[Foivos]

I imagined a solution where authorities would notify the hosting company of the IPs that are streaming. It should be obvious for the hosting company which customer is using these IPs for streaming illegal content just by studying the traffic pattern, no need to actually look inside the packets.

Then they can just ban this customer. That way the authorities will not have a reason to ban IP ranges affecting the other customers.

[regularjack]

Wouldn't the traffic pattern be similar to watching Netflix?

[Foivos]

I think live video has a bit different pattern than video on demand.

But aside from it, it should be very obvious: A) you are notified by the intellectual property holders that somebody is streaming pirated content, B) a specific customer or set of customers, who are not a known streaming service, are serving tens or hundrends of IPs with video and C) these customers do not have much activity during other times.

[joseda-hg]

So not Netflix, but Twitch?

Plenty of people stream commentary to matches without showing the game itself, so that would flag as guilty too

[Foivos]

These are not peer to peer connections. These people would send a single stream to twitch and then twitch, a known streaming service, would stream it to their viewers.

In theory someone might rent a server and do the streaming directly to his viewers, without using a known platform. This would be a legitimate false positive as you describe. But this would be so expensive I doubt anyone would do it when the alternative is a free platform with built in community and monetisation tools.