On December 2nd, I interviewed Rainforest Scully-Blaker, an MA student in Media Studies. Last year, he worked on a project at Concordia’s MLAB that investigated communal play and spectatorship. Half of the experiment was conducted in a couch co-op (two players sitting next to each other sharing a controller) and the other half of the experiment was broadcast over the livestreaming service Twitch. I was interested in how livestreaming opens up the space of the experiment to online communities. We discussed the physical layout of the MLAB, the research methods in this experiment, and how the lab functions as a game lab.
CW: What motivated your research at the MLAB?
RSB: We wanted to investigate methods for studying play. The first part of that was a huge review of the literature to see what had been done. What we found were mostly generic ethnographic interview methods adapted to games. We were interested in the kind of things we could do uniquely with games. We wanted to do player studies that were more game-y.
CW: That led you to a live-play experimental environment?
RSB: Yes. One of us was into Let’s Play studies, [Let’s Plays are casual playthoughs of a game, usually with commentary] and my interest with speedrunning [a speedrun is a playthrough with the intent to finish the game as fast as possible, within a specific category of completion] had attracted me to Twitch and livestreaming so we decided to do both. One group (of ten ten participants) would play in pairs, passing a controller back and forth. The other group (of five) broadcast their play on Twitch for whoever was watching. One of our interests was in the family resemblance between a couch co-op and a livestreaming situation. In both cases, you’re not just playing for yourself. Either with the person sitting on the couch next to you and sharing your save file, or with potentially hundreds of people watching. We would also participate in the experiment, to simulate the chat interaction. Participants would interact with that, either with typing or over the microphone.
CW: How did you formalize the results?
RSB: The way we ended up turning the experiments into a way to produce knowledge was: once all the sessions were done we conducted exit interviews with everyone. We then broke up the results into four different papers. My research specifically was about Twitch and livestreaming.
CW: What game did you choose?
RSB: Ultimately we decided to pick Dragon Age. We had a couple people who were interested in Dragon Age and Bioware games and it was a popular enough game. The game has a lot of dialog trees and moral dilemmas.
CW: What was the viewership on Twitch?
RSB: The viewership averaged about 2-3 across the play sessions (the sessions were 1.5 hours long). The most viewers we had at any one point was 12, at which point it was the most-viewed Dragon Age channel on Twitch.
CW: That seems like a low number of viewers. Were you expecting more?
RSB: We wanted a meaningful viewership, so we chose a new game. By the time we got approval, Dragon Age wasn’t as new. The general consensus at the time was that we didn’t anticipate having that many viewers. We decided to do everything in our power to get viewers. If we got a lot of viewers, great, but we would work with what we had.
CW: How would the experiment have changed with, say, fifty viewers?
RSB: Having a smaller audience changed things in a number of ways. One of the claims that we make is that having a small audience is similar to couch co-op, because you’re playing for other people. In our interviews, the participants felt like they were playing along, even if the chat wasn’t active. From that, we drew on a couple concepts about spectatorship and performativity through Twitch as compared to couch co-op.
CW: So the most important part of the livestream was having an interactive chat?
RSB: The more viewers, the more spam you get, but at the same time, a quicker moving chat would have changed the experience. The chat moved at a rate that allowed people to read it and respond to it. If one participant had a session where they were interacting frequently with chat, then in subsequent sessions they would be more open to talking even if there was less chat participation. So the base level of interactivity with a hypothetical audience was still there. With a larger audience they might have gotten less done, and been more inclined to participate with the chat.
CW: How much experience did the participants have with livestreaming?
RSB: Only one of the five had streamed before. So they would have been the only one to know how to verify how many people were watching.
CW: They didn’t have the feedback of the viewer count?
RSB: They had it, but I don’t think they knew how to identify it. For the most part, I suspect they didn’t know how many people were watching.
Speedrunning and Communal Knowledge Production
CW: I’m interested in how Twitch facilitates a kind of mastery of these games that wouldn’t be possible otherwise. I imagine that having an audience to interact with affords speedrunners a kind of motivation that wouldn’t be possible otherwise.
RSB: For sure. The community has evolved from the early days of yore of the internet, the pre-YouTube age. There was speed demos archive, which was based off of an early Quake speedrunning site where you would record a speedrun, get it encoded on a CD and then send it to the archive, and then they would vet it and encode it. It would take forever. It was a slow, almost tedious process: there was just you and your TV and the game. Maybe you had a friend over, but there wasn’t a sense of an audience. Even when Youtube came along, and you didn’t have to worry about the draconian standards of how to get a run approved for an archive, you were still recording it by yourself and putting up your highlights and hoping that people would watch. With Youtube, because of search priority and google analytics, there was always a chance that the video wouldn’t get picked up. So there are definitely people saying “I wouldn’t go back to doing it without an audience”.
CW: So how did Twitch change speedrunning?
RSB: Twitch has made a huge difference in terms of the way that these communities’ knowledge grows. The more people you have watching, the more people are going to ask questions like:
“Why don’t you try this thing?”, “Why do you jump over that thing instead of just running?”, or “If you did this would this glitch happen?”
It facilitates an immediate interaction: “I don’t know, let’s try it out”.
If someone streaming on Twitch finds a glitch, it gets disseminated on Twitter and Youtube so quickly.
CW: Looking at Twitch, Twitter, and Youtube as a nexus for knowledge production about streamed games is fascinating. For the research that you were doing, it seems like there was minimal interaction with these livestreaming communities.
RSB: Yes, we were more interested in play-studies rather than player-studies for this project. For practical reasons, particularly participation, we were much more interested in observing play in a controlled lab setting. Because our end-result was a couple of theoretical concepts for how we talk about play, it would have been backwards to start with the communities on Twitch then move into theoretical concepts and then finally into a discussion of play.
There’s a whole other range of things people do on Twitch. If we wanted to be more involved with Twitch, it would have been a much longer experiment to establish some kind of viewership. For example, studying a Twitch personality become a Twitch personality, could be an interesting project. We weren’t really interested in Twitch other than as a platform for a live audience.
Space and Methodology
CW: What kind of technical knowledge was necessary to conduct this experiment?
RSB: The only real demanding part of the experiment was the livestreaming. For the couch co-op, it was as simple as putting the disc in the Xbox One and starting the game. For the livestreaming setup, I had a rudimentary knowledge of what software worked well, so I got it working and showed others how to set it up. It involved:
– Figuring out how the twitch interface works
– Installing the streaming software
– Getting a webcam and microphone setup (we used a headset)
– Setting up the drivers for the controllers so people who were more comfortable with controllers wouldn’t have to use the keyboard
CW: Can you describe the space of the experiment?
RSB: There’s a nook in the back where there’s a leather couch and a TV where we did the couch co-op.
CW: Who was present at the MLAB during the experiment?
RSB: People could come and go as they please. People were aware of the research so they would work quietly at the table. For the couch co-op, people felt more comfortable having a conversation with the people playing. For the livestreaming it was a much quieter environment. Even if you knew the person, there was the assumption that by talking they would be interrupting.
CW: Were there people in the lab who could observe the play, but who weren’t participating in the experiment?
RSB: There are four computers that are visible from the table, and two that are facing the windows. We had the participants stream on the two hidden computers. So, you couldn’t see the screen of the person playing, but the stream was visible on at least one of the four. I think people felt more comfortable looking over our shoulders than the participants.
CW: How would the research you’ve done change in a different setup?
RSB: The people who responded were interested in doing game studies, they weren’t empty vessels who were there to play a game for us with no prejudice. The incentive was that you were exposed to what the game lab was like, and there was the potential for networking, though we didn’t make it explicit. One of the things we asked was: “Do you feel that you did anything differently because you were streaming? Did the potential of being watched affect how you played?” and the corresponding couch co-op question “Did playing the game collaboratively change the way you played?”. Some participants said that it didn’t, but by and large, the response was that it did affect their play. It would always manifest itself in some way. But, as with every piece of ethnographic research, just our presence potentially impacted the way the participants played. One common response was: “I didn’t want to be boring for the person next to me, or for you guys”, but I don’t think it got the point where people felt “I need to play to be interesting from a scholarly perspective”. I think the MLAB as a space is probably the best case scenario for avoiding that. It’s a pretty casual space.
MLAB as a Game Lab
CW: What defines the MLAB as a game lab, as distinct from other kinds of media labs?
RSB: The MLAB is Mia Consalvo’s lab. She is a media studies/game scholar. When Mia set up her lab, it was cut and dry that it was going to be a game lab. We don’t have a maker-corner like TAG does, however, the game collection is much better. Whenever a game is donated, the game goes to the MLAB. It’s a bonafide repository for games. The only tools we have for doing research are computers and game consoles. I think TAG is a game lab, but not purely a game lab. A lot goes on at TAG that is not game-related: 3D printing and rapid prototyping, some very artistically minded and creative projects. I don’t think anything happens in the MLAB that’s not game-related. A game lab is a space where work is done on games. The facilities are there to conduct research on games: by either playing games or watching people play games.
CW: Thanks for this interview!
RSB: No problem, I hope it was helpful.