Streamers Blue and Queenie recount being botted prior to Twitch crackdown

The platform is developing its machine learning algorithm to beat the bots

Twitch's purple, white, and black Twitch Glitch logo

Twitch recently announced that it is “taking action” on accounts found to be linked to follow and view-botting. According to an announcement tweet from its support account, over 7.5 million suspect accounts had been implicated in the sweep.

Botting has historically been one of the platform’s biggest issues, which manifests in a number of ways. Some fraudulent channels have purchased services which artificially inflate their following, making them appear more popular than they really are – be it to try and game sponsorship opportunities, or for other reasons. Similarly, view-botting boosts channels to the top of a category, making them more likely to be clicked on by users as they browse the multitude of creators the platform has to offer.

Bots have also been used to brigade other users – sometimes in the hopes of helping a channel out, and other times with the intent to have a channel flagged and taken down. Fortunately, it seems that Twitch has finally caught up with the practice. “A majority of these accounts were detected through ongoing machine learning technology that will continue to operate going forward,” a follow-up tweet reads. Those found engaging in botting may even face legal action.

Eager to educate viewers on the potential damage bots can cause on the platform, Twitch Partners Simon ‘Blue’ and Rowena ‘Queenie’ Flynn recount their own experiences with being targeted by bots to The Loadout.

Some time last year, Twitch flagged four accounts in the couple’s channel as bots with the manual scanning system it had been using up until recently. This naturally came as a surprise to them, considering they had no intention of cheating the system. As it transpired, a rogue viewer had been running the bots to possibly try and boost the pair’s place in the Twitch directory.

However, such was the gravity of the situation that Blue and Queenie spent the following weeks painstakingly combing through chat logs to verify the legitimacy of their other viewers, before presenting their findings to Twitch. After a public plea to their community was made for the botting to stop, the message was fortunately heard by the anonymous user, and the bots were removed.

Although the incident was resolved, it could have been disastrous for the duo. Had Twitch not been in contact with them through the Partner program, they would have swiftly been blacklisted from any form of exposure on the platform, including access to front page spots and callouts on social media. Effectively, their Twitch career would be hampered considerably.

The couple, like numerous other channels on the platform, have also been subjected to the other form of botting referenced by Twitch’s statement, follow botting. With the latest sweep, they say they have had 1,000 fake accounts removed from their channel. However, this is less of a concern to them, with Blue noting that Twitch is on top of the issue, and routinely does “an absolute purge” of millions of these accounts every couple of years.

For those impacted by follow botting, there are also community tools out there – such as that produced by ‘CommaderRoot’ – to swiftly remove them without having to wait for Twitch’s own sweep.

On the other side of the coin are those channels utilising bots to gain their own unfair advantage. Blue says this is “destroying the eco structure” on Twitch, with those looking to cheat the system for their own gain actively getting sponsorships from even some of the larger brands working with influencers on the platform.

In fact, such is the scale of the issue that Blue and Queenie have been actively reporting flagrant examples of channels self-botting to Twitch – something which they urge others to do should they come across one. Additionally, the pair have been consulting with brands on what they should be wary of when choosing and approving potential influencers.

“We would like for companies who sponsor streamers to actually go and look at the analytics of the channel they’re about to sponsor,” Blue says, “because it’s super infuriating to see these brands sponsoring botters.” Classic signs of a botted channel include abnormal, vertical growth spikes without a clear cause (these can be seen using tools like Twitch Tracker), and/or a high viewer count with a completely empty chat.

With the development of Twitch’s machine learning algorithm, however, the jig could soon be up for bots. Blue and Queenie’s experience with being viewbotted showcases the efficacy of even the antiquated system Twitch originally had in place. As Queenie says, it is “spot on.” Now, with automation layered on top, sweeps of the scale we saw last week are now possible, and will hopefully stamp out the problem for good.