Back after the summer break, this Game Kitchen meet-up looked at games to be observed. Artist Adam James writes larps which include an audience, and improv games are sold as theatrical events to a paying audience. What does a game designed for an audience look like, and what does that audience bring to the experience for the players?

Before the meeting, some relevant links were posted to the Facebook event page, which are worth preserving here (with the explanations of what they are):

  • ROOM: An individual audience member is blindfolded, narrated to, and asked what they want to do next etc.
  • Coney’s salon: Jenny attended this minicon about participatory theatre. The page includes links to various other artists working on ‘playing theatre’.
  • Early Days (of a Better Nation): A Coney event which uses actors as facilitators and has the audience/participants creating story for each other by taking decisions mutually in groups.
  • Sanguine Dreams: An artist filmed a Vampire larp and edited it into a short film. Separately he interviewed one of the participants, who was explaining about the clans in the short segment that Michael watched.
  • Pandemic Live: Comedians play Pandemic in front of an audience. Hilarity (presumably) ensues.
  • The Dark Room: Audience direct a victim through a choose-your-own-style adventure
  • Choose Your Own Comedy Adventure: What it says.
  • MegaGames: Audience members playing board games against one another under an MC’s supervision, not to be confused with:
  • Megagames: An evolution of ‘Model UN’-type wargame/debate games.
  • A Mother’s Heart: A larp designed to be played with a participating ‘audience’.

Classifying audience involvement

Richard had prepared a diagram with “Audience effect on plot” on one axis vs “Audience level of activity” on the other, on which these and other such participative events were plotted.
Draft version here.

Version as annotated at the meeting:

Audience axes

[Hopefully these annotations will get updated across to the legible version!]

The activity scale goes from Passive (audience just spectates) to Suggestions only,  to Limited choices (like in Choose Your Own Adventure), to Free interaction with performers, to Free interaction with other audience members. While what might be called the plot-agency scale goes from No effect on plot (the audience are witnessing and/or exploring predetermined material) to Lots of effect (outcomes are largely determined by the audience’s choices).

It was pointed out that Early Days of a Better Nation (in which the audience, assigned into interest-themed small groups, think about policy issues and then put their points of view and decision preferences to each other) had quite a bit in common with the larp Mad About the Boy (in which players, cast into interest-themed families, think about etc and so on). And the way that (all-knowing) performers can be used to facilitate the audience in these experiences has some similarities to the ways that NPCs (‘crew’) can be used in trad larp. But the word ‘larp’ (and role-playing in general) has traditionally been taboo for interactive performances of this sort, whether because it’s a stigmatized activity or because it would strike fear into the potential audience.

Richard suggested a useful division in games-with-audience might be between those where story is presented, and those with individually experienced narratives in which the audience were controllers/creators. Michael suggested adding a division between pre-planned and spontaneous narratives. He said that performing a play could in a sense be seen as an extreme of story-game, in which each character’s actions are very firmly determined by the game designer(/author), but the director and ‘players’ still have some freedom as to how to perform the material. Storygames like Ashley Griffiths’ unpublished A Family Affair, or Witch, where you know exactly what’s going to happen but can still enjoy the process of getting there, are points towards that end of the spectrum.

Enabling the audience

In a game-with-audience where the audience is just a passive recipient of entertainment (eg. Tabletop), the only important task for performers is to be entertaining as they play. When the audience is active, though, how do you disinhibit them, to get them to participate? Comedy impro has a well-established tradition of loosening up audiences to get them involved. Keith Johnstone talks about techniques for this in his books, said Michael, including the important aspect of not encouraging the audience to involve themselves in harmful/destructive ways which will ‘break’ the experience, such as trying to make comedy improvisers fail. In The Drowned Man, crews of stagehands lurk to seize and expel any audience member who tries to interact too closely with the performers or the set.

Graham felt that for impro (and other active-but-not-too-active-audience events) a firm structure was necessary to make it interesting. Unstructured games would struggle to retain audience interest. Where ‘structure’ is about the experience rather than necessarily about the plot. Tommy  said that as a performer, having a structured plot was a useful way of signalling to the audience that there would be satisfactory closure: and Ed generalized this to signalling audience expectations. Including expectations of behaviour: when should the audience laugh, how long for? In Spain, comedy audiences only laugh and applaud at the end of the piece, not after each gag.

Michael said that from a game point of view, rules can signal expectations in a similarly useful way. Graham related this to flamenco: a highly-prescribed form, but with space to improvise within the constraints. Audience members use conventions (signals) to indicate when they wish to step up and take a spot. Michael said that some immersive theatre similarly used signals to show when to ‘join in’.


Graham facilitated a quick round of a cut-down version of the Vincent Baker pirate role-playing game Poison’d. Each player briefly defined a pirate and we played out a scene in which we decided what to do with the murderer of our captain, and decided who should be captain next. The interesting thing about this game for the discussion’s purposes was that it had a very low threshold for participation – each pirate needed just a name and two statements about themselves – and then level of activity could be set by personal preference. Some players might just listen to the others debate entertainingly, some might join in fully.

This turned out to be quickly engaging and lots of fun, both for the people who said and did a lot and for those who said and did little. Maybe in a larger group it would be frustrating if you wanted to say things but couldn’t get a word in?

Audience providing emotion

Ed is interested in a CYOA-type experience where the audience start by co-creating the character, to give them a sense of investment in them in advance. Michael raised the importance of context – the audience must know (enough of) the situation before they are asked to provide creative input. Otherwise, they will feel disconnected. So you would need to establish the normal state of the character’s world, before they start doing the extraordinary thing that the event is about. Kentucky Route Zero is an unusual (and beautiful) computer game that asks you to determine the inner emotional landscape of your character.

Ed wondered how to have the audience collectively determine the character’s emotional stance: is there a way of ‘reading the room’ to get a group emotional feel? Richard raised the Rain warmup exercise from Knutpunkt 2014 as an example of ways for audience members to signal emotions in a physically-apparent way. Graham suggested pitched humming might also work. Is there a danger that the performers will interpret and so ‘steer’ the audience response? wondered Nick. Bez suggested mapping parts of the audience onto parts of the character’s psyche.

Tommy said that with suitable guidance, audiences can be got to do all sorts of things. First ask them all to say the word “Dog”. Then ask them all to bark like a dog. Then ask them all to think about their favourite animal, and make a noise like it. They will happily do so. But if you’d asked them that first, you’d have got no response.

“Audience” and power

Ed felt the term “audience” implied a hierarchy. Bez said that at Megagames, the facilitators explicitly focus attention onto the players, to make it clear they’re not just an audience. Nick raised When the Dark is Gone as a storygame in which the facilitator similarly is enjoined to foreground the players. But Tommy felt that if the performers know in advance what’s going to happen and the other participants don’t, that creates a de facto hierarchy however you dress it up.

Ed mentioned Red Bastard, an anarchic clown-type performer at whose shows audiences expect to participate actively in some fashion, but it’s not known in advance how they will do so. Interaction level is expanded in a workshop-like way, forcing the breaking down of barriers: the performer’s role is to enable and to disinhibit the audience.

How much interaction?

If it’s desirable to have a range of levels of interaction available to the audience, how do they signal their willingness to the facilitator? Spiritualists and hypnotists are entertainers (of a sort) who use their judgement as to when and how to involve specific audience members more actively. Richard felt that Nine-Worlds-style interaction badge clips could be used by audience members to demonstrate visibly what level of interaction they were up for. (More traditional signals such as folded arms to show unwillingness, sitting on the front row to show willingness, were also still effective.)


Bez’s experience of publicly playtesting his party game In a Bind made it clear that making a public fool of himself with the game first was an important way of getting other people to take part, rather than putting them off as he had initially feared. Michael said that in Punchdrunk’s productions, masks were worn by audience members partly so as to distinguish the actors, but also so as to provide cover and allow them to interact with the show less self-consciously.

The Office Party

Mo wondered about a larp which started with a couple of people playing characters, everyone else as audience, but the audience members could become as involved as they wanted to be: assuming characters, interacting, and becoming fully performers if they wished to. Because this would be billed as a game rather than a performance, and because it would be improvised rather than prescripted, there wouldn’t be that expectation of hierarchy. Current larps with audiences (eg. A Mother’s Heart) restricted the level of involvement that the audience members could have, to a greater or lesser extent. But why not a larp which starts with performers and an audience, but by the end might have just all performers?


We set out to design such a larp, and thought a party might be a good setting: it would feel natural for people to drift around by themselves or in small groups, and they could engage with other people or not in a natural manner. Furthermore a party typically provides alibi for uninhibited behaviour. It was suggested that an office party, at which co-workers could be present along with clients, guests, family members and so on, and in which a natural hierarchy and set of tensions among the characters would be easy to set up and understand, might be a good one to try. Meta-techniques could be used to spotlight particular interactions or reactions.

Players would write their own name badges as they assumed characters. Furthermore, as they improvised, they would write ‘hooks’ on index cards. So for example if two characters are conversing and one mentions their partner, they can write “Alex’s partner” on a card, and place it on the table. This can then be picked up by an audience member who wants to play that family member at the party, and who writes a badge for themselves accordingly: “Pat, Alex’s partner”. Or less directly, if a player decides their character is secretly a fraud but someone should know this, they can write a hook card saying: “You know that Taylor is a fraud”. An audience member might create a new character for this hook, or they might combine it with another, such that it’s Alex’s partner who knows that Taylor is a fraud. And so on.

We thought it might be helpful if people could discard characters and assume new ones as they please, rather than being stuck in the first one they hooked into: by taking off their character badge, Alex’s partner leaves the party, or goes into another room. (Also, they could withdraw from performing and return to just observing if they wished.) This naturally extended to other people wanting to take up that same character later and play Alex’s partner returning. Bez suggested keeping a little notebook for each character, so players could write down the important things that other players might need to know if they picked the character back up. This seems essential for continuity, but fiddly.


So, we playtested The Office Party for about an hour: at the office of Zak Steele, superstar archaeologist. The initial characters were Zak and his office manager Jeremy. All ten attenders took part, with Michael being in the role of person who wishes only to observe and not playing a named character at all. Jenny had to leave partway through (to go to the immersive theatre production Father Dagon!), but was able to do so neatly in-character.

Debriefing the game afterwards, it was felt that it was a bit unstructured and chaotic, and might have benefited from a preannounced timetable and endpoint. (This would effectively be taking it somewhat in the direction of What Happened in Lanzarote.) An internal-monologue space would also have helped, to let people update others on how that particular character was developing/progressing internally. It was hard for observers to hear enough speech to get a good overall picture of what was going on – they had to instead choose to tune in to individual conversations. Music cues, and possibly an overall background soundtrack, might have been good. And physical props – eg. a bunch of different hats, scarfs etc – would help establish and broadcast character identity. Overall though it was felt that the game was a fun experience both for people who played characters most/all of the time and for those who didn’t: and that it might be worth developing further.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s