For instance, imagine how a RPG session played "story now" might actually produce the Moria sequence.
<snip>
One PC has, as a goal (whether formally established, in the manner of a BW Belief or a Cortex+ Heroic Mileston, or informally flagged as it might be in 4e) I will meet my cousin Balin in Khazad Dum.
Another has the goal "Having escaped from Saruman [in an earlier episode of play], I will thwart his desire for the ring." And also has the character descriptor (again, in BW this would be a Belief, in Cortex+ Heroic a trait, in 4e it might be an element of a theme or paragon path) "I am a wielder of the Secret Fire!"
A third has the goal "I will show that I am fit to be king, and leader of the Free Peoples".
A fourth has the descriptor "I am an elf of the woodlands, a peerless traveller".
(For the present, I ignore the hobbits and Boromir. The journey through Moria is not such an important part of their stories.)
Through whatever mechanism is being used (eg in BW it would most likely be an Orienteering check; in 4e it would probably be a Nature check in the context of a skill challenge), the players fail a check that corresponds to their safe travel from Eregion into the Vales of Anduin. The GM narrates, as the consequence for failure, that the Mountains steand before them as a significant obstacle.
The next thing that occurs in play is that it is established - the method is something I'll elaborate on - that the PCs know of two ways to get to the other side: the Path of Caradhras, or the Mines of Moria.
There are multiple ways this might be established, depending on system, mood, whim, etc. Eg the players might declare knowledge-type checks; the GM might just tell the players; etc.
Let's suppose, for the sake of this example, that it unfolds in the following way.
To begin, let's take it that it's already established in the fiction that Moria is known to offer a path under the mountains (eg this seems implicit in one PC's established goal). Gandalf's player then declares an Ancient History check, with a buff from Legolas, to establish some useful bit of knoweldge about the mines. But the check fails - and so instead (the GM explains) Gandalf and Legolas recalls that there is terrible danger in Moria, awoken by the dwarven miners. (The GM is getting ready here to play with Legola's identification as an elf, with Gandalf's identification as a wielder of the Secret Fire, and with Gimli's goal to visit Balin in Moria.)
Gandalf, therefore, cautions another way. The GM calls for another check (in 4e it would be Nature; in BW it might be Mist Mountains-wise). Again, it fails, and the GM narrates, "You know of the Pass of Caradhras, but the snow seems to have set in early this year. It will be hard going." (In Dungwon World that's what they call a "soft" GM move.) Aragorn's player advocates for passage through Moria, but Gandalf's player encourages the group to take the pass. In BW, this could be resolved as Duel of Wits between the two PCs (which would obviously implicate Aragorn's Belief about leadership); 4e doesn't have a comparable mechanic. In any event, the group resolves to take the pass.
The players then make a group Athletics or Endurance check (maybe both, maybe one, depending on system and what the GM calls for), but it fails, and the GM narrates the snow all around. And, picking up on Gandalf's character elements about Saruman and about the Secret Fire, adds in a hint of magicsal malice to the description. Gandalf's player, playing the fiction in the sort of fashion that [MENTION=16586]Campbell[/MENTION] has talked about upthread, declares an Arcana check in response to try and tame the storm. This also fails, and the GM informs the players "Your hobbits will die unless you turn back." So the players decide that the party turns back. Now they have to try Moria.
Free narration gets us to the west gate, but the GM describes it as closed. A failed History check by Gandalf's player is narrated as him not knowing the password. Frodo's player offers an after-the-fact augment (in our 4e game that is acceptable if an action point is spent), but the success comes at a cost: the watcher stirs and attacks the PCs. There are different ways to do success at a cost - in BW, it is one way of establishing "fail forward" narration; in 4e, it could similarly be part of the narration of a skill challenge.
The PCs retreat into the mines without defeating the watcher, leaving it free to block the door behind them. Gimili's player then makes a Circles check (in BW) or perhaps a Diplomacy check (in 4e) to make contact with the dwarven colony. But this check also fails, and so the GM narrrates an undesired consequence instead - the dwarves are dead, killed by orcs with drums in the deep! (This narration plays on the Belief written for Gimili, and also is another "soft" move that cumulates with the earlier one establishing the danger in Moria.)
It is now clear that Moria is inhabited by bad things, and so the players declare, or the GM calls for, a group Stealth check as the party crosses to the east. The player of Pippin fails, and the GM narrates this as him carelessly dropping something down a shaft. The drums start up! (Another "soft" move, that further ratchets up the stakes of failing in the attempt to travel through Moria.) But the group as a whole succeeds on the check, and so they aren't immediately attacked.
There are different ways to imagining the fight scene being framed. One is that another check - perhaps a Dungeoneering check, for successful navigation - is failed, and it is the "hard" move made by the GM in response. Alternatively, the GM just frames it as a consequence of what has already taken place, but - because the group Stealth check was on balance a success - allows the players the advantage of being attacked in a defensible position (a room with a door) rather than pinned in an open hallway.
The fight is a success for the PCs, although the GM is now pouring on the pressure, and Gandalf's player has him cast an Arcane Lock spell to hold the door against the implied hordes beyond. The spell is broken, though. There are different ways to imagine that happening. In 4e, the GM is free to have introduced a monster into the situation (in this case, a balrog) with a "spellbreaker" ability. In BW, this would more likely be the result of another failed check - perhaps not everyone succeeded on the Speed check to make it to the final bridge.
However exactly it comes about, the final scene of Moria is framed as the PCs trying to flee across the bridge while Gandalf holds off the balrog. In BW, Aragorn's player makes a Command check to break the hesitation the other PCs suffer from the balrog, so they are able to flee. (And this speaks directly to his leadership Belief, earning him a fate point.) In 4e, Aragorn is probably statted as a warlord or hybrid warlord, and uses some power to buff his allies' movement, so they are able to flee.
Neither BW nor 4e has a "pyrrhic victory" rule which would enable Gandalf's player to buff his attempt to hold off the balrog by risking his own life (but such a rule is not purely speculation - HeroQuest revised does have one, and Cortex+ Heroic has options in the neighbourhood). So we have to assume that Gandalf's "shatter" effect is subject to an interrupt from the balrog (which is part of the 4e mechanics; and in BW a lot of action resolution is simultaneous following blind declaration, so the balrog can declare "ensnare with whip" while Gandalf declares "shatter the bridge"). And so is dragged down even as the other PCs get away.
What does that example show?
First, it illustrates how important failures are in "story now" RPGing, as they generate the unwanted consequences that drive things forward in ways that are unexpected, in some sense undesired, and yet continue to speak to player-established concerns.
Second, and related, it reminds us how the trip through Moria is a story of failure upon failure - as Aragorn later laments. By my count (with fails and successes bolded in my account) there are at least 5 failures, interrupted only by a success with a cost, before the players eventually succeed at a combat. The final confrontation is then another success with a cost (ie Gandalf dies). It would be quite unlucky to get this happening in 4e, as 4e is quite a mathematically generous system. BW is capable of giving this sort of thing, though. It is mathematically pretty brutal.
Third, it shows how "no myth" works. From a bit of backstory and some Beliefs/descriptors, the participants at the table have all that they need to establish a setting with mountain passes, magically sealed gates, watchers in the water, orc-and-balrog infested halls where a dwarven colony has perished, etc. But at no point is any of that stuff pre-given: had Gimli's player's Circles check succeeded, for instance, then the fiction would have unfolded completely differently. The dwarves would have been able to guide them through Moria. Because of the earlier "soft" move in which the GM established that there is danger in Moria, some sort of check would still have been required - there's no real point speculating what sort of check, because we don't know how the ensuing interaction with the dwarven colonists would have gone. All we can say is that the story would have been very different.
<snip>
Failed checks is how, in "story now" RPGing, adverse consequences become part of the fiction. This is why PCs don't always get what they want; and why players' plans don't always work out. (It's not because they guessed wrong about what is in the GM's notes.) The imagined Moria recount shows how this can happen.