PCs get the chance to expose the whole evil plot and all its backers! But then that sounds interesting enough that it would seem better just to frame the PCs into that choice, rather than make it depend upon going left or right.
There's also the issue of the mysterious stranger. Do the players know about this person? Is s/he of interest to them? If the answer to either is "no", then why is the GM trying to push something that has not got traction with the players? If the answer to both is "yes", then why is the GM making an interesting episode of play - the PCs confronting the mysterious stranger about his/her allegiance to the bad guy - contingent on the players choosing to go left rather than right?
Returning to the issue of "Schroedinger's map": I don't use all that many floor-plan type maps in my game, outside of the context of particular encounters where I draw up the immediate surroundings. When I do, as I said upthread, their main function is to preserve coherence of backstory. An example that comes to mind is The Well of Demons in module H2. As I ran this, there was first an encounter in an entrance chamber, then a big encounter with some gnolls and their pet demon. Then a couple of NPCs who had been prisoners of the gnolls were dealt with.
Then, in the next phase of the module, there are three or four rooms to explore, the idea being to gather the items that will unlock a portal. The sequence of these rooms makes no difference - there is no particular continuity or sequence between them, and the main contribution they make to play is to give interesting settings for the PCs to strut their stuff (though I did build in some links to the Abyss which had broader thematic relevance in my campaign). The only reason for following the map, and so doing them in the order the players have their PCs open the doors, is because it saves embarrassment later over whether the Demon Blood Room was behind the left-hand or right-hand side door.
EDIT:
it still reeks of railroading to me. But scene-framing always treads that line, in my opinion. It's like soft railroading; you're not saying what the outcome will be, but you are forcing specific scenes on the players regardless of their decision. No matter what, the PCs will find the incriminating documents, or they will show up to the ritual just before it's completed, or whatever. Where it differs (from my perspective) from hard railroading is that the GM doesn't force an outcome, and that the GM rolls with player decisions (so if the players just abandon the ritual, it's never presented, and they never arrive just in time).
It's not "regardless of the players' decision."
The reason the PCs encountered the niece as she was in the process of accidentally reviving Kas wasn't independent of player decisions. It was in direct response to player decisions (decisions about the attitude of their PCs to the apprentice they rescued, to the Baron, to the fact that his niece was missing, to the niece's fiancé, plus also stuff about the Sword of Kas that I didn't mention in this post but that is outlined in the actual play post I linked to).
