You are viewing the Wiki. To play the game, visit the forum.

Interactive Tells

From MafiaWiki
Jump to: navigation, search


Original Publication: August 7, 2013 by Mastin2

Last Revised: August 24, 2015 by Mastin2

Original Thread.


Recently, I was requested to give interactive tells. I told the person asking for these "partnership tells" doing so would be difficult, due to the nature of tells:

When any credible tell is published, it stops being a tell. Scum become aware of it while town don't think of it, making town caught by them and not scum. If reversed into a towntell, scum use it for convenience and towncred, nullifying the tell. (Note: There's also the question of whether the tell was ever credible to begin with.)

However, some things never become obsolete. The problem is, the reason they're not obsolete is that they're subtle, subjective, and highly circumstantial. So I can't exactly give a full guide on them. But this article was created to give a few pointers.

Finding town's important.

Got a ton of nulls? You're doing it wrong. Got twice the number of maximum possible scum in your scumlist? You're doing it wrong. The town makes up the majority of the players, so you should be paying attention to player interactions and looking for town and scum. Why?

Because Process of Elimination is a DEADLY tool in your arsenal. Many games aren't won by lynching scum, but rather, by not lynching town. Being able to tell who is town and seeing town interactions is key to this method. Now, obviously, just because you're aiming to not lynch town doesn't mean you don't want to lynch scum. You're still aiming to hit the most-likely scum player, but that can sometimes be a secondary objective compared to the primary of not lynching town.

Mindset's tied to interactions.

The two are practically the same. Get into the players' heads and think, "Why are they interacting with that player?" Is it because they're town? Because they're scum? Or is it something they're doing because they are who they are? (Meaning, regardless of alignment; thus, null.)

Understanding how each player thinks is a vital objective in knowing what to look for. If you can see where they're coming from, it's easier to decode their posts. Recognize their intentions and you'll know what spawns their interactions.

Tangent: Meta

This is why meta is so prominent on the site. It's much, MUCH easier to understand the mind of a player if you've seen their previous games. They give you an idea of how that player works in general (things unique to them as a person, thus not alignment-specific), as town, and as scum. That said, tread carefully--

Experience trumps reading.

It's one thing to read second-handedly a player's past games. It's an entirely different thing to have been in that game first-hand as things were unfolding. Generally, I trust meta more if the player giving it was personally present during the time events unfolded, because it gives a more accurate view of the player.

The more the better!

The larger your sample size, the more accurate it'll be. If you've only got one game of town/scum, that's not meta. That's just a random town/scum game, which doesn't give a greater indication of the player as a whole. Yes, sometimes there's nothing else, but you should use what you can, because meta is meant to show a consistent trend across town/scum games, for town/scum mindsets and thoughts in general, made easier with multiple examples of both.

If lacking a large sample size, you can still use meta, yes, but use it with the understanding it'll inherently be MUCH weaker than with a larger sample size, thus, should be taken with a grain of salt.

Metas change!

ESPECIALLY if sufficient time has passed, the player is new, a Village Idiot who wants to stop being one, or most critically, is aware of their meta. This is another reason large samples are handy: it helps show what doesn't change.

Ultimately, meta's a tool.

Not the end-all of scumhunting. It has value, ESPECIALLY at reading the mindset of a player, but it is only a guideline. It is not absolute. It's there to help you read a player, not to be the entire basis for a read on a player.

See Also:

Meta: Analyzed, an article written detailing meta in greater depth.

"How" is as important as "why"!

As is the "when", for that matter. Circumstances are everything, and the WAY a person interacts with someone is arguably even more important than the WHY behind them doing so.

Scum love bussing.

And hate being caught on town lynches. Yes, there'll always be exceptions. But generally, when I look at a town lynch, I'm going to mainly focus on the players OFF the wagon to see why they were.

"This is counterintuitive.

"Surely, scum would be absent from scum lynches and present on town lynches!"

Not in current site meta--scum are overly fond of bussing, fear being caught lynching town, and obsess about maximizing towncred. Should they be, No; they should keep long-term objectives in mind, but I'm speaking from experience: most don't. They advocate lynching scumbuddies by thinking them scummier than they are, always viewing themselves as having a weakest link.

Basic hints:

A lot of the time, scum know a player is town and didn't want to be on. So, some advice--

  • If the player encouraged the townlynch wagon (even subtly) but stayed off it, they're far more likely to be scum.
  • If the player consistently defended the townlynch wagon throughout the day, it's nullish-town. Many players (myself included) find would-be-mislynches easy townreads, becoming frustrated that their words are being ignored and that a townread's being lynched. Look for that frustration. Look for that resistance, that fight. It's often the distinguishing factor between a player knowing a player's town and not caring about the lynch (scum) and "knowing" a player's town and fighting to prove it (town).
  • If the player was going back and forth on the wagon, analyze why. Look for those same signs of frustration! And add in confusion. Scum when they "can't get a read" are generally "am indecisive", but lack the emotion. Whereas a town player not getting a read HATES it, and cannot settle down on one. Look for that.
  • Scum manipulate. Look for signs of actions not matching words, if their ultimate goal is different from what it seems. Elite scum players will seem like they're pushing a town agenda, while secretly furthering a scum one, often evident in wagon formations.
  • Don't rely on generalities. Sometimes, all the scum WILL be on a wagon. Sometimes, all the scum will be OFF a wagon. Don't make arbitrary calls as to how many scum are on or off. Instead analyze the reasons why a person is on a wagon, or (very vitally, especially in the case of a townlynch wagon) why they're not on the wagon. (But, uh. They're called generalities for a reason. Keep it simple.)

The guidelines (not rules!) for lynches on scum are different.

  • Bussing scum want VCA towncred. If a player backs out at the last minute from paranoia, generally, they're not scum making a last-ditch effort to save their scumbuddy. Instead, usually they're town who suddenly got a bad feeling about the lynch. (Note: again, emphasize GENERALLY. This also only applies to current site meta.)
  • Bussing scum seriously want VCA towncred. Thus, are slightly more likely to be on the earlier half of the wagon. (Note: This changes every year or two, though. In 2015, it has swung back around a bit.)
  • The above tips act as guidelines for running through the entire procedure, of analyzing why a player is on or off and looking for emotions, manipulation, and signs that this player is scum bussing.

Note that while all these tips are more useful in single-faction games, they aren't invalid in multiball; they simply don't give the same accuracy. Also, cannot stress this enough: These tips only work in the current site meta. If scum played the way I actually advised them to, these tips wouldn't work at all. And via their publication, these tips are likely to be obsolete within a year or two. (Maybe sooner.) They'll work fairly well RIGHT NOW, but scum WILL adapt.

Scum really love bussing.

Cross-bussing, double-bussing, triple-cross-bussing, double-decker-bussing, I'velosttrackofhowmanytermshavecomeupaboutit-bussing, you get the idea. Never assume the fight was so intense that it could never be scum bussing.


Though the above tip should be kept in mind, don't violate occam's razor and assume they're scum. Keep in mind the mindset. Sometimes, scumbuddies really don't get along and flat-out hate each other. They are the type most prone to these elaborate busses, so think about the players' playstyles and if they're heavily in clash with one another.

In general, the simpler interaction will be correct, so generally, you can conclude that what might be a triple-cross-double-decker-whatevertheheckyoucallit-bus, probably isn't. Yet always keep in mind the circumstances, because sometimes, it CAN be.

"What about other tells?"

Things like buddying? Emotions? Rage? Arrogance? Surely, there are tells for those, right?

Unfortunately, none that can be generalized. They are highly, HIGHLY circumstantial, more than any other type of interaction, and are largely player-dependent.

However, I do have this:

Scum usually don't buddy scum.

So if a player buddies up to someone that flips scum, chances are generally much higher that player is town. (Multiball excluded, for obvious reasons.) Yet that's it. Town buddy town just as much as scum buddy town. Town have emotional outbursts just as much as scum; it's mostly person-dependent.

Does this make those things impossible to analyze, though? Heck no. Just keep in mind the tips for mindset, meta, the how/why, and you can use said interactions to potentially read a player. But there are no predefined tells for it, none that will work often enough to actually be called a tell.

Don't discard lynched town players.

Ask yourself why they were lynched. Sometimes they had good reads, and good reasons! So don't blindly assume, "Oh, they were lynched, they must be awful; I can ignore them." On the contrary, analyze their posts closely now that you know they're town-aligned, with a town-driven mindset, so look at their town interactions with others.

Yeah! Sometimes, the results will be, "They were an idiot that deserved to be lynched. There's nothing of use in here."

...But often, the results WILL surprise you; they actually had some valid points, points which in confirmation bias were ignored by others. And if so, then you have to carefully wonder how they got lynched, despite their valid points.

The result is, more often than not, scum interference. That subtle manipulation.

NKA's just as vital as lynch analysis.

Yes, tread carefully; nightkill analysis is a wifom minefield. It's easy to get lost, since usually, there's no single reason a player dies; generally, it's multiple factors combined. So it's impossible to be certain.

...But it's not something to be flat-out ignored. If you can figure out the reasons why that player died (common ones being: obvtown, reads dangerously close, is charismatic, avoiding the dangerous roles, is a rational player capable of changing their reads, and might be a PR), then you can better figure out who killed them. Who had the most to gain from their death, and why? Who is more likely to kill them? What does the kill accomplish? Was it random or planned? Things like that.


Don't put too much stock in it.

You don't want to ignore NKA, but you don't want to put much faith in it. Let it contribute to a read, not be the entire basis of a read.

These tips are lategame-geared.

They're mostly for after you've had at least a couple of flips, a lynch or two and a kill or two. Interactive tells before a player's flipped is dangerous territory. A lot of players (myself included) DO it, but I take great caution!

Reads to evidence!

My golden rule applies here; always fit the reads to the evidence, not vice-versa. So if you DO use interactive tells pre-flip, you don't hold bias.

Keep things simple.

Not necessarily "the simplest explanation" as in occam's razor, but simple. When it comes down to it, scum aren't master plotters. They're humans, who have basic instincts. Figure out those basic (that is, simple) instincts, those basic patterns, and you'll be much better off.

(But, uh, do apply occam's razor. The simplest explanation may not always be true, but it's going to be true more often than not.)

Work with other players.

Often, they'll have a perspective on a player that you lack. Their input can be quite valuable.

Grains of salt!

Players giving reads can be scum (who have an agenda) and/or their experience may not be accurate, especially if their sample size is too small. So don't blindly accept their input, just as you shouldn't blindly ignore it. Filter it; take it into consideration.

The Big Picture:

Never forget it, never focus on it. There's more than one scum in the game, so be aware of that, but also don't expect to realistically catch the entire scumteam(s) early-on. You're not a scumhunting god; you WILL make mistakes. That doesn't give you an excuse to only have one scumread, though. You should have at least as many as you expect there to be in a game. (Two for a micro, three for a mini, and numbers after that are all approximations: ~four for 16-20 players, ~five for 19-24 players, and ~six for 22+ players. Yes, there's overlap. Again, these are approximations.)

If you have too few scumreads, you're being equally as negligent as if you have too many.

Yeah, it's okay to have too many/too few for certain stages, but the key?

Work HARD to fix faulty numbers.

If too many, narrow down. If too few, work to include more. Never, EVER leave yourself with that number of too few/many, because it'll lessen the value of your reads SIGNIFICANTLY. (Also, it opens the door for others to call you scum. Too few, "Because they can't think of other players to mislynch" or similar. Too many, "Because they're scum keeping their options open!" Admit it, you've used that accusation as a player yourself, so you're not immune to being accused of it.)

Another part of having the bigger picture in mind is obviously interactions. Let's say you're in a game with four scum, and you have four names which individually look like they make a perfect scumteam...but their interactions make no sense.

Chances are ASTRONOMICALLY high that you're wrong, because interactions typically trump individual scumminess. (Hence, why not forgetting the bigger picture is important.)

But also keep in mind that you're not omnipotent. You can't know how the scum are thinking for certain. You can't read their quicktopic/PT. You can't know how they're treating each other for certain. You can guess, and guess reasonably well, but there's always that element of doubt. (Well, SHOULD always be that element of doubt.)

Meaning, you could be wrong. They could actually interact that way, even if to you it makes no sense, because you lack info that they have.

"You're contradicting yourself!

"I'm confused! What do I do?!?"

Circumstances are everything. Analyze, and analyze carefully based off of that context. Ask if it's possible that you're wrong about the interactions making no sense. Ask also if you're right about the interactions making no sense and wrong about the scumteam involving the four. Explore both equally, and form a conclusion based off of your findings.

Often, it'll be a bit of both!

Depending on how good a scumhunter you are, you could have one, two, or even three scum in the original four. (Yes, it's possible to have all four, but not probable.) And with those name(s) you keep in the scum pile, ask who should replace the name(s) you're throwing out.

And when you've done that, run through the process again. Do the interactions make sense? If not, then how likely is it that these four are your scumteam and you're wrong about the interactions, how likely is it that you're wrong on the names but right about the interactions, and so on, and so forth. It's a cyclic process, one which should never end.

One thing all of these have in common?

The middle ground!

Don't be too extreme in any part of your play. Ideal town play is to keep that middle-ground perfectly, and to hold things mostly in balance. Yeah, there's going to be exceptions, and yeah, it's pretty much impossible to fully achieve, but you should be striving for it, that middle-ground, because it's that middle-ground which spawns great scumhunters who are incredibly rational and capable players.

"I disagree with..."

Go ahead and use what you can, even if you don't agree with the whole thing. This guide is always going to hold controversy for the exact reasons I mentioned when starting: because it's highly circumstantial, both subjective and subtle.

Because of that, the tips I've presented are largely my own opinion and my own process. Your own may be different, and that's not wrong! There's no truly entirely-right or entirely-wrong way to approach the game, this being no exception. So feel free to share your own take on things, and take from this guide whatever you find useful.