Maybe I should say "how you'd determine the extent to which you should forgo scumminess for information on a given day", or smth.
I was prompted to ask this when you were talking about the hypothetical dude with loads of game-relevant posts. And Helz said his flip could unravel the game, which suggested lynching him would be worthwhile irrespective of how you read him.
One of the biggest is the game state. If its early in the game and I scum read a player but recognize them as a strong player I will be more hesitant to lynch them because I recognize the opportunity cost of being wrong is much more significant. Later in the game I hold a much lower value on their future contributions.
I also look at the game solving aspect which generally centers around how much I focus association tells. If I scum read one player more than the other but I have some very strong association tells on the second I think of it as 'lynch 1 scum vs potentially lynch multiple scum.'
Another and probably the most obvious is my level of certainty in my read. There have been a few times I was absolutely dead certain a player was scum although its been rare. Generally when I get that level of conviction there is nothing that will shift my focus.
My townreads reads is a factor that comes to mind. I like to consider my bias and triangulate my reads as much as possible and I often favor lynching a player who is not my top scum read when my town reads highly value the information gained from the lynch as well as their certainty.
I sometimes focus on gathering information from a player I know is scum especially when I have the knowledge shared. If the town is in a good position I highly value the idea of poking a known team scum for information. Smart scum shut down once they know they are going to be lynched so prods for information are more effective while the pressure on their slot is not overwhelming.
Finally I consider lynch viability. Regardless of my reads or my level of influence a lynch takes a consensus and I accept that sometimes I can just not make a lynch happen. Even as town I sometimes work to set up a player for a lynch in the future while favoring a current lynch that will yield information. I do prescribe to the thought process that lynching should be favored over no-lynch in most situations because of the information gained from lynches. Actions speak louder than words and votes are an action : )
To be fair I think I would lynch to reduce entropy if I signed up for a low activity game that was significantly disrupted in the early game by 1 player but I honestly do not think I would ever sign up for a low activity game.. I even have a list of players that I wont play with simply because their low effort hurt my experience in the past. Theres probably too many names on it..