Very interesting dispersal.
Conversation
Replying to
What could be the next experiment to run when the bots will start to add random.radint(a,b) delay(to fool this kind of approach)? To check that there’s a uniformly distributed delay?
1
1
Replying to
Off the top of my head:
1) NLP approaches
2) Distribution through the day (humans sleep, but this wouldn't eliminate human bot farms).
3) Checking the uniformity of the distribution would probably work if there is enough data.
4) Cardinality of subreddit activity.

