Mafiascum.net

Well, it’s been a while, but I got inundated with school work, welcome to the second addition of TSQ talks about mafia.
The first installment, on arguing in the game of mafia can be found here:

http://mafiascum.net/forum/viewtopic.php?t=12228





The beginning of a game of mafia is in many ways difficult. Mafia is inherently a game of information. The more you have of it, the better off you are. Because of this, you want to get as much information as soon as possible. At the beginning of the game, you have no information other than your role pm, and therefore must start playing almost blind, with much less information than the scum, who have high levels of information because they know who the bad guys are. This is what gives rise to what has been popularly dubbed the “RVS, or random voting stage.” This article will be in two parts. First, I will discuss what the random voting stage IS, and what it is not.

This discussion should give us a decent starting point to discuss misconceptions about this part of the game, and also what you should be trying to accomplish during this part of the game. Second, I will discuss what I see as optimum strategies. That is to say, I will discuss what I think ought to be done to accomplish the goals set out in the first part of this article, and how practically you can accomplish this. The random voting stage need not be as random as we are making it.


It is important to understand what the random voting stage is in order to discuss it. More importantly it is important to understand what it is not.The random voting stage is defined by a low amount of town information, so that all actions taken are a little bit random due to not being as informed as they could be. It is NOT an area of the game where it is ok to BE random, and where actions and votes do not matter because they are random.

Remember, all actions in a game of mafia are structured by your role pm, and since you already have that on day one, all actions are still inherently readable. Scum can only hide in the RVS if you let them be unaccountable. In a game of mafia we vote on motivations, as well as on actions an role based information. At the beginning of the game, there is a distinct lack of the latter, and a steadily increasing, yet comparatively low amount of the first two things. To get around this, players resort to many tactics which they feel get them to information to lynches. Some of these work, some of these do not. It’s important to understand what you should be trying to do in the random voting stage: Getting the town out of the random voting stage as fast as possible. This is the towns main goal at that point. No one is going to leap up and scream “I AM SCUM!!!!!” The ones that do not have at their core some basic misperceptions about the random stage.

First, that you cannot scumhunt without information. This assertion seems on face ludicrous, but I hear it made in almost every game I am in. To understand why this is wrong, we need to have some sort of basic understanding of what scumhunting is. This article defines scumhunting as “Any action taken that makes it more likely for the town to find scum.” By this logic, asking questions is scumhunting, discussing relevant things is scum hunting, and any action you take which might give you a clearer indication of the alignment of any player is scumhunting. Given this definition, the claim that you can’t scumhunt on day one seems patently ridiculous. Furthermore, it would seem to me that you would want the times in the game where you have the least information to be the times you scumhunt the most. This is what forms the basis of fully informed lynches later on.

Second, that the random voting stage is random, so its ok for me to be random. This one is most pervasive. It is the idea that since we have little information, then the best play for the town is to just be random as well. We can see pretty clearly that this isn’t true. In order to force information into the light, the town must be as non random as possible, they must attack leads, ask questions, and create situations that bring information to light.

Ironically, this misperception of the random voting stage is what CAUSES the random voting stage. By not making players accountable for their actions, and acting in a manner that is entirely random, I have seen players make the stage of the game with decreased town motivation last almost all of day one, or well into day 2. With the right methods I have seen the town exit the time where they have almost no information as early as page 2.

So what sort of actions fall into these catagories? In the first one we have things like lurking through day one and not voting on day one. These strategies are completely incoherent in terms of playing to win. Since we need the most information on day one, when players lurk through the day, or don’t vote during that day, they are actively hurting the towns quest to find information, and thus are playing in a manner that is completely anti-town. Notice I use the phrase anti town, and not the phrase “scummy.” Because in my experience town players are just as likely to take these actions as scum players, even though they detriment the town. In order to play the RVS effectively we need to stop letting these strategies be mainstream for town as well as scum. By doing this we take away a scum crutch and force them to take positions much earlier than they would like to.

In the second one a whole myriad of things are done. The two that I find most objectionable are completely random votes, as well as self voting. When we vote with dice, or for ourselves, we are playing right into the scums hands, because we allow them to be unaccountable. They can vote completely randomly as well, and that puts us in a situation where It takes much longer to get some sort of coherent vote record. Voting with dice is probably the most socially acceptable version of this approach, and there is often little to no discussion of why these votes hurt the town, allowing the scum to hide under a shield of randomness that can be avoided.

On the subject of self voters, you should never EVER EVER vote for yourself in a game of mafia unless by some way it helps achieve your win condition. A good example of this is Mr. Flay voting himself to achieve a lynch as scum, so that he could end the days discussion and not let the town have information. As a town player, when you vote for yourself, you are voting for the only player in the town that you know to be not scum: Yourself. This means that your vote is inherently unreadable and inherently unhelpful . Some mods go so far as to ban self votes from their game, which is not ok either. You should be able to vote for whomever you want, and when town does this it is bad play, and not out of the purview of the game. We have to punish such behavior to create less places for scum to hide day one.


So we’ve established in the first part of this article what you want to do in the RVS: Create as much information as quickly as possible. How do we do this? Are forever going to be lost in a sea of no information until some scum player magically does something stupid? Fear not, there are ways out.

First, you want to make everyone accountable for everything. A good town player never accepts the excuse “it was in the RVS, so it doesn’t matter!” By making players accountable for their actions you turn everything they do into a piece of information. Do not let people get away with anti town plays just because it is the random voting stage.
A good thing to note here is that in the RVS, your threshold for what constitutes something voteworthy is lowered dramatically. The less information you have, the less scummy something has to be for you to vote for it.

Remember, bandwagons are good, pressure is good, votes are good. These are all things that create situations that are serious in game, and require reactions from not only the player who is being bandwagoned, but the other players in the game. Just because something is based on a low threshold for a vote does not mean it can’t be serious and have in game repercussions. By creating band wagons over a minorly scummy thing, you take the game out of the area of randomness, and into the realm of information. As people react to these bandwagons, as well as votes, you will get more information, which will create less random votes, and less random bandwagons. This is how we progress to better information.

Second, by asking questions. When players make actions, ask them about them, ask other players to comment on them. By asking questions that relate to the game even tangentially you are creating information about the other players in the game, even if it is theory based. These lines of questioning may not yield directly voteworthy information, but the other players in the game will comment on the answers as well as your questions, and these comments and stances DO lead to vote worthy information.

It is important to note here that not all questions are “good questions.” Fully theory based questions that do not in any way pertain to the game bog down the town into theory discussion that is not in any way role relevant, so avoid that kind of questioning. Theory questions that relate to the game are however a great way to get players to take positions which are relevant later on. For instance, the question “What do you think is optimum scum strategy in this game, and who’s play most closely models that play?” is a personal favorite of mine, because it makes a player state an early suspicion, as well as create a reference to look at their play later on.


The RVS can be frustrating, but that is not a reason to perpetuate it. Many of the actions town takes when they are in the RVS are completely incoherent given the aim of getting the town more information so that we can leave the random stage. Misconceptions about, and they allow scum to hide within the shadow of our misinformed actions. Stop playing the game as if day one has to be random, it does not. By making actions that decrease the randomness, and scumhunting on day one, you actively bring the town around to real discussion and lower the amount of areas for scum to hide.
I hope you enjoy this article as well as my previous one, and I look forward to any feedback you have to offer on it. I will hopefully continue this series, and anyone with a subject they would like me to write about is more than welcome to PM me their thoughts on that matter.

Thanks,
TheStatusQuo

Excellent article. Has made me think about the way I approach the RVS. Thanks, TSQ.

I might read both in several hours. When I'm not bitching about my face begin frozen.

Really good read. I think games would stall much less if all the players had this kind of mindset from the start. I'll certainly be taking a more active stance in not letting players get away with 'random' actions.

that is the "negative" approach. You could also take the "positive" approach and make sure information is created at the start of the game.

Great stuff. My one problem is that I expected you to discuss more of the theory behind RVS, namely that it shouldn't even exist as a seperate stage--having a term for a low-information stage in the game often makes players think that there is some magical threshold past which the RVS suddenly ends and the town is out of the woods.

I would really argue that discussing the practical side of RVS without a disclaimer along the lines of the above only perpetuates the mental barrier that the concept of RVS has become. Players should naturally view the game as a fluid progression of information without seperate stages other than the days and nights.

Korts wrote:Great stuff. My one problem is that I expected you to discuss more of the theory behind RVS, namely that it shouldn't even exist as a seperate stage--having a term for a low-information stage in the game often makes players think that there is some magical threshold past which the RVS suddenly ends and the town is out of the woods.

RVS has ended, so .

Yeah, Korts spotted the only complaint I have (well, and a few grammatical things). If we could look at it as a characteristic of the game, rather than a 'phase', we'd be a lot better off IMO.

Otherwise, nice essay TSQ.

I will preface this post by saying I've only skimmed over Shea's article. It looks pretty okay, and I'll probably respond directly to the article later. But I wanted to post some general personal thoughts, becaue I am naive and egotistical enough to think that anybody might care what I have to say.

Korts wrote:Great stuff. My one problem is that I expected you to discuss more of the theory behind RVS, namely that it shouldn't even exist as a seperate stage--having a term for a low-information stage in the game often makes players think that there is some magical threshold past which the RVS suddenly ends and the town is out of the woods.

I agree with this. There are a few things that bother me about the current way RVS is generally looked at.

1) People seem to think that there needs to be a distinct cutoff between "RVS" and "Serious."
2) People seem to think that everybody should accept that the RVS ends at the same time for all players.

And as a less-serious, more-nitpicky sidenote:
3) I generally prefer to think of page-one voting as "arbitrary," rather than "random." MoS is the only person I can think of who would consistently vote. (That is, he would always use dice tags to determine where his vote would land on Page 1.)

The second one bothers me especially. Back when I was but an IN DEBUT I SPEW, the "random voting stage" was nearly seamless. People would start off jokingly voting/wagoning each other, and then someone would vote for another for a non-arbitrary reason. People would examine that reason, agree/disagree, and wham! Suddenly we have people taking sides and interacting, and the game progresses as usual.

Eventually people began to notice "hey, these Page 1/Page 2 votes are for pretty silly reasons, and eventually people get serious," and so they labeled that early phase. Through evolution of game, playerbase, and meta, we arrived where we are today... that people seem to think that there is like a set real-life time period or number of pages/posts where random play is "acceptable," and everything after that point is :badposting: (or, worse, automatically "scummy").

In one of the games I played recently (can't remember which, too lazy to post), I actually saw someone attack another player for making a random vote "after the random stage had ended." I practically did a spit-take, because I couldn't figure out why the attacking player had decided that the RVS was universally over. Some players had found non-arbitrary reasons for votes, but not every player had posted, and not everyone was taking the game seriously. I certainly hadn't been up to the point where I read the attacking post. But it got me thinking "what makes one player decide that the random stage is or 'should be' over for everyone else?"

There are two other reasons that I like the Arbitrary Voting Stage (suck it, RVSers!).

First, it gives me information that I probably wouldn't otherwise get. If there's one concept which "works" for me when catching scum that I wish every other Scummer would follow, it's this: Around D3/D4, re-read D1 completely and in great detail. Shea makes a point about being better off with more information, and this could not hold true more than when examining D1 play. Because one thing that I don't think most players fully grasp is that . Scum A will treat Scum B diffrently if B has heavy suspicion on D1 than he will if B has heavy suspicion on D5. I don't think it's ever a conscious thing, but it's almost as though scums assume that they won't be caught by interaction with a scumbuddy (whether bus, defense, counterwagon, or whatever), because they assume that the town doesn't have the knowledge to root out these interactions. On D1, that's completely true. But by mid-game or endgame, when you have more information, these interactions become much easier to point out.

The second reason I like letting early-game play out is because it helps a wide host of metas. Some people are silly to different degrees and for different lengths depending on their alignment. Like with scum interactions, a ton of people don't even realize that they behave differently in early-game. While I understand that general mafia meta and individual player metas change over time, it is inherently impossible to play the exact same way with fewer unknowns and a different set of win conditons. And in this light, the RVS is a wealth of information waiting to be exploited by a keen eye and a sharp mind.

So how does this relate to Shea's musings on the RVS? Well, I think we can safely agree that any protown player should want to utilize the early-game, low-information stage as much as possible. Kudos there. We want information, and we want to be able to utilize that information. However, I think there's one subtlety with which I disagree.

I think it's okay for people to be random in the random phase, but I think that too many people want, try, or happen to be too random for too long, whether this is a conscious effort or not. When you are protown, think to yourself "what kind of posting will help me and/or the town find scum?" While it may be amusing to you personally to banter with X for a page and a half about changing your gender, chances are this is going to clutter the thread for most players and create a lot of white noise. Scums white noise. Love love love love. It provides them with a whole host of excuses to attack people, to lurk (actively or inactively), and to be largely disinterested with "finding" scum. Or it can give them incentive to "look the part" of the person trying REALLY REALLY HARD to find the scums in spite of your own most random efforts. I know that we all love a little praise, and that each protown player wants to find and catch scums themselves. Believe me; I understand. But the players aren't just the ones who can find scums themselves. They're the ones who enable their entire townmates to find scum, too. They're the ones who, just by being in the batting order or by playing on the court, make their teammates hit and score and hustle that much more.

So what does this mean? How do you lend your teammates a hand? Two things. One, be the instigator. Be me in Sesame Street Mafia and take a stand so strongly that people have no choice but to react and interact. Two, make people take sides on whether X is scum or whether Y's vote is good. Shea hit the nail squarely on the head when he said "make everyone accountable" and "ask questions." Suppose I bandwagon a player. Ask me why I did so, and ask two or three other people what they think about my vote, and about the person I bandwagoned. This is the type of behavior which not only makes the RVS useful to the town, but also leads us out of it in a natural and productive way.


tl;dr version
Some randomness is fine, and may even be necessary in any given mafia game. But don't let random play breed random play. Your goal as a protown player should be to make the "random" phase useful to the town not by eliminating it, but by using it to force players to interact with one another. Take it upon yourself to bring the town to the point where they can use gameplay to come to a productive lynch. If that means sticking your neck out a little bit, do so. If it means asking the hard questions, ask away. But this notion of "D1 is so random, we might as well be silly" leads to above-average mislynches, makes re-reads that much more painful, and sets a tone which strongly favors the scums.

EBWOP: Sorry, that turned out much longer than expected. Even my tl;dr is lengthy. :/

First, it gives me information that I probably wouldn't otherwise get. If there's one concept which "works" for me when catching scum that I wish every other Scummer would follow, it's this: Around D3/D4, re-read D1 completely and in great detail. Shea makes a point about being better off with more information, and this could not hold true more than when examining D1 play. Because one thing that I don't think most players fully grasp is that scum give off more about themselves when there is less information to go around. Scum A will treat Scum B diffrently if B has heavy suspicion on D1 than he will if B has heavy suspicion on D5. I don't think it's ever a conscious thing, but it's almost as though scums assume that they won't be caught by interaction with a scumbuddy (whether bus, defense, counterwagon, or whatever), because they assume that the town doesn't have the knowledge to root out these interactions. On D1, that's completely true. But by mid-game or endgame, when you have more information, these interactions become much easier to point out.

I think this is an important point. It is my firm belief that the most difficult part of scumhunting is catching the initial scum, precisely because you have no reliable associative tells to go on. Similarly, it can be very hard to catch a lone scum. Associative tells are often by far the strongest tells; catching x by how he interacts with his scumpartners as opposed to how he reacts to townies is pretty much expected.

As such, day 1 (and to a lesser extent day 2) are the engine room of the town's scumhunting effort. Day 1 is the day when everything kicks off, when the players who are viewed as "scummy" are split off from those viewed as "town", where most players initial reads are determined, focused, etc. Everyone is alive, so everyone has the possibility of interacting with anyone, and who and how they choose to interact with those players is often crucial.

By later days, it's very rare that new information is being produced; most of what is being considered is based on the previous days. So getting day 1 right, avoiding it being a mess of noise and garbage, forcing players to take stances and put down their positions, and making sure it doesn't last so long that rereading it is an unpleasant task is absolutely crucial to the town's scumhunting efforts. How day 1 goes often influences how the rest of the game progresses to an impressive degree; not in necessarily the outcome of the lynch but in the info gathered.

Anything anti-RQS is good by me.

I like RVS, leave it alone >.>

Haylen wrote:I like RVS, leave it alone >.>

Anyone who does not agree should be hung to the nearest tree.

yabbaguy wrote:Anything anti-RQS is good by me.

Haylen wrote:I like RVS, leave it alone >.>

CSL wrote:

Haylen wrote:I like RVS, leave it alone >.>

Anyone who does not agree should be hung to the nearest tree.

It'd be nice if you guys deigned to weigh in on the debate rather than just flat out saying "NO" or "YES" with no more justification than that.

Split the first post up into easy to read bullet points, and I will

Haylen wrote:Split the first post up into easy to read bullet points, and I will

Don't post in MD if you're not prepared to do the intellectual work.

Sorry, I didn't realize there was a rule against quickly replying when you wanted to show initial support for the article after a skim, with the intention of following up with a better response at some point later in time. That said...

D1- the first wagon has more often than not started on the guy who just behaves the most screwy of the bunch (VI, someone having a bad day, socially awkward, whatever). Who that happens to be at that moment in time is random. Players' reactions to the first wagon (who's complacent, who's complaining, who's computing...) are what make the game, so that's why this definitely makes sense.

I would argue that by revealing actual information (a claimed investigation result, a flip), the prior existing information gets a new perspective on it. I would disagree w/ Shanba therefore that D2+ is lacking in the "new information" department because by being forced to look at past events, wagons, arguments in a different perspective.

The other thing is that people's reactions to these claimed roles and results can be telling. People who are oblivious to proper pro-town strategy are probably not town, I would claim. To those who've done Thread Reading Club and seen the Friends/Enemies assignment we see Either way you look at it, both deserved the lynch, since they were clearly not an asset to town.

Shanba wrote:I think this is an important point. It is my firm belief that the most difficult part of scumhunting is catching the initial scum, precisely because you have no reliable associative tells to go on. Similarly, it can be very hard to catch a lone scum. Associative tells are often by far the strongest tells; catching x by how he interacts with his scumpartners as opposed to how he reacts to townies is pretty much expected.

I fully agree with this. A lot of scum-side awkwardness comes from having to deal with other people. With a lone SK or no flipped scum, it's more difficult to see that awkward behavior (pairing theories between two living players almost always end in tears).

I also agree with how RVS silliness is fun but really counterproductive. Making silly votes to be silly just clouds an already cloudy stage of the game. Then again, it is a decent scumtell...

I disagree with Glork about how artificial it is to say there's a "clear division" between RVS and serious business. If you random vote after nonrandom discussion has already started, you're effectively ignoring it... why?

Probably my biggest gripe about the early game is how people say "oh you, it's only the RVS, don't get worked up". The game doesn't get started if someone doesn't take a stand. I would prefer if it didn't take something this to get discussion going <.< (sadly, this worked very well if you read on)

Random Anti Town Stage fails miserably, for sure. You're trying to help the town by being... anti-town. (Oh, RATS.)

Vi wrote:I disagree with Glork about how artificial it is to say there's a "clear division" between RVS and serious business. If you random vote after nonrandom discussion has already started, you're effectively ignoring it... why?

Well, now that's the million dollar question. The answer to said question will be different in literally . But to mandate that once a non-random choice has been named, the RVS "should" be over for everyone. Allowing a degree of individuality and interpretation is actually a great way to find scum. You let people be themselves, and that fosters healthy player interaction.

I liked the article, Shea. A little heavy on the rhetoric, but not so bad that I started skimming.

I've still never random voted. My MO from my very first game has been to introduce myself, exploit theme mechanics, and talk about meta. When I started here, this approach brought me a lot of heat but now it's more accepted. I'm not sure if that's a change in the site meta or if people have gotten to know my playstyle but it is refreshing. There are some other good players who have been taking strides over the last year or so to turn the RVS into an information creation phase, and I'm glad for it.

I pretty much agree with whats been said.

RVS has become random though. I have read far too many games where a player gets asked about their vote and their defense is that it was random so doesn't need reasoning.

The point behind RVS is to create information, when players go against that a part of me dies.

I agree with Ythill lots, when I joined the RVS was not looked into, the trend went to a point where you could analyse them but now it's going back down hill. So it'll probably continue going between the two as waves of players join and question the last generation's methods.

A much better way to generate large amounts of information early on is to just massclaim...

RVS has become too random though. I have read far too many games where a player gets asked about their vote and their defense is that it was random so doesn't need reasoning.

...which is scummy. I'd be up for serious voting that person.

Pie, I presume several others have already squawked at you about this, do you still deny that the information for scum is greater? I hate individual players who powerclaim D1, it's reckless, and it's more often than not unnecessary.