14 comments

[ 4.4 ms ] story [ 47.4 ms ] thread
The blog post notably downplays the severity of the linked lawsuit. (http://cl.ly/013D2U0t2U0C)

It is thorough, especially with an invocation of the CFAA.

I thought they didn't let the CFAA turn private ToSes into criminal acts. Am I missing something here?
Although technically in the criminal code, 18 U.S.C. §1030(g) allows a civil cause of action.
A big lot of hot air (and wasted money) over nothing.
I don't see how it's either of these things? Seems like they are at the very lease violating the TOS.
Fighting them legally might work temporarily, but eventually new viewbot services will arise in other countries that can't be reached legally.

I'm very interested in what measures they use to detect viewbots now. What's stopping twitch themselves from buying a viewbot package for a fake streamer, then shadow banning all the accounts and IPs that show up? There also may be clever statistical methods to detect them.

Ad companies and big websites like youtube also face this problem. It's not unique to twitch. In fact I think most sites with user content have had to deal with spam bots at some point.

I think they rely a lot on users reporting channels that are viewbotting. It is usually pretty easy to spot even for a human. You'll notice a channel that is for some reason pretty popular despite being some unknown person. You go check out the streamer and often times in these cases there is very little activity going on in the channel itself. Not many people chatting. This is odd if a stream had a lot of viewers.

It is hard to punish a channel for viewbotting though because you usually can't prove they hired the company to view bot for them. Because someone else could easily try to get a channel in trouble that way by view botting someone else.

I think the biggest giveaway is the low quality content. Two of the biggest Hearthstone streamers are so unfunny and not entertaining at all but still get the top spots.
Oh yeah for sure! I have seen so many people like this. It is pretty obvious when you enter a streamer's channel and it looks like they just started yesterday and have no one talking in chat yet have a couple thousand viewers.
I think a good way to fight them is to use IP reputation services like http://getIPIntel.net since the vast majority of bots will be on hosting IPs or compromised systems.
Can Twitch force the defendants to give out information about it's customers?
I think eventually this will work its way out with people using popular streamers as consultants for how to improve streams, potentially with them taking a cut of the stream revenue. It is something that will probably happen as some streamers no longer want to stream or do it as a side job or something to do to reduce the overall days of streaming. I assume that although one might enjoy streaming, after enough time on twitch I am guessing streamers would like to take a few months off and this seems like the most relevant way to maximize the expertise they have gained.