Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[notquitethere]

Mafia Theory

I'm sure we all have scumhunting theories, so let's air them. I've got a lot of broad precepts based on voting patterns but I've not managed to concretely demonstrate the truthfulness of these precepts yet. However, I've just discovered something of note.

The Player With The Highest Post count Is Never Scum

I checked the last mafia dozen games on this subforum with Zombie Urist's Lurker Tracker and I discovered that in no game did any scum-team player have the highest post count. They were occasionally the second highest poster, but never the first.

My guess is that the scum-role promotes an inherently reactionary and cautious playstyle that fosters a lower playcount. Or, inversely, the player who posts the most is the most engaged with the game, and engagement with a game more readily flows from a genuine desire to root out scum than a pretended desire.

Some players have a tendency to break their posts up into a number of shorter, frequent posts, and this can artificially inflate post numbers. Even taking this into account, this rule held for all the previous dozen mafia games on this site. This precept doesn't rule out the top poster from being a third party: this isn't wholly uncommon especially in games where there are many third-parties.

Once in BM XXXVIV, a scum player replaced another scum player and collectively they had the highest post count. However, in this case, both were IC (who are expected to post more) and all but two of the other players had replaced other, less-active players, and the highest individual poster (myself) had replaced a player who had no posts for the first four real-days of play. Even with these mitigating factors in mind, the highest individual poster was still town. This leads me to give more credence to the idea that the person most invested in the game is never scum.

[Lenglon]

That depends on several factors NQT, mostly having to do with individual players. for example, if I remember correctly from something Dariush posted in the most recent Toon Mafia, Solifuge has a pattern of posting more as scum than as town. additionally, in BM XLI the player with the highest post count of the living players on the last day was actually scum - the player with the highest post count was an IC and was lynched the day before. although your pattern seems to hold true for the game as a whole, it is not necessarilly true of the game at every single point in time. I'd be carefull about using this as a major towntell.

[notquitethere]

That depends on several factors NQT, mostly having to do with individual players. for example, if I remember correctly from something Dariush posted in the most recent Toon Mafia, Solifuge has a pattern of posting more as scum than as town.

I'm not sure I'd give much credibility to that source. Obviously, some people are going to be more engaged with a game if they're playing as scum, but even taking that into account a non-scum player had the highest post count in the last twelve games. Remarkably, this was the case even in very different mafia games, like Toon Mafia, Trust Mafia, Mafia & Masons, all the BMs...

Additionally, in BM XLI the player with the highest post count of the living players on the last day was actually scum - the player with the highest post count was an IC and was lynched the day before.

That's wholly consistent with what I've been saying. If people had known and taken into account this precept in that game then maybe Vector wouldn't have been lynched.

Although your pattern seems to hold true for the game as a whole, it is not necessarilly true of the game at every single point in time. I'd be carefull about using this as a major towntell.

I've just gone back and checked three more games. These fifteen games are all the mafia games played this year (so, all games excluding Revolution and Elements). In only one of them did a scum player have the highest post count and that was Webadict in Paranormal 22, which is a bit of an outlier because:

This game blows. Two people talked during the Day. That's a frickin' record.

So one game this year, or 1/15 had a scum player with the highest post count. Is there any town-tell more major?

[webadict]

Did you compare these scum players against games where they were not scum? Because that seems to be a vital miscalculation on your part.

[Leafsnail]

It's too easily manipulated to be reliable tell.  Also 14/15 isn't that fantastic a hitrate when you're looking for something that's true over 3/4 of the time by chance.

I post more when the day isn't going the way I want it to, and that's true for either alignment.

[notquitethere]

Did you compare these scum players against games where they were not scum? Because that seems to be a vital miscalculation on your part.

No. What I did was I ran each game from this year through the lurker tracker, sorting by post number (making sure to start each game from the official game starting post, rather than the OP). In all but one of these games, the top poster wasn't scum. This is true even for players who post loads regardless of alignment (like Tiruin). This seems to indicate that in any given game the person who posts the most is most likely not scum. Where's the vital miscalculation?

It's too easily manipulated to be reliable tell.  Also 14/15 isn't that fantastic a hitrate when you're looking for something that's true over 3/4 of the time by chance.

I post more when the day isn't going the way I want it to, and that's true for either alignment.

Uh... 93.33% is still a lot higher than 75%, and probably if I go back and take more games that will be even higher. Can you point to a more reliable town-tell? You might post more when things aren't going your way, but a seriously engaged non-scum player (and there's almost always at least one in a game) posts frequently regardless of how things are going.

[Tiruin]

So one game this year, or 1/15 had a scum player with the highest post count. Is there any town-tell more major?

Alternatively, and not to reword the past two: Did you compare the players and their respective playstyles? That is where my opinion judges the theory at hand as subjective-post/player = scum/towntell.

I mean, look at me. 

Addendum to your theory: There is a certain voice which people adapt in a certain alignment-related to their RL worldview and/or their personality. Certain wording patterns and how they are delivered contribute as tells. These are also known as 'meta-tells' where the deviation in their credibility lie in how precise the tell is.

Also I'd love to (re?)state the note on statistics. Here, they act more like a guide than a baseline on the generality in regard to a conclusion.

[Leafsnail]

Considering the sample size that isn't too large, I don't think it's huge difference.  I think there is some truth in your observation in that scum are often able to coast to victory, but turning it into a hard and fast rule means mafia members can easily get around it by just posting a lot.

[notquitethere]

Alternatively, and not to reword the past two: Did you compare the players and their respective playstyles? That is where my opinion judges the theory at hand as subjective-post/player = scum/towntell.[

I mean, look at me. 

The thing is, even when you played scum you didn't post more than the top posting town player and you usually post the most. Obviously if one were to try to make use of this precept in-game, taking into account meta-tells might be handy, but I rarely see that happen in actual play.

Addendum to your theory: There is a certain voice which people adapt in a certain alignment-related to their RL worldview and/or their personality. Certain wording patterns and how they are delivered contribute as tells. These are also known as 'meta-tells' where the deviation in their credibility lie in how precise the tell is.

Also I'd love to (re?)state the note on statistics. Here, they act more like a guide than a baseline on the generality in regard to a conclusion.

Undoubtedly there are meta-tells but from what I've seen, players are usually not very good at picking them up. I'm not sure what your last paragraph is trying to say: I think this is a more helpful tell than the usual things that get bandied around (OMGUS, WIFOM etc).

Considering the sample size that isn't too large, I don't think it's huge difference.  I think there is some truth in your observation in that scum are often able to coast to victory, but turning it into a hard and fast rule means mafia members can easily get around it by just posting a lot.

The sample size is every game this year... How many years would constitute a reasonable sample size then? I was actually a little concerned about telling you guys about this as it might hinder my ability to use it to scumhunt but the beauty of it is worse case scenario scum feel an incentive to post more content, which gives town more rope to hang them with.

[webadict]

Considering the sample size that isn't too large, I don't think it's huge difference.  I think there is some truth in your observation in that scum are often able to coast to victory, but turning it into a hard and fast rule means mafia members can easily get around it by just posting a lot.

That's not the issue. The sample size is plenty sufficient to come up with a generality. However, the problem is that the data is not parsed correctly.

And yes, knowledge of the scumtell allows for better counterplay to it. However, this is true of every single scumtell. That's why it's a tell, and not a rule. The same goes for miller claims. They are, in a large part, a towntell. No one is claiming that they are always town though. Just that they are likely town.

[Imp]

I don't even care if this theory is true or not, because I like its possible implications.

(I like it.  I think it's interesting.  I think it could be true.  I'd like it to be true.  Statistics do seem meaningful to me if prepared so they accurately answer the question being asked.  The little I know about how this statistic was prepared seems meaningful.  Small sample sizes do not necessarily mean invalid statistics, it simply increases the range of error, the spread over which the 'true number' is likely to be found.)

Now that highest post count is being talked about, it could affect players' investment in the game.  I don't see it as likely to reduce postcounts.  Currently, I support anything, true or false, that increases or even may increase players' involvement and posting in the games I play/will play/might play/may one day play.  Woo NQT!  Keep thinking and talking, I like what I know of your thoughts and words.

[ToonyMan]

Yeah, miller claims are towntells.  They don't guarantee a town player, but I always take it as a strong town read.

[Max White]

Except as soon as you say the player with the most counts isn't scum, he scum start posting the most...
There is no such thing as a hard tell. There is no 'always' or 'never'. Any discussion of possible current meta about tells will just change the current meta to be out dated.

[Jim Groovester]

Considering the sample size that isn't too large, I don't think it's huge difference.  I think there is some truth in your observation in that scum are often able to coast to victory, but turning it into a hard and fast rule means mafia members can easily get around it by just posting a lot.

But people will notice the guy posting for the sake of getting to the top. It's harder for scum to be at the top while making useful-looking posts than it is for town.

I think the metric is meaningful for that reason, but the pure post count on its own doesn't mean much.

[Tiruin]

Except as soon as you say the player with the most counts isn't scum, he scum start posting the most...
There is no such thing as a hard tell. There is no 'always' or 'never'. Any discussion of possible current meta about tells will just change the current meta to be out dated.

New theory: The most pervasive notion on being town/scum (This is why I hate labels :I) is the one to be most debated upon and inconsistent in definition.



I am half-joking.