Every once in a while, the storyline of an episode of the Buffy the Vampire Slayer TV series would unfold in such a way that I never saw the resolution coming. Generally in these episodes, even the characters in the show are stumped until the very last minute (and sometimes even beyond that) about how to wrap things up. An example of such a storyline is episode 2.19 I Only Have Eyes for You. Stories like that can make for an excellent written tale, where the author has full control over the characters' actions. But these stories are very difficult to tell through the medium of a role-playing game, where characters may not say, or do, or even conclude, what the story's framer wants them to.
I knew from the outset that this adventure was going to be a challenge because it tells such a story. Because of this, I classified this adventure as a "literary experiment." I had come up with an interesting story that I felt might have made a good book. However, I wasn't sure it was workable as an RPG.
In this particular case, one of the problems was that the players had to reach specific conclusions at fairly specific times about the nature of the central conflict (the tie between Ghost and Coyote) and how to address it. Another problem was that players of RPGs tend to want to fix everything, which means that once they discovered the tie between the two characters, they would likely put their energy into trying to keep Coyote alive, or even to cure him, rather than helping him to go gently into that good night.
In other words, the adventure in my head was heavily scripted, which tends to make the characters in the RPG little more than puppets in my psychodrama. That's one recipe for a really bad RPG, and almost every GM falls into this trap at some point or another. My primary challenge then was, "How do I get the players to tell a story about 'letting go' without running their characters for them?"
As I'm writing this, it occurs to me that I could have simply told the players at the start of the game session, "This Episode, we're going to tell a story about letting go. Keep that in mind as you tell me what your characters are saying and doing." I didn't think of that. I probably wouldn't have done that anyway, because like most GMs, I like the central conflict to be a surprise to the players (at least to some degree) when it's revealed. I subconsciously dismissed this option because I thought the element of surprise was more important than securing the players' cooperation through full disclosure. If I run a similar adventure in the future, I think I'll try the full disclosure route and see how that goes.
This time, though, I handled things differently. At the very beginning of our game session, I warned my players that this Episode was "highly choreographed." By that, I meant that the adventure was headed toward a fairly specific outcome. It was therefore likely that when the characters were exposed to some of the stimuli in the adventure, the general bent of their reaction might have to fit a certain mold, although the details of their actions and dialog could vary considerably.
In order for this goal to be achievable, I'd have to make it clear to the players when we had arrived at such "linchpins" in the plot, and I'd try to emphasize (through the actions and dialog of nonplayer characters) what type of outcome we were going for. We also established that the characters would have the ability to spend a Drama Point for a "flashback" to a relevant scene if they needed help figuring out what was the appropriate reaction for their character.
So with those understandings, we played the game. I think it worked rather well, considering that it was very different from the type of adventure we normally run. However, it'll be awhile before we try something like this again, because it is so different from what we're accustomed to.