Uh-oh, it’s happening again. The players are going off-piste. You thought they were going to zig but they zagged. They’ve got the bit between their teeth and on top of that they’ve got the wrong end of the stick – except that it’s only the wrong end according to your notes, because whatever the PCs do, that is the story.
And this is all good, because now you’re forced to improvise, and improvisation is the key to pulling down the lightning so that, instead of just lying inert on the slab, your scenario comes alive. Nobody knows what will happen now, including you. That’s what makes roleplaying fundamentally different – and often much more exciting – from traditional scripted media. It’s storytelling in retrospect, but right there in the moment, it’s life.
(If you want a good example of how real life is much more engrossing than most fiction -- in short, how it spawns better stories -- compare the documentary Beautiful Young Minds with the excruciatingly lame and formulaic fictional version of the same characters in the movie X+Y. With all its genre tropes, the latter resembles a preposterously proportioned toy breed of dog in comparison to the natural honesty of the real-life version.)
"[David Lynch] is not following some screenwriting template where the characters are going someplace, they arrive someplace, and then it closes. Storytelling doesn’t have to be that way at all, especially if you’re interested in delivering a different type of truth." -- Oisin Fagan
But for improv to work its magic, you can’t just apply it as a sticker slapped on the chassis of the planned scenario. For example, the characters are going to a secluded mansion to interview an NPC with important information. Instead of going by road, as you expected, they park on the far side of the woods and walk to the mansion that way. OK, this is a great opportunity for improv. You can build atmosphere and threat. You could have them find something sinister in the woods. Maybe somebody has tried to burn something out in the woods? Something heavy was dragged this way? Is that a mound of freshly-dug earth? And what about this makeshift tent as if somebody has been living rough out here?
But whatever you do it will be more effective if it connects to the core events rather than simply being a closed side-loop for the characters to deal with that has no bearing on anything else. Whatever emerges from this improvised woodland trek should tie back to the adventure. Maybe the burned papers contain fragments that hint at the NPC's secrets. Maybe the drag marks lead to a shallow grave that changes everything the players thought they knew. The goal isn't to create a self-contained side quest that gets resolved and forgotten, but to deepen and complicate the main narrative thread. Use the woodlands sequence as an opportunity to extemporize from what you know is going on at the mansion.
Improvising is easy for the players. They just have to think in-character and do what their game alter egos would do, and events unfold automatically from there. "What is incident but the illustration of character?" as Henry James put it. For the referee (or GM if you must) it's harder. He or she is not reacting from within the world except when improvising an NPC's choices. When in authorial mode, the risk is that the referee will improvise something clichéd -- as all writers tend to do when under pressure, which is why we see the same story tropes over and over again in TV/movie drama. Here's a trick for avoiding that. Whatever you first think of, put that idea aside and improvise something else. I ran a whole Krarth campaign that way and avoided all my bad habits. It was like somebody else was running the game, and the results were as surprising to me as to my players.
In this episode of The Good Friends of Jackson Elias, the chaps describe an example of ‘extreme improv’. A player goes upstairs in the empty house and returns saying, ‘I found the missing woman’s diary and it explains everything she was up to in the days just before she vanished.’
Now, this sounds like a passive-aggressive way of telling the referee: ‘Too slow, bored now, get on with it.’ Obviously you can’t block by saying, ‘You didn’t find a diary,’ because that discourages improvisation. So how about: ‘The diary is interesting because it’s so uninteresting. The entries are all brief and pedestrian. Could she really have led such a boring life? Then you find an entry where she’s written something that’s been heavily crossed out. Holding it up to the light you can see she began writing, “At last! Exciting news from – ”. And she broke off there and scrawled out what she'd written.’
Now the players might realize: ‘She made a mistake writing that. So she must have had two diaries, one for show and one for secrets. We need to find the other one, where she recorded the important stuff.’
We could call this improv aikido, using the player's energy and redirecting it productively. The player’s improvisation is rewarded – it’s told them what they need to do – but you don’t have to drop the clever stuff you had planned for how they find it. Of course, if the player then says, ‘Yeah, I found the other diary too; it was in the same drawer,’ then she really is hating this game and you might as well play something else.
The best improvisation doesn't feel like improvisation at all. It feels as if the story was always meant to unfold this way, even though you and the PCs are making it up as you go. When you achieve that seamless integration of planning and spontaneity, you create something that no author or writers’ room could ever achieve – the plot that emerges out of contingency and character in a wonderful cascade of events that nobody could have predicted.
