IIP-7 Playing Online: Twitch Broadcasting.
This is how we do it: We read two (or three, if we’re keen) articles on Twitch streaming from different perspectives. Then, we watch 3 different Twitch streams (particularly gameplay streams) and consider what from the texts was evident in the streams, and what from the streams might be missing from the texts. We’re meant to focus on nut just the viewing side of the stream, but also how it was produced. Ready? Here we go!
Introduction & approach.
In this Twitch comparison between Twitch features included in three articles and three Twitch channels, I have taken a specific approach to analysing the Twitch streams. In Taylor (2018), one streamer in personal correspondence to the author said the following: “I really do believe you can watch two different people broadcast the same game and have totally different experiences and totally different stories.” Given the nature of this comparison and inspired by this idea, I decided to watch three different live streamers play the same game.
I chose Elden Ring assuming many live streams would be occurring concurrently and that I’d see a lot of variety in styles of play and chat interactions. I could also control for streamer, type of play, numbers of viewers, and amount of interaction between streamer and viewer. Interestingly, the first two Twitch streamers were focusing on boss fights, so I could also compare how that affected their demeanor, entertainment value, and the reactions in the chat.
First, I collated the Twitch features mentioned in each article and created the matrix below (fig. 1) to summarize what I’d be looking for in each Twitch stream. This information has been organized based on data collected in tab 1 in the appendix. Then, I created a second matrix from my observations of the three Twitch streams to tease out the relevant aspects to focus on. These are collected in figure 2 and are based on the information collected in tab 2 of the appendix. These matrices provide side-by-side comparisons of each Twitch feature found in the articles and seen in the streams. Based on these matrices, we’ll focus below on two of the most interesting points shared among the streams: Parasocial relationships as well as the chat feature, and the economics of Twitch.
Image: Streamer 1, TeosGame
Figure 1: Twitch Features (Articles)
Figure 2: Twitch Features (Streams)
Parasocial relationships.
Each article had a focus either on parasocial relationship specifically, or on parameters of sociality. Wulf et al. (2020) took issue with the conception of parasocial relationships fully fitting the Twitch platform. They broke their concept into two parts: Virtual friendship and respectful interest defined respectively as:
An intimate relationship in which persona appears like and old friend and important companion
A relationship based on impressions that the viewers have gained during exposure to media.
The other specified definition from de Wit et al (2020) considers parasocial relationships as:
One-sided intimate connections with a media performer based on repeated encounters.
These can be based on wishful identification wherein viewers picture the streamer as a role model and want to imitate them and via emotional engagement with the streamer and with other viewers.
Sociality in Taylor (2018) is more about chat interactions, monitoring, and viewer input.
In stream 1, it seemed as though the first definition of virtual friendship was most fitting. Viewers seemed to be intimately familiar with the streamer and with each other. At only approximately 1, 200 viewers in the feed, the streamer was able to keep track of the chat and felt comfortable chastising them if he saw something he didn’t like. Likewise, the chat seemed comfortable mocking his failures and spamming memes of his face. Stream 3 had a similar feel, taken to an even greater extreme. The streamer gave personal information freely, and at approximately 1, 300 viewers, actively engaged with viewers the most of the three streams (even with bots in place to provide an assist). Notably, streamer 3 was the only female streamer I watched. Streamer 2 was the most professional and removed of the three and did not fit well in the above definitions of parasocial relationships. His interactions were loosely sociable with bots monitoring most interactions, few call outs to viewers, and rigorously monitored viewer input. He had over 12, 000 viewers on his feed perhaps making it difficult for him to do any more than that.
Image: Streamer 3, RachelKay
Chat.
The chats on Twitch are worth their own study. There were a lot of interesting things happening in there from allusions to the kind of coping described by de Wit et al. (2020), to the chat features described by Wulf et al. (2020) and Taylor (2018).
In the first two streams, the game play focused on boos fights. The first was doing a combat experiment with a difficult boss, while the second was doing a complete boss run. Due to the nature of the play, both streamers restricted the chat to no backseating (giving unwanted suggestions) unless explicitly asked. Even when asked, the viewer input was not always accepted. Particularly, streamer 1 resisted a lot of input by explaining he simply didn’t want to play in the suggested way.
Conversely, streamer 3 was happy to leave the chat open. As a result, there was a lot more input from viewers and some borderline inappropriate comments made about her gender. She didn’t react to these comments but interacted a lot with friendly viewers and was willing to act on suggestions as they came up. She had a clear persona that felt like friendship though I only watched her for a few hours and her chat felt much more like a community than the other two streams.
Image: Streamer 2, CDawgVA. Note the duplicated chat graphic overlay on the left-hand side.
Economy of Twitch.
All three streamers had each of the ways mentioned in the articles for earning financial support, such as graphics and pop ups for indicating new subscribers. Streamer 1’s was the best as it was an animated GIF with an entertaining sound effect. Streamers 2 and 3 were more subtle but still had graphic pop ups if without sound. Streamers 2 and 3 had Amazon affiliate links at the bottom, while only streamer three had ad breaks. Streamer three also provided the most support in acknowledging viewers contributions, and in giving free emotes to followers. Streamer 2 had a very strange feature where voiceovers read text written by maybe new subscribers. I couldn’t quite work out what the triggering action was for the voiceover, just that it was an acknowledgement of something.
It was evident in watching the three streams that each streamer had a different approach to acknowledging viewer support. As usual, streamer 3 provided the most fervent vocal acknowledgement as well as a graphic/sound trigger. She made a very big deal over anyone subscribing, but especially those subscribing for more than one month. It was easy to see how viewers might want to subscribe for the sheer pleasure of receiving this much gratitude.
Image: Streamer 3’s Ad Break
Evaluation.
Each of other three articles had a unique goal. For Wulf et al. (2020) it was to show what features of Twitch correlated with enjoyment, loosely defined. Taylor (2018) focused on qualitatively describing the production side of Twitch and exploring the platform’s relationship to copyright and intellectual property, as well as the public and private spheres. De Witt et al. (2020) focused on wellbeing and social support from Twitch if players were going through a difficult time. The articles collectively hit on most of the notable features of Twitch, but those focusing on the social aspect of the platform (Wulf et al., 2020; de Wit et al., 2020) might have considered spending more time breaking down events taking place in the chat and in thinking about how the chat can provide more nuanced information about parasocial relationships on Twitch.
Economically, each article mentioned how Twitch users earn a living, especially Taylor (2018). It would be worth exploring further how these economics feed into the parasocial relationships on the platform. There was a single mention of this in Taylor (2018) but it stood out quite a lot in watching the Twitch streams that the acknowledgements, bot posts, and subscription notifications are an inherent part of social interaction on Twitch. The positive feedback a viewer receives from the streamer is another way in which closer ties are built. You feel as though your support is important and invested in the livelihood and success of the streamer. This connection was surely lacking in each of these articles.
Image: Streamer 1 and his subscription GIF animation at bottom left. Animation was accompanied by a triggered audio, “Yeeeesssss!” recorded by the streamer.
References:
de Wit, J., van der Kraan, A., & Theeuwes, J. (2020). Live streams on Twitch help viewers cope with difficult periods in life. Frontiers in Psychology, 11. 1-16.
Taylor, T. L. (2018). Twitch and the work of play. American Journal of Play, 11(1), 65-84.
Wulf, T., Schneider, F. M., & Beckert, S. (2018;2020). Watching players: An exploration of media enjoyment on twitch. Games and Culture, 15(3), https//:doi:10.1177/1555412018788161
Appendix
-
Article 1:
Wulf, T., Scheinder, M. F., & Beckert, S. (2020). Watching Players: An Exploration of Media Enjoyment on Twitch. Games and Culture. Vol 15 (3), 328-2018.
Twitch governed by algorithms generated by users’ bowser history, metadata, and searching behaviour. Therefore, Twitch offers navigation through several streams via a sidebar with an overview of: followed streamers who are online; streams friends are currently watching; recommended streams; a search bar for further stream search.
Media enjoyment: Cognitive, affective, personal integrative, social integrative, and tension release motivations. Information seeking, learning to play.
Live character – suspense of a real-time event with an unpredictable outcome.
Twitch encourages streamer sympathy in that viewers experience arousal and suspense hoping for them to accomplish a challenge.
Chat: Streamers interact with users, answer questions asked in the chat, address their audience, and explain why they follow certain strategies.
Digital learning paradigm: Teachers (streamers), learners (viewers), and co-learners (other viewers) may be responsive with each other to optimize the understanding of strategy.
Set schedule of streaming times, like tv.
Parasocial relationships: Virtual friendship (VF) – an intimate relationship in which persona appears like and old friend and important companion; Respectful interest (RI) – a relationship based on impressions that the viewers have gained during exposure to media.
Social communities: membership (subscription); influence (interaction with other viewers and the streamer), need fulfillment (rewards like sociability, status, success, and gaining knowledge and skill), emotional connection to others (common history, identification with others shaped via continuous participation).
Chat can be used to gain specific information, adjust the personal chat, or adjust settings.
Findings:
Suspense = small, positive connection with media enjoyment (nonsignificant)
PSRs = positive association with enjoyment and suspense
Chat = higher levels of media enjoyment
All other factors removed, suspense was significantly correlated with media enjoyment.
Article 2:
Taylor, T. L. (2018). Twitch and the Work of Play. American Journal of Play, 11(1).
Raid – when viewers flip from one finished stream to another en mass.
Layers of production
Set design:
- Audio
- Graphics
- Green screen
- Cameras
- Triggered events such as notifications of new followers
- Chat bots
- Custom emoticons for the channel
- Customized channel page
Performance:
- ‘Think aloud’ method of externalizing thought processes
- Physical expressions and gestures
- Esport an exception
Critique and evaluation:
- Reflecting on mechanics, design, game play, “feel”, etc.
- Expert evaluators of systems
Sociality:
- Ongoing chat
- Streamers monitor this as well as the game
- Audience can give input on gameplay choices which generate a lot of engagement.
Materials and digital infrastructure:
- Technical components provided by Twitch (video codecs, storage, servers, transmission nodes)
- Hardware level - Computer, audiovisual hardware, furniture, lighting
- Software level – graphics, audiovisual processing software, bot and notification trigger systems, network functionality
Economic and commercial frameworks
- Channel subscriptions
- Revenue from ads and game sales
- Bits donation system
- Third party donation systems
- Sponsorship deals
- Amazon affiliate links
Streamer leverages social and emotional cues with language of support, appreciation, and increasing chat functionality. Set design also specifically geared toward certain types of interactions and engagement.
“I really do believe you can watch two different people broadcast the same game and have totally different experiences and totally different stories.”
Does content creation like this violate copyright? Streamers encounter problems with legal structures and understandings of games as narrowly construed intellectual properties.
Streaming an example of a constitutive and transformative activity not only a referential or descriptive one.
Public/private debate:
Caillois’ opinions on contamination of play. Gaming being colonized and our agency limited. Is there a moral argument here about the purity of our work and play? Taylor accountable to situating player practices within participants’ own descriptions of the pleasure, creativity, social connection, aspirations, and authentic experience accompanying the work of play.
Work of play transformative: difficult pleasures, enjoyable instrumentality, complex negotiations between system, self, and others.
Streamers create content and experiences for their own fulfilment and for the pleasure of other and the community can provide insight into these mechanisms at play.
Article 3:
de Wit, J., van der Kraan, A., & Theeuwes, J. (2020). Live Streams on Twitch Help Viewers Cope with Difficult Periods in Life. Frontiers in Psychology, 11. 1-16.
Multidisciplinary media – collision of televisual, computer games, internet, and computer-mediated communication.
Audience – mostly male between the ages of 18-34 years old. Might be going through mental health issues re: transitional period.
Twitch an adaptive or maladaptive form of coping.
Active participation of viewers and streamers – similarities with live sports platforms.
Features shift focus from gameplay to viewers – people create their own identity through user names and channel pages; streamer is usually visible and audible on the stream alongside the game.
Money comes from playing advertisements, sponsorship deals, contracts with live streaming platforms and game publishers, and donations from viewers. Viewers can subscribe for a fee. Both kinds of donations are usually accompanied by a pop up and acknowledgment by the streamer.
Parasocial relationships – one-sided intimate connection with a media performer based on repeated encounters. Wishful identification: viewers picture the streamer as a role model and want to imitate them; and via emotional engagement with the streamer and with other viewers.
Compared to playing games, viewers lose sense of agency in shifting to role of observer, but the presence of the streamer and the possibility of interacting with other viewers add new sources of entertainment and community.
Chat – adds option of discussing serious and personal topics unrelated to the game.
Parasocial relationships b/w streamers and viewers could put them in the position of a role model, allowing them to improve their mental health awareness and literacy.
Twitch as coping mechanism: tension release, distraction, entertainment (emotion regulation), social interactions.
Platform designed to emphasize people present on Twitch, facilitating the development of communities.
Parasocial relationships provide social support and simulate self disclosure.
Findings:
For some viewers, the streamer and the community were more important than the game.
Main motivations:
1. Entertainment
2. Ability to follow streamers
3. Become part of the community
76% always or almost always read the chat
42% actively participated
43% never or sometimes participated
Personal connection to viewers follows similar distribution.
Coping mechanism:
1. Distraction (adaptive or maladaptive)
2. Entertainment (adaptive)
3. Community (adaptive or maladaptive)
Distraction:
Evidence: Spamming
Entertainment:
Evidence: Streamer’s positive attitudes, sense of joy in community
Community:
Evidence: Streamers play games with viewers; streamers invite viewers to participate as if in a talk show; streamers view and respond to the chat; streamers acknowledge donations or people subscribing to their channel on stream. Offline chat/page features.
Invitations by streamers to share issues in the chat.
No substantial finding in viewers increasing their viewing during difficult times.
-
Twitch Stream 1:
TeosGame – Elden Ring
https://www.twitch.tv/teosgame
Observations:
Backseating seems an issue. One person in the chat mentions their comments being removed and didn’t know why but later realized the problem was that they’d offered advice without the streamer asking for it.
Lots of spamming in the chat. People get really worked up with these boss fights. The spam seems to be memes of the streamers face as either very frustrated/angry, sad, etc.
The chat loves when he gets upset and dies in some spectacular manner.
Says goodbye to viewers he seems to know.
Streamer seems pretty firm in the chat. He openly provides counter points to unwanted advice from viewers. He seems like he has a perspective on how he wants to play and is firm on it.
There seems to be some fuss about predicting but I don’t know why it’s a big deal.
Streamer asks chat for advice on how to find certain resources to defeat a specific enemy (Malenia) from a specific follower @fency_jake
There was a subscription pop up but he didn’t acknowledge it, it just flashed on screen with a sound. Acknowledgement came much later. Around 1200 people watching.
No green screen, home background. Set of stairs and a neon sign advertising a sponsor maybe? Corsair?
He has a mic on an arm extended on front of his face, and a head set on. I think there must be a light on him in the front upper right. His partner comes into the frame at one point and is on camera asking if he needs anything from town. His dogs Luna and Sunny also comes in to the frame later on.
One person in the chat does mention procrastinating on a stats exam.
His dog has separation anxiety and the chat seems to know about it.
I genuinely want to know when, with his current build, he’ll be able to get past Malenia. He is iterating with his weapon but not changing anything in the overall build. Like a controlled experiment.
He definitely has a persona that seems well-defined based on the chat. Seems famous for facial expressions, sounds and voices, regular freak outs, and puppies.
Twitch Stream 2:
CDawgVA – Elden Ring
https://www.twitch.tv/cdawgva
Observations:
Around 12300 people watching. I can see subscribers pop up on the upper right, and top gifters for donations above the chat. Amazon links below the video feed. Sound notifications for something, but no accompanying pop up to indicate what.
This person is maybe less entertaining overall than the first stream I watched. The deaths and combat are remarkable similar – they even seem to still be using the same build for weapons in the boss fight and seem intent on not cheesing any of the combat.
Does have a nice sense of wry humor – lots of ‘thinking aloud’ and talking to the game and the characters.
Because he’s focusing on a boss run I wasn’t that invested in the gameplay. I also didn’t find his narration all that compelling so the suspense for me wasn’t working here.
He has a graphic chat overlay which is a little distracting for the viewer given that we see the chat on the right anyway, but I’m assuming this is for him to keep track of a crazy busy chat.
Graphic overlays to count up subscribers on screen but no sound notification.
Mic set up is crazy good and I think a virtual background maybe from a game interior?
There was some strange voiceover as well from maybe a subscriber describing how the streamer had really been cheering him up after his other favorite streamer had ‘gone blind’? I’m guessing this means they haven’t been online.
I’m so confused about these voiceovers from the viewer – how and why are they happening?
A lot less spamming in this chat, again being monitored for backseating. I think there are bots and monitors both keeping track of date, time, chat behaviour and level that continually update the chat.
Twitch Stream 3:
RachelKay – Elden Ring
https://www.twitch.tv/rachelkay
Observations:
About 1300 people watching. First time seeing an ad break! Weird! Amazon links. Interesting pop up with sound effect for subscriptions. Vocal acknowledgement. I followed her and got emotes for free and acknowledgement from the bot in the chat.
Republic of gamers Gamepass?
This streamer is paying a lot more attention to the chat and the chat is much more conversational than the others.
Interesting – she says people clip her when she stands up #ladyproblems. She blocks the user when that happens. There are some weird things in the chat – asking how old she is, etc.
“prettiest girl on twitch” eeeeew.
“How is she so wholesome?” eeeeewww.
She’s spending some time describing Twitch Con and how to join Discord to plan a meetup. So, she’s planning a meet up.
Spends a lot of time talking about personal stuff like food preferences and eating out.
Mentions her Discord community a lot, uses a lot of positive, reinforcing language. Mentions having movie nights on Discord.
Instead of a little window overlay, she’s a cut-out directly on the screen. Earbuds wired in, large mic. Consistent rolling overlay of her other feeds – Twitter, Instagram, Tic Toc, with her handles.
She does actually go out of the frame without standing up on the video, probably to avoid clipping. She honestly sidled out from the chair so carefully.
I find her much different from the other two streamers. She isn’t focusing on bosses, but also stops and listens to the dialogue and explores for loot in between battles. I find her persona interesting but also a little uncomfortable. She’s an amazing gamer but I feel like she’s playing into some of the weird stuff in the chat and playing a bit dumb.