Quoted By:
Using some simple calculus, you can do a little math to figure out how viewers entering and leaving the stream correlates to live viewers. Anyone who's taken a calculus course knows about these rate-of-change problems. Like how quickly a basin of a certain shape fills up given a certain inflow, and outflow of water
One simple model is to assert that viewers leave the stream in a direct proportion to the number of people watching. This resolves to a nice curving line that comes to a plateau, and looks very much like the live viewer graphs. However, this pattern likely emerges on those graphs because youtube takes a long time to update and verify the number of live viewers, thus a long curving slope coming to a plateau.
In reality, it's probably a different model.
A) There is a certain baseline viewership for every stream. These are people who put the stream on their schedule and watch it all the way through. This is the solid core of your audience, it will never erode unless the streamer seriously scuffs their stream. The stream topic, and time of day, certainly affects this number.
B) On top of them, are iterant viewers who saw the video in their recommended, and clicked. These are either subscribers, or soft-subscribers (people who aren't actually subscribed but have your channel in their watch history). These viewers behave differently, and are constantly coming in-and-out of the stream, like a lake that is fed by one river, and drained by another. The maximum plateau of these viewers is is directly proportional to two factors: 1) Retention time. Which is the average time they stop by before leaving. 2) The rate of influx. This is probably a constant factor which depends on the time of day, it's the rate that people stumble across the stream.
Having a bigger subscriber base will thus increase your live viewership. But, surprisingly, viewer retention does the same thing. If you can get viewers to stick around for 2x as long, you will get 2x the live viewership....for iterant viewers at least.
