I would have thought it would be 90% engineering. Latency is the problem which causes the horrible UX for time-sensitive entertainment like sports. It's not just the stream from the server, syncing cheers from all parties would be tough.
Discussion
Syncing cheers from all parties is a UX choice, as is the fidelity of the streams.
Something like WhatsApp but where the live event is the background would a huge hit.
There are things like that which have been tried, but yet they are not popular. I'm arguing that syncing cheers would be a ux requirement. Nobody wants to have the name revealed to them by a desynced cheer.
Would you say Discord is probably the closest thing to that?
Plenty of gamers playing rt first person shooters and talking via discord. Seems that should transfer to live events?
It might be a little different. With FPS chat there is still latency, but it doesn't mater because each player's perspective is relative to their own first person view. For esports, the broadcast of the player view and commentary are coming from the same location. I think the difficulty arises when the media is broadcast from one location, commentary sent from another, and viewer receiving both separately.