Mastin Academy

This forum is for discussion related to the game.
User avatar
GuyInFreezer
GuyInFreezer
Magical Girl
User avatar
User avatar
GuyInFreezer
Magical Girl
Magical Girl
Posts: 18048
Joined: January 23, 2013
Location: In your wall

Post Post #325 (ISO) » Wed Jun 28, 2017 2:18 pm

Post by GuyInFreezer »

Also, funnily, the counter for "WIFOM for everything" argument is really "risk vs reward, learn to think you noob."
Show
"I used to think you had this elegant-trolly, minimalist playstyle. Then I realized the playstyle is ~Lazy~
The true enlightenment was realizing that they are the same thing."
~fferyllt

"who the fuck fakeclaims Tracker like that
WHO THE FUCK DOES THAT"
~Alisae
User avatar
mastina
mastina
She/Her
False Prophet
User avatar
User avatar
mastina
She/Her
False Prophet
False Prophet
Posts: 16670
Joined: October 7, 2016
Pronoun: She/Her
Location: Between Snohomish and Monroe, WA
Contact:

Post Post #326 (ISO) » Wed Jun 28, 2017 4:39 pm

Post by mastina »

In post 325, GuyInFreezer wrote:Also, funnily, the counter for "WIFOM for everything" argument is really "risk vs reward, learn to think you noob."
The two articles are closely linked, yes. :P
Geralt of Rivia
Geralt of Rivia
Goon
Geralt of Rivia
Goon
Goon
Posts: 235
Joined: May 20, 2017

Post Post #327 (ISO) » Wed Jun 28, 2017 4:52 pm

Post by Geralt of Rivia »

I'm interested on this, to start off with my first and now completed newbie game on this site: viewtopic.php?f=11&t=71951

What could I have done better?
What should I not have done?
Also, any helpful comments are appreciated.

I'm looking for advice on improving my game play in general for starters and one other point which I will disclose via PM once I see a reply to this post.
User avatar
mastina
mastina
She/Her
False Prophet
User avatar
User avatar
mastina
She/Her
False Prophet
False Prophet
Posts: 16670
Joined: October 7, 2016
Pronoun: She/Her
Location: Between Snohomish and Monroe, WA
Contact:

Post Post #328 (ISO) » Wed Jul 19, 2017 2:02 pm

Post by mastina »

I may need to reword this article some time to better explain this concept, but for now...
Resolving Simple versus Human

So in the game of mafia, there exists a fundamental argument which is a guiding principle of the game: keep it simple. Occam's razor dictates that the simplest explanation is often the correct one.

...And yet. This runs into a small problem. Humans are complex; we
aren't
simple. We have all of these things which complicate the equations we would run: we don't all think the same way; we don't all express our thoughts the same way; we don't all experience the same emotions; we don't all have the same stimulus for emotions.

What we do is influenced both by thought and by emotion, by feeling, and yet these are things which cannot be universally applied, boiled down to what's the simplest answer for everything.

All the same, when it comes to scumhunting, you can often still differentiate between the two. When a person's actions are simplistic, they are more likely to be scum; when a person's narrative is simplistic, it is more likely to be accurate.
These are not a contradiction
. Having a simplistic narrative is following Occam's Razor.

The former, however, takes a little more explaining. Essentially, it boils down to humans being difficult to pin down. We're all fucked up individuals. As a result, it's difficult to pin down what we tend to think, right?

...Kind-of. In my experience, a scum role PM tends to make a person be incredibly objective-oriented. They know their wincon. They are the informed minority. And so, they have a bias in their actions. The job of a scum player is to do actions which help the scum while looking like they are helping the town. So a scum player's interests are rather simple. They are interested in a specific way of portraying things. Their posting is often more simple, because their goal is simpler: look town, while not helping town.

In contrast, a town role PM actually tends to keep people closer to their default human status: they lack
direction
. A town player, as the uninformed majority, doesn't have an idea what they are doing. They are inherently random and chaotic. As a result, they are more complex and harder to figure out. This is why it's so easy to misread players:
scum are better at giving a good image than town are
. They will have posting which there's only a few select interpretations of. With town, a town player doesn't have that same interest in looking town, so they don't have a focus on their image. Without image control, it can be hard to understand what they are doing and why it's town. (Often, even they don't know!)

Town players tend to be more human, more complex, whereas scum players tend to be more simple in their posting. And yet, their narratives are the inverse. A scum player's narrative requires an extra step to make sense; a town player's narrative is what makes the most sense.
Because scum are aiming for a goal opposite of town, their disguise to look town requires more explanation
. This extra explanation, this complexity in the narrative, is where they begin to violate Occam's Razor, making them stand out more.

In summary:
A town player tends to act in complicated ways which are more complex, but without an agenda which makes it easier to read their narrative; a scum player tends to act in ways which are simpler, but with an agenda which requires a more complicated narrative. Confusing the two is disastrous and will lead to many town losses, as they are distinctly different things.
User avatar
mastina
mastina
She/Her
False Prophet
User avatar
User avatar
mastina
She/Her
False Prophet
False Prophet
Posts: 16670
Joined: October 7, 2016
Pronoun: She/Her
Location: Between Snohomish and Monroe, WA
Contact:

Post Post #329 (ISO) » Sun Nov 26, 2017 3:30 pm

Post by mastina »

So something I've wanted to do for a while is a glossary for me: a mastina dictionary of my terms, where I define and explain them in detail. I figured I might as well post them here.

Playerlist Synergy
:

The ability of a playerlist in a given game to synchronize. How well they mesh together. Usually, when I use this term, I am referring more specifically to how they do so
with me
, and vice-versa, in that I am going to make it well-known how well I can work with the players in the game (or more likely, complain about the inverse).

This is how well different players can work together. Too much of the same style is bad, as is a conflict between two entirely alien styles of posting. The best level of playerlist synergy offers a mixture of the familiar and the unknown, of people you know well and call friends, people you know but don't really interact with much, and people who you don't know. (It is best to avoid having people who you strongly disagree with most often.) It is closely tied to situational awareness, specifically as town.

Gamestate
:

The state of the game. It is directly tied to situational awareness, in that situational awareness is essentially knowing what the gamestate is. A gamestate is how the thread is progressing. It is the way the game is moving, for better or for worse. It includes type of content and also direction of momentum. Mood is both influenced and a driving force behind gamestate, in that the emotional state of players will determine how things go, and yet how things have gone will influence the emotions of players. Playerlist synergy (or lack thereof) has the strongest influence over it.

Situational Awareness
:

Simply put, it is self-descriptive: the ability to have a level of self-awareness about your situation in any given game. As town, this manifests as knowing the playerlist synergy in your game (or lack thereof) and having a good feel for what your role in the game will be/needs to be: what role you need to fill, what you need to do. This is from a combination of your role PM, your confidence level in your reads (or lack thereof), and your interactions with your fellow players.

As scum, this mainly manifests in psychologically profiling players and creating plans: knowing what the likely obstacles you are to encounter to endgaming town, and producing the best path to overcome said hindrances as quickly as possible. You get a good idea of how players are going to interact with each other, you get a good grip on how your team will play with one another, and then you form a strategy wherein you manipulate circumstances such that events always favor your faction. It is THE most important skill to have as scum, hands-down, because if you lack the awareness to know where you stand in a game, then you fundamentally cannot make decisions be it day or night which move the game in the direction you want.

Psychological Profile
:

More or less exactly what it sounds like: profiling your playerlist. It is a compilation of their meta, which provides you with a predictive algorithm of how they are likely to act in any given game as the given alignment (usually town, because usually you use psychological profiling as scum). This is how you guess how players will act, and interact, with one another. You take everything you know about a player, and then given the introduction of elements you control, you guess what will happen to them.

This skill isn't useless as town, but it is mostly a tool for scum. It is also exactly that--a tool. It is not infallible because no matter how good your profile of a player is, you cannot perfectly predict their every action. It can and will be wrong. Mostly, you use it to form an initial plan, and then change the plan based off of new data. Profiles, when the skill is mastered, can be adapted on the fly, where when that new data comes in the profile is updated.

As town, I do tend to have fun with it. A large part of my RVS voting relies on psychological profiling: I will look at posts in the RVS (even if just confirms), and then off of all the data I have, I run a guess on whether the player in question would be town or scum, combining this with situational awareness. However, I am of course quite aware that my accuracy rating there is low; it is mostly a way to kickstart serious game discussion.

Town Triple-Threat
:

This is a term I use near-exclusively as scum, however, I will make occasional references to it as town. (Largely when describing what I would think of a player as scum, but even outside of that it has an occasional use.) The town triple threat is exactly what the name implies: the stronger these three aspects are (especially in conjunction with one another), the more threatening that town player is and the higher (if nothing else on policy alone) that player will be on the kill list.

Obvtown:
Exactly what it sounds like, when a player is obviously town. Of the three points in the triangle, this is actually the one I fear the most as scum. A player can have shitty reads and utterly lack charisma, but if that player is impossible to lynch because they are so blindingly obviously town to literally every town player, then scum are in danger of losing via POE. ONE player being obvtown, on their own, is actually not much of a problem (unless said player holds a key PR and is thus an unlynchable PR). But when you leave that one obvtown player alive, and then there's another obvtown player created, and you leave that obvtown player alive to have two obvtown players alive, and then a third is created...

...It doesn't take long before when crossreferencing roles with obvtown, the scum start to get cornered. So while one obvtown player might not be threatening (thus making it the weakest of the three), get to a certain point and it becomes problem number one. Yes, with that many unconfirmed players considered obvtown, doubt does creep up. They are, after all, unconfirmed. However, if scum failed to successfully weave their way into the inevitably-resulting townbloc from said obvtown, then the townbloc will lynch those outside before they start to dissolve it.

Being obvtown is also a necessity for those who have the other parts--it doesn't matter if you have the charisma and accuracy to lynch scum. If the town turns on you and thinks you were scum bussing, you get mislynched out of paranoia and the final scum who you had pinned gets to walk out free. It is also not a one-off thing, usually. To ensure success in a game, a town player needs to CONTINUOUSLY be obvtown, especially in a day and age where town players have the memory of a goldfish (figuratively speaking that is). They won't remember that amazingly awesome towntell you did three weeks ago. They'll remember that towntell you did yesterday...until a week or two later that is.

Charismatic:
Next down on the list and closely tied to being obvtown is being able to convince players. A player who is able to get others to follow them is an absolutely
terrifying
force to be up against. And, sure. Say that player is off in their reads NOW. What happens when their inaccurate reads with that charisma suddenly turn into ACCURATE reads with that charisma? If they're still in the game, and if the town still trusts them enough to not lynch them, then you can kiss your scumteam goodbye.

Of course, it's possible to lynch charismatic players if they are not obvtown, because getting people to follow you doesn't mean they'll think you're town and you can see a battle between two charismatic players which means ONE of them
has
to lose (unless there's a read reversal that is), but in general, a town charismatic player is the greatest enemy of a scumteam. Charismatic players often have the illusion of appearing as tunnelers with confirmation bias, but frequently they possess the ability to reassess at the most inopportune of times. They are usually smart enough for instance to piece together that "I'm charismatic-->It's lylo-->I'm not dead-->I'm alive for a reason-->My reads have been wrong", a chain of thought you absolutely cannot afford.

However, that being said, accuracy is important because if a charismatic player IS wrong, and doesn't reassess, they cost the town a game which was easily winnable with just a little more skill involved in scumhunting.

Accuracy:
As scum, I actually don't fear accuracy that much when it comes to players, because accurate reads mean nothing if you can't convince others they're accurate, and/or you get mislynched before you can push them through. HOWEVER, when in tandem with other factors, accuracy can be a huge hindrance.

An accurate scumhunter with a vigilante role is going to trash your scumteam regardless of their charisma. An accurate scumhunter WITH charisma is going to lead lynches on scum, and good luck figuring out how to deal with that down the road. (It's possible, but incredibly difficult.) An accurate scumhunter who is obviously town is someone you can't get the town to discard the reads of by mislynching. (Because towns will in fact discard the reads of mislynches no matter what.) So in of itself, accuracy may not be a threat, but when paired with other factors, it amplifies the threat level of a town player considerably.

Notably absent: Roles. I do not take roles into consideration when creating threat values for players. Roles are more of a tiebreaker. When all other factors are equal between players, the greatest role threat is never a bad idea to take out. Roles CAN contribute to a threat--players who are conftown are more of a pain; an obvtown player with a role you don't want obvtown is a huge problem; accurate player lands vig; etc.--but for the most part, roles don't influence my evaluation of a player.

Keep in mind that the town triple threat also is variable by game. Players have a "set value", so to speak, in all three fields...
however
, those values increase or decrease on the merits of the individual game. A player who is borderline-lynchbait in one game might be the closest thing to conftown in another. A player who is a scumhunting god in one game might be a village idiot reads-wise in another. You can average these, but only after observation is it possible to generate accurate data for the CURRENT game.

Three Axes of Reads
:

This is a relatively new terminology I've started to use, even though the methodology behind this has been refined for over nine years. The three axes of reads are exclusively something I use as town, and are how I scumhunt these days. This is, in essence--the three driving factors which are how I generate my reads when I play a game.

Play-based Merits:
This is, essentially, reading a player off of their play in a game, evaluating off of the available information I have and referencing it against my mental playbook of town versus scum play. That will take some explaining--basically, I have mentally compiled and archived a set of "tells"...but not in the sense people think of the word. When people think "tell", they think 'scum are more likely to do this' or 'town are more likely to do this'. That methodology is flawed and doesn't work. Mine's different. The simplest way of saying it would be, I suppose...
"In the current situation, given the gamestate, a town player of this sort would be more likely to do this, whereas a scum player of this sort would be more likely to do that".

There are a couple of key differences. Most prominently: my 'tells' are
circumstantial
, NOT universal. They apply only in specific situations. They rely on context. They cannot be boilerplate applied across any/all games. They depend specifically on the individual situation. "Given the play I have witnessed thusfar, scum are more likely to have this as their aim, where town are more likely to have that as their aim". My internalized tells do rely on a few assumptions and make a series of links, chaining things together which can (and often will) be a mistake.

That being, meta and interactions tie into these tells. I will see how a player interacts and off of the data I have stored in my mind, I file the play as increasing or decreasing certain scenario's likelihoods. These play-based merits typically don't adjust for meta, especially not when I feel I lack familiarity with the player(s) in question. A major flaw in my system is when I lack data to the contrary, I often assume others will have a similar subconscious tendency to act in a way I would. That is, my tells for situations operate largely on what I assume to be the optimal play given what I know.

If what I know is massively wrong, and/or the scum simply aren't operating on a wavelength I'd consider even remotely optimal, it makes them much harder for me to pin down and identify.

OF NOTE: My VCA methodology? It largely works off of play-based merits. In fact, my VCA can be thought of as play-based merits from the summary of votes alone to tell the story. I predict the patterns of what I think scum are likely to have done
in the current situation
, and from there map out the people with the highest probabilities of being scum.

Play-based merits also get stronger the later into the game I can survive. They work off of information. They work off of gathering what has happened, and from what has happened, predicting what was most likely the cause of what happened. The more data points I have on hand, the better. Up to a point, of course. The human brain can only handle so much information at once and while I do my best, my mental archive for any given game is only so big.

Meta-based Merits:
This is mostly what happens when I take what I know about a player and crossreference it with their play. It changes the equation, so to speak. There's a big difference between Ginngie voting me and Nachomamma8 voting me, even though by play-based merits they would most likely in the given scenario have incredibly similar game profiles. It is neither superior nor inferior to play-based merits, in that it is an extra dimension to a read, rather than a replacement. I use play-based merits by default since I don't have meta-based merits on everyone, but when I use meta-based merits they do not supplant the play-based merits.

Meta-based merits are simply an adjustment to the equation. It's quite fully possible and rather common for me to have a null read off of play-based merits. Having a meta-based merit can break the stalemate and tip the read into a particular direction. Meta-based merits MAINLY see use, however, in changing the chances behind actions: "Given what I know about playername, this circumstantial tell which would normally tell me that is less likely to apply", "given what I know about playername, this circumstantial tell informing me of that happening is far more likely to be the case", and the like.

Meta-based merits aren't, therefore, strictly what you would think of as meta, per se. They are more a modification of a model, with the model in question being the read of the player off of their play. That modification, just like the play-based merit, will come with a strength rating. That being, some metas are better than others and I try at all times to have an awareness of how good or not-so-good the meta is.

Crowd-based Merits:
You might think an appeal to majority is a fallacy...but in actuality, outsourcing my read to others is often the secret to having a semblance of success in games. I am well aware of the weaknesses in my play-based merits and my meta-based merits. So, often, when I have doubt in my read, instead of trusting my read, I will...trust the read of some other player in the game. This other player--often it's actually more than one player--that I choose to trust isn't chosen blindly. The trust is earned, off of my reading of them by both play- and meta-based merits.

That being, "Given this situation, in these specific circumstances, I feel that playername(s) have a better read on *player(s)* than I do". It is me placing a fundamental level of trust in the competency of (an)other(s), and often not unjustifiably so. Thread consensus reads might not be accurate always...but that's why once more, this is not a replacement for the other two. It is an augmentation, supplementing existing analysis. I will use my best judgement on whether to trust the thread's/the player's judgement.

This is also somewhat where I place faith in dead players. If we're later in the game, I will reanalyze those that have fallen, and look for clues that might have led to their demise. Nightkill analysis is a hard tool to use well, because there are simply too many possible factors determining the scum kill and even if you correctly identify why they died, you then have to convince the town that you have actually done so. (And they often won't listen, if for no other reason than "but...WIFOM!".)

Basically, each of the three is a core part of the process. When one fails, I rely more heavily on the other two. I have a base model. I modify that model first off of the circumstances behind play. I then modify the model further off of my knowledge about players in question. Then, I make final modifications off of what I see from others and the level of trust I place in them. I've had people be much higher/lower in my readslists than they otherwise would be solely off of seeing what another has said and silently marking it (without commenting) as something I felt was potentially important.

Soulread
:

For me, a soulread isn't simply "I can read this player 100% of the time". For me, soulread instead means something akin to "I know EXACTLY what drives this player". This is knowing how they think, such that when they perform an action you are able to identify the why with a 100% accuracy rating. And that applies to
every
aspect of their process. It's not just 'they are scum this game' or 'they are town this game'. It's "they are doing this thing as scum, for that reason, following this process". It's "this thing they are doing is driven by that factor. This thing makes them think that, and do this. Which leads to eventually, that. And from that, we can conclude this".

A soulread is
knowing
a player. A soulread is
getting
them, on a fundamental level. It's understanding what drives them, what makes them, what they do, how they think. If you can't predict the process behind a player's actions perfectly, you didn't soulread them. Having an ability to identify and hone in on exactly what makes them do something and what that means for their alignment is a soulread; having only the alignment is not. Because a soulread is reading their soul. A soulread is the ability to see their actions and to see them unfold every step of the way. It is seeing their plans, their motivation, their mindset, their intentions, and how those shift off of factors in play.

Thinking In Concepts
:

This is a phrase I've taken to using which refers to how I process information. I am autistic. When I think, I do not process information in English. I process information as ideas. Ideas come in, and out. And when they come out, I have to attempt to translate them. This has led me to the saying that English isn't my first language; concepts are. And it means that instead of thinking "green", I will have the idea of a color in my mind, analyze it, then have a need to find the word which most closely matches the description, and then that word becomes what I thought, in this case, 'green'. But it could also have been "forest green".

Or when thinking of 'light blue'. I first flashed to an image. It was only then that I diverged in paths: 'light blue', or 'cyan'? They were one and the same in my mind as a concept, but then I had to unpack the concept, and when it diverged, there were two possibilities, only one of which I could select and yet I have no clue which would be the "right" one...possibly, neither.

It is me thinking in the abstract, and then trying to translate what that thought was into something more concrete. When I think "I am speaking", I don't think "I am speaking". Instead I had the idea of something involving my throat, which I then had to translate into an action (speaking), and then from that action make words. In that case, English isn't the second language; it's the
third
.

This process has a conversion algorithm of sorts. I take information in, and I spit information out, as efficiently as I can. Translating what words mean into a format my brain can functionally recognize them, and translating the concepts into a form where they can be given for others rather than just myself to understand. However, while this conversion algorithm is
good
, a fair amount of information is lost in translation, quite literally. Ideas in my head get lost when I try to put them to words. Concepts which so clearly shone in my head suddenly stop, cease to exist, with no explanation for what they once were. I see "empty data", and recognize the file was corrupted, but cannot run an effective recovery process most of the time.

Scenario(s)
:

A scenario is, essentially, my autistic brain running through a given situation, usually hypothetical. These are the things which I am determining the possibility versus the probability of. They are largely on the conceptual level, in that I will run through dozens of scenarios in seconds flat, yet in the time it takes me to run through those scenarios mentally, I can maybe verbalize a single one of them in spite of having done them all. Literally hundreds to thousands (if not millions) of scenarios exist every game, and are rapidly revised, removed, and invented as the game continues.

Every time I make a post, I am running through the scenarios: how people have acted and what that means, and predicting how they
will
act off of my new content. Scenarios are essentially me mentally acting out events in the game in my mind. They are never accurate 100%, nor could they ever be. They use everything I know, to make models off of what I think went on and will go on.

Possibility
:
A scenario which is possible, but not very likely to be true. It can be shorthand for a few things. It can mean "I am considering this, and quite seriously". It can mean "I think pushing the idea of this as true is likely to come from scum". But at its base level, it simply means that something isn't impossible, but yet also isn't something I feel to strongly be the case.

Plausibility
:
A scenario which has a stronger chance to be true, but is not strongly likely to be true. It is between possibility and probability in terms of scenario strength. I don't use it very often, but when I do it is often to help reaffirm those that know of what I mean by 'possibility' that I am not dismissing them even remotely. So instead of 'seriously considering', it would be "that's actually something which resonates, but I have some doubts keeping me from fully supporting you".

Probability
:
A scenario which isn't confirmed to be true, but is more likely than all or almost all other possible scenarios. It is something strongly likely to be true, to the point where using it as a working theory for what happened would not be a terrible mistake. It can still have doubts, but if so, they are at minimal levels.

See Also:
Possibilities Versus Probabilities.

Play > Role
:

A self-descriptive driving philosophy of mine. Roles are meant to augment play, not the other way around. If your read on a player is based off of a roleclaim--especially in an era where moderators give scumteams a plethora of fakeclaims/safeclaims of really cool/strong/useful roles--then you are likely to lose. Roles CAN influence a read...by paying attention to what that player has done with the claimed role. If their play matches, and the play for that role appears to be helping the town, then it is probably a safe assumption the role actually is in fact town.

If there's ANY sign of the play not matching the role as they should be playing it as town...well then you're a moron if you overlook it. How a player plays is, barring conftown, a far more reliable overall indicator of alignment. Role != alignment is the accompanying statement behind this philosophy.

Role != Alignment
:

A saying directly tied to Play > Role. Whereas Play > Role is the driving philosophy behind my beliefs, this is the driving statement therein. A roleclaim does not guarantee the player is town, no matter the role (barring extenuating circumstances where, yes, it is a 100% town role. There's no scum Innocent Child; there's no scum Friendly Neighbor; there's no scum mason; there CAN be a scum sane cop). It is a statement both because the roleclaim could be a fakeclaim (and thus, bullshit) AND even if real, it is not the role but how it is used which demonstrates alignment.



This is by far not a complete list. It's majorly a work in progress, too. I feel many of these descriptors are inadequate. But it's at least a START.

Notably absent: Gut. That, even when I get the glossary to standards I consider acceptable, will be reserved for a separate far more detailed post.
Also, I did not talk about Meta/Interactions/Mindset nearly as much as I would prefer, but oh well. I can get them later.
Post Reply

Return to “Mafia Discussion”