TL;DR EU regionals weekend 1 viewership was still fine, but legit viewership never cracked 80k even though the reported peak was over 150k.
Psyonix was cooperative enough to leave the Sunday stream shenanigan-free, so we have a natural comparison between the two days. Both the Saturday and Sunday streams had the same stakes- half of the EU teams played each day trying to qualify for next weekend- so one would expect viewership each day to be pretty similar, and in reality, this was true. Twitch displays total viewers for everyone to see, and the number of people in the chatroom is available through the API, and I track it. (NOT the number of people actually chatting- you’re in the chatroom if you’re logged in, viewing normally, and didn’t go out of your way to close chat. The fullscreen button doesn’t do it, and it appears that closing chat while viewing in squad mode doesn’t remove you from the chatroom either). Only looking at people in the chatroom on Saturday and Sunday gives the following:
That’s really similar, as expected. Looking at people not in the chatroom- the difference between the two numbers- tells an entirely different story.
LOL. Alrighty then. Maybe there’s a slight difference?
Large streams average around 70% of total viewers in chat. Rocket League, because of drops, averages a bit higher than that. More people make Twitch accounts and watch under those accounts to get rewards. Sunday’s stream is totally in line with previous RLCS events and with the big events in the past offseason. Saturday’s stream is.. not. On top of the giant difference in magnitude, the Sunday number/percentage is pretty stable, while the Saturday number bounces all over the place. Actual people come and go in approximately equal ratios whether they’re logged in or not. At the very end of Saturday’s stream, it was on the Twitch frontpage, which boosts the not-in-chat count, but that doesn’t explain the rest of the time.
The answer is that the Saturday stream was embedded somewhere outside of Twitch. Twitch allows a muted autoplay stream video in a random webpage to count as a viewer even if the user never interacts with the stream (a prominent media source said otherwise last year. He was almost certainly wrong then, and I’ve personally tested in the past 2 months that he’s wrong now), and, modern society being what it is, services and ad networks exist to place muted streams where ~nobody pays any attention to them to boost apparent viewcount, and publishers pay 5 figures to appear to be more popular than they are. Psyonix/RLCS was listed as a client of one of these services before and has a history of pulling bullshit and buying fake viewers. There’s a nice article on kotaku detailing this nonsense across more esports.
If the stream were embedded somewhere it belonged, instead of as an advertisement to inflate viewcount, it’s hard to believe it also wouldn’t have been active on Sunday, so it’s very likely they’re just pulling bullshit and buying fake viewers again. If somebody at Psyonix comments and explains otherwise, I’ll update the post, but don’t hold your breath. Since a picture speaks a thousand words:
2 thoughts on “RLCS X viewership is being manipulated”
Hey, I appreciate your work very much. Is there any way for me to keep track of which streams were botted and which ones weren’t? That would be very interesting to me.
If you want to track it live, the BTTV extension has a /chatters command or http://www.tidyxgamer.com/realviewers.asp is half-broken but still shows chatters. If the chatters are a significantly lower percentage than 80% of the viewers on the main stream, something is up. If you want a historical list for RLCSX of streams that were significantly manipulated,
8/1 EU Regional 1 day 1A
8/9 EU Regional 1 Finals
8/15 NA Regional 1 day 1A
10/16 SAM Major day 1
10/23-10/25 SAM/EU/NA Major Finals