GDX 2009 – Improvisational Performance and Games

by Darius Kazemi on April 16, 2009

in conferences,gdx2009,transcript

Here are my raw session notes for Brian Shurtleff’s GDX talk, Improvisational Performance and Games. This is my best attempt at a transcription of what he said. Any mistakes or misinterpretations are mine and mine alone. My comments are in square brackets.
Basic summary: game developers who are interested in narrative should try improv, they’ll learn a lot about pacing stories on the fly and making the player look good (among other benefits).
Improv Games
Small moment to moment goals that are highly unstable
Some types of improv games have goals (guessing games)
Two goals are always common: entraining people and becoming better at improv
One type of game: When an actor says a line, the host/judge will do a signal sometimes and that player has to make things more interesting. One time an actor said “here’s my wife, I’ve had three kids with her”, I decided that wasn’t interesting so I gave the signal, he wasn’t expecting, panicked, said, “I’ve had four kids with her”. That’s not interesting, do it again: “I just have birth to demon babies and we ghave to go kill them right now” — sends us on an epic quest.
Improved skills: think on your feet, creating situations that add rusprises, tension, obstacles, or progress the story. 
Entertain: yeah, it’s pretty awesome. (If you’re doing the above, you’re automatically entertaining.)
People often think improv has to be funny, but it doesn’t always have to be that way. There’s a trafition of longform improv with serious scense. But the  comedy in improv comes from its randomness and surprises.
The entertain/improvement games are meant as training but are now entertainment in their own right.
Now I’ll talk about the rules of improv games. There’s a notion called “The Rules of Improv” — not rules but guidelines. As such they’re useful tools for game devs to look at. Note that everyone has a slightly different list.
I think the most foundational rule is: make your partners look good. All other rules are based on this. partners are the other actors in the scene, but for games, maybe just “make your players look good.” We’re already good at making players look good (Tomb Raider, etc).
Another important rule is the “Yes, and…” rule. Accept a contribution, and then enhance it.
Bad example: scene with two players. A man and a woman. The woman contributes: “I am a giraffe.” The man: “That’s stupid, you’re not a giraffe.” Now woman needs to re-assert, or back down. In either case, woman looks dumb, man looks like jackass, whole situation is a narrative dead end. Improvization abandons the notion of competition. Denying a contribution is dominating them, denying them their agency in the scene.
Even when teams do Comedy Sportz, the competition is a metagame. The joke in Whose Line: “The only game show where the points don’t matter.”
Here’s a better example: “I’m a giraffe.” “That’s interesting, because I hunt giraffes.” He accepted, and added tension to enhance it. Tailor the obstacles to the player’s contributions.
More rules: don’t ask questions. Don’t die. Both are part of “Make your partners look good.” When you ask a question, you’re putting someone on the spot. When you die in a scene, you’re dead weight the actors have to carry.
There are improv games that force you to break the rules. There’s a game where you talk only in questions, several games where everyone must die.
There’s an audience-performer feedback loop in improv, where the audience is given special powers of prompting things in a scene. In this sense, the audience has agency and it makes them a kind of player. Most improv games are multiple people collaborating to tell a story.
Takeaways for game dev: improv helps you make your knowledge of story structure instinctual. In a scene you learn to change pacing on the fly.  Some games teach you to learn the exact right moment to end a scene. Specifically you have to do all this while maintaining the agency of your players.
Too often we think of stories in terms of how movies tell stories, but improv is inherently interactive storytelling (so are some aboriginal storytelling traditions). Being involved in game dev and improv I want to combine them. How do you say “Yes and” in a video game. Base AI on improv actors. But that’s far away. What if we use improv performers in games. People are always talking about drama managers — that’s what an improv actor does, but that’s basically just a dungeon master, nothing new. Why not handle improv actors to play characters in games [yeah but you record performances offline]. At a recent ren fair, there were lots of games where you throw knives at targets, etc. But there’s another game mechanically similar, but you threw tomatoes at a guy in the stocks. This actor was paid to have tomatoes thrown at him and he worked at insulting people, kind of a human attack mode. Funny, clever, vicious, really insulting: a real villain. Why throw at a target when you can throw at a villain.
ARGs have experimented with this. Urban Interactive runs arg-like experiences, I was asked to worke for them. I was a pirate in full attire and asked people to solve piratey riddles. In a sense we paced players through by being a human difficulty adjustment system: providing hints for the stuck, more puzzles for the skilled.
Other types of games: MMORPGs. We had a convo about improv stories in games on the writers sig list. In the Matrix Online, we’d write a script and have live people performing them. The guy playing Morpheus had to say certain things, but could also respond in real time via improv. When a named character directly applies to a tell, it doesn’t have to be crazy interesting, it makes you feel special. There was another team at the Shadowbane launch, but it was a very expensive feature whose profitability was hard to prove. 
Problems with improv actors for games:
Post-launch staff of improv actors is expensive. How do you handle art assets, or in the case of AI how do you program it. 
Player-run stories can be good, but improv actors are TRAINED, so player run stories will mostly suck.
You can overcome this with systems to help out players making bad stories. I have no idea how to do this!


David Sahlin April 16, 2009 at 7:02 pm

For some reason, I am now totally geeked to start GMing my pen-and-paper campaign, now.

I really liked that rule, “make your partners look good.” It reinforces the virtue of humility while still being productive as a group.


Brian Shurtleff April 18, 2009 at 5:44 pm

Thanks for coming and posting notes here about my talk, Darius. :)

Your notes are pretty spot on (although sorry if I spoke too fast and made it hard to take notes… I tend to do that.)
The only thing I see is that I actually said “attract mode”, not attack mode, trying to make the comparison between a video game’s attract mode and how the Vegetable Justice guys encourage the audience to play the game by intriguing them as well as insulting them into wanting to pay money to throw tomatoes at them.

I’ll link to this on my blog. :)

Comments on this entry are closed.

Previous post:

Next post: