On Scumhunting

Part of Mastin Academy.

Original Lecture: On scumhunting

I'm never really able to pinpoint scum as town myself, and have always just followed up on cases other people started.

That's actually a great start!

Scumhunting doesn't need to--and shouldn't--be done from scratch. My reads are massively, MASSIVELY influenced by the words of others. Building off of what others have done is an entirely natural thing to do! I'm generally good at reading people I engage with heavily, so when I townread someone, when I truly trust a player and that player has a read on someone I don't have a strong read on...it makes a HUGE difference on my read.

Of course, I am not a sheep; I do my own analysis and form my own conclusion. But I still read the words of others and take them to heart. I listen when they voice their opinions. This isn't blind faith; their reads may not be better than mine. But they influence mine, heavily. My analysis is done knowing what they have shown, building off of their own analysis: what I see, what I don't, and what could be there, weighing the scales to form an ultimate conclusion.

Now that said! I don't exclusively rely on others. I value their input and prefer having it. But it's not necessary for me to have player feedback. Ultimately, I will call things as I see them, making my judgement off of the content of players. Said opinion may be painted partially by others' interpretations of it, yet when all is said and done, I am still analyzing everything by myself.

It's never perfect, but this is obviously easier and more accurate if there's a reasonable amount to work with: neither too much content to sift through nor too little content available. Too much/too long, reads get corrupted as biases and paranoia set in. Too little/short, and I lack the grounding to form anything solid.

For this reason, I typically aim for the whole picture. When forming scumreads, ideally my scum list is under double the number of scum in a game. (E.g. 2 scum, ideally 2-3 scumreads; 4 scum, ideally 4-6 scumreads.) If your scumreads exceed that amount, something has gone horribly, HORRIBLY wrong in your approach; scum aren't the majority; town are. So you should always have just as many if not more townreads than scumreads.

I also recommend delving into scumteam theories from both interactions and general content when given sufficient levels of information at your disposal. Of course, be realistic; you can't expect to know the full picture or catch all the scum. Just find what's individually most likely and cross-reference it with what teams seem likely and figure out what causes any clash between the two. (You do need to put that thought into what causes the clash; never discard one and/or arbitrarily assume one is correct.)

Just as important as discerning what is true is figuring out what isn't true. If you have a reasonable theory that seems plausible enough yet a flip brings its validity into question, you need to investigate if the theory was completely wrong, or simply slightly off and in need of a tweak.

Simply put--the strongest aspect of my play is that I play the WIFOM game; the weakest aspect of my play is that I play the WIFOM game. Nightkill analysis, Votecount Analysis, I do both, but they're educated guesswork at best. I try to read things for how they are, but ultimately, it's only my own take on things.

My perspective on the game is that everything is WIFOM, and the job of a town player is simply figuring out what "level" is actually true, including alternatives: most people think of wifom as one glass poisoned and figuring out which, yet a good town player can recognize it's possible BOTH, or NEITHER, are. Sometimes both ideas are (at least partially) correct; sometimes neither idea is correct.

Going on a slight tangent (vaguely related to the concept), I consider gut and logic to be one and the same, different aspects of the same concept, where I run through things to make sense of them. When I read a post, I might have a gut reaction to it--logic is what makes me search for what. When I compile a bunch of analysis, it's not my head which tells me what the analysis means. (That requires an objectivity that simply doesn't exist in mafia games; what is true in one game isn't true in another.) It's driven by my feeling, my "gut", of what's right.

So logic, reasoning, looks at posts with the overall picture in mind, keeping an eye on interactions and mindset, looking for patterns (patterns are everything!), then analyzing all of it...and then going with what feels right. Though there are a few basic rules I work by, mostly they're more accurately defined as "guidelines", giving my general outlook on games, and it is through my gut I am given which path to follow.

Ultimately, there's no universal method for finding scum. These are just tips and guidelines. Among my most important? Recognizing scum thoughts, since picking up on their mindset/motive/intentions are incredibly important to identifying them. A scummy player can be town whose intentions were good; a townie player can be scum who looks good but is pushing a scum agenda. Having meta knowledge helps refine guesses, and noting key interactions can assist even further.

By reading how others interact with one another, and by seeing what they flip, and by guessing at what their mindset was when making those interactions, you can form increasingly-coherent reads that make progressively more and more sense as fitting. You'll never get a picture-perfect scumteam; there's no such thing. (If it looks too good to be true, it's not true.) But in general, you'll get a scumteam that looks like it's plausible enough to maybe actually be true.

Further Reading

Interactive Tells talks about interactions you can pick up on.

Recognizing Reads talks about the process of giving out your reads, which ties into...

Mafia as a Social Game, where I talk about, well...talking to others. The article there mainly focuses on convincing players, but I also recommend it for reading on how to work with other players and incorporate their feedback into your own reads.

I realize those articles are a bit long, but that's because they go into these things more in-depth than I can in this session. They're for reading if the tips I present in here aren't enough.