I think a lot more 3-person pairings would make sense if Also, to be fair Theorel, I haven't looked at any specific 3 team combinations including you to see if any "make sense". But we should be wary of avoiding combinations that don't make sense -- good scum players are skilled enough to manipulate town into making sure that their specific combination doesn't make sense, whether it's avoiding to vote a certain player to manipulate the scum placements, or something similar. It's very possible that our actual scum team is a combination that doesn't make sense. For this reason, I think we're better off looking at scummy individuals, rather than trying to find the entire scum team in one go.
This. I've been following along on mobile and keep thinking "why are we looking at 3 person scum teams already"? We don't have a scum flip yet, and I think it would be easier to find scum teams knowing one member of the team.
I don't disagree that it's easier to find scum teams knowing one member. But I think that analyzing combinations of players is a useful tool for scumhunting, as long as the set of players being analyzed is sufficiently small. In this instance, I'm not looking for the scum-team. I'm trying to determine if certain events are likely based on the necessary implications of it (i.e. what that says about a scum-team).
Regardless this comes down to scum-hunting theory. Since I'm the one primarily analyzing combinations of players, I'll elaborate on the theory behind it. If you're not interested feel free to ignore the rest of this post. I find theory interesting to talk about, and have a tendency to get sidetracked.
The goal of scumhunting is to differentiate town players from scum players. So the question is, how do we differentiate them? Well, in order to differentiate them, we have to know what's different about them. So, what's different about them?
1. They have a different goal.
2. They have additional information.
Now, consciously they are trying to hide both of those facts from the rest of us. But just because they're trying to hide it, doesn't mean we can't look for it.
Those two things are kind of broad, so I break them down a bit.
1. Different Goal:
a. They're trying to lynch town. Here's where we look for flimsy cases, misdirection, votes on town players.
b. They're trying to survive. (Note: Town doesn't need to survive as badly as scum). Here's where we look for things which don't help town, but help the player avoid lynches.
2. Different Information (ultimately this is just knowing who's town and who's scum. But I'm going to talk about specific situations)
a. They know when they're wrong. Either trying to direct town to a town-lynch, or defend scum. They know everything they're saying is false, and sometimes that ingenuineness comes through or they try to find ways to cover their poor reasoning.
b. They know when they're right. Either when bussing or when defending a town player. They know that they're right, and so sometimes they show excessive conviction on a flimsy basis.
c. They know who their partners are, and what they're doing. They can avoid or join wagons accordingly, and often will, because the names of their partners will stand out to them.
d. They know they're scum. This knowledge can make them act differently from normal.
Anyways, there are various things people do to detect each of these cases, and ways that scum tries to cover themselves from such detection. I would argue that 2c. is the hardest for scum to cover for. Their interactions can seem forced, or be non-existent. They avoid each other's wagons or join them ON PURPOSE. And ultimately, that's what I'm looking at in those analyses. I'm looking for purpose, and if any purposes make sense. Now sometimes scum do senseless things, sure. This isn't the be-all end-all of scumhunting. But, I'm terrible at telling if people sound genuine. I am not good at all at reading people. I am good at analyzing data, and searching for patterns and purpose. So, that's what I do.