Most of Burning Wheel's GMing advice - for action resolution, for scene-framing, for backstory and campaign design - is aimed at this. It interacts with a PC build system that requires players to provide clear signals about goals for their PCs.
"Fail forward" is a fairly important element in this, because it is the narration of failure by the GM that introduces the additional content into the fictional situation that establishes the conflicts between a player's goals for his/her PC.
<snip>
What is maintained is not "momentum towards a goal". What is maintained is
momentum, pure and simple.
<snip>
In my actual play example, finding black arrows rather than the mace doesn't maintain momentum towards the goal of recovering the mace. It significantly reduces that momentum. But it generates momentum in another direction, namely, bringing the potentially conflicting beliefs and goals of the PCs closer to a crisis point.
<snip>
The only thing that "fail forward" allows players to rely on is that their PCs will always be confronted by interesting and engaging challenges. And that is something that I
do want the players in my game to expect. If I'm not delivering this, then I'm failing as a GM. (More on this below in this post.)
<snip>
(given that systems which emphasis "fail forward" are also likely to use some version of Let It Ride, or scene-based resolution, which means that repeats aren't possible), every failure is costing something vital to the PC and his/her goals.
<snip>
Note that this does not depend upon multiple competing goals, because the goals are only competing
as a result of the failure. I think that that is clear in my actual play example.
<snip>
If the mage PC in my game had been successful in finding the mace in the tower, he would still have had complications and difficulties. Just different ones, related to whatever goal the player authored for the PC to replace the "get a mace" goal. (In BW all PCs have three Beliefs at all times, and a player is free to change any Belief at (almost) any time.) The difference between success and failure isn't about whether or not the PCs have challenges in front of them, but whether the unfolding path of those challenges is broadly reflecting the PCs' desires and goals, or thwarting them. A dramatically satisfying story tends to need a bit of each - constant failure can generate bathos, just as constant success can generate Mary Sue-ism.
<snip>
The fun of "fail forward" play is the creation of dramatic narrative by way of RPGing. Eero Tuovinen
puts it well:
The fun in these games from the player’s viewpoint comes from the fact that he can create an amazing story with nothing but choices made in playing his character; this is the holy grail of rpg design, this is exactly the thing that was promised to me in 1992 in the MERP rulebook.
I would hope that my actual play examples also illustrate that this is what the fun consists in. (Of course "amazing story" is a matter of degree - but B fiction is much more aesthetically engaging when you are spontaneously generating it as a participant with your friends.)
<snip>
What happens if, from the point of view of the players, the GM sets the wrong stakes? The answer is, if this happens repeatedly then the game will suck. That's why, as Eero Tuovinen points out in the same blog I already linked to in this post,
The GM . . . needs to be able to . . . figure out consequences. Much of the rules systems in these games address these challenges, and in addition the GM might have methodical tools outside the rules . . . and pure experience (helps with determining consequences).
Ron Edwards also
addressed the issue in a post about scene-framing techniques:
It's not the distributed or not-distributed aspect of situational authority )ie who gets to frame scenes and set stakes] you're concerned with, it's your trust at the table, as a group, that your situations in the S[hared ]I[maginary ]S[pace] are worth anyone's time. Bluntly, you guys ought to work on that.
<snip>
As I noted upthread, there are also some GMs around who build boring dungeons. But that's not an objection to classic D&D in general. All games that are based around a GM require that GM to have the appropriate skills. (In my own case I think I'm not especially good at dungeon design, but am not too bad at narrating "fail forward"-style consequences that keep the game moving in a manner that engages my players.)
<snip>
flexibility with respect to backstory, in combination with the general frailties of human preparations and anticipation, mean that there are absolutely always consequences that can be imposed without contradicting the established backstory.