|
On March 17 2015 23:14 plotspot wrote:
Thanks for the answers so far. I have a few other things I was wondering about like:
Is a minimalistic game possible, where everyone is silent except for daily activities of voting and nightly activities of dispatching and other non-verbal functions?
Assuming a yes, would such a game not favor the mafia, as no information is baited out from them through any sort of interaction whatsoever?
I hope I can understand you correctly.
Rules can be made such by the moderator in some of the veteran normals, but it's not the norm in most games at all, and silent nights (where posting is disallowed but night actions occur) never happen in these student games. But yeh I'd say they favour scumteam, and thankfully are rare.
Also I think making goals for each game is a good idea Tronak, it's something I still do for myself. Also style of posting will change from student games to normal minis, so I think adaptability is key. But either way, right mindset in trying to make yourself more readable.
|
I think you understood me partly. But I was not talking about rules to forbid posting. I meant a minimalistic game out of free will, where no one chooses to talk at any time. The game with the smallest amount of posts possible. What would it look like? I just want to grasp the lowest possible boundary for this game, regardless of whether it makes sense or is the norm in the first place.
|
On March 18 2015 00:14 plotspot wrote:
I think you understood me partly. But I was not talking about rules to forbid posting. I meant a minimalistic game out of free will, where no one chooses to talk at any time. The game with the smallest amount of posts possible. What would it look like? I just want to grasp the lowest possible boundary for this game, regardless of whether it makes sense or is the norm in the first place.
Of course it is possible but that would make it a "guess right or wrong" game instead of mafia game.
As the game's goal is to talk and figure out who is talking out of their ass instead of genuinely figuring out who is mafia and who is not, it makes the chance of "minimalistic game" pretty much 0%.
To properly answer your question, it would look like "randomly vote for someone" game. In other words not a mafia game.
|
On March 18 2015 00:14 plotspot wrote:
I think you understood me partly. But I was not talking about rules to forbid posting. I meant a minimalistic game out of free will, where no one chooses to talk at any time. The game with the smallest amount of posts possible. What would it look like? I just want to grasp the lowest possible boundary for this game, regardless of whether it makes sense or is the norm in the first place.
Let's create a basic game based on the Student Mafia setup...
Setup:
10 Town
3 Mafia
(0 Power Roles and/or 3rd parties)
Rules:
Town randomly lynches one person a day.
Mafia kills one town every night.
Mafia's wins if [# Mafia] >= [# Town]
Town wins if [# Mafia] = 0
I'm too far removed from my collegiate statistics class to come up with a "neat" solution to this. The idea of "night kills" makes setting up a problem out of my depth. The only way I can think of is to brute-force the problem: find all possible combinations of 3 mafia lynches and add 'em up.
There can only be between 3-6 days in this "game", which makes the calculation manageable.
+ Show Spoiler +Day 1 Mafia Lynch:
(3/13)*(2/11)*(1/9)
(3/13)*(2/11)*(8/9)*(1/7)
(3/13)*(2/11)*(8/9)*(6/7)*(1/5)
(3/13)*(2/11)*(8/9)*(6/7)*(4/5)*(1/3)
(3/13)*(9/11)*(2/9)*(1/7)
(3/13)*(9/11)*(2/9)*(6/7)*(1/5)
(3/13)*(9/11)*(2/9)*(6/7)*(4/5)*(1/3)
(3/13)*(9/11)*(7/9)*(2/7)*(1/5)
(3/13)*(9/11)*(7/9)*(2/7)*(4/5)*(1/3)
(3/13)*(9/11)*(7/9)*(5/7)*(2/5)*(1/3)
= ~.08125
Day 2 Mafia Lynch:
(10/13)*(3/11)*(2/9)*(1/7)
(10/13)*(3/11)*(2/9)*(6/7)*(1/5)
(10/13)*(3/11)*(2/9)*(6/7)*(4/5)*(1/3)
(10/13)*(3/11)*(7/9)*(2/7)*(1/5)
(10/13)*(3/11)*(7/9)*(2/7)*(4/5)*(1/3)
(10/13)*(3/11)*(7/9)*(5/7)*(2/5)*(1/3)
= ~.06260
Day 3 Mafia Lynch:
(10/13)*(8/11)*(3/9)*(2/7)*(1/5)
(10/13)*(8/11)*(3/9)*(2/7)*(4/5)*(1/3)
(10/13)*(8/11)*(3/9)*(5/7)*(2/5)*(1/3)
= ~.04262
Day 4 Mafia Lynch:
(10/13)*(8/11)*(6/9)*(3/7)*(2/5)*(1/3)
=~.02131
Add this all up and we get a ~20.778% chance of a town victory in this setup (assuming random lynches).
I hope that someone more mathematically minded can provide a "cleaner" solution (or just check my math), but this should illustrate that the chances of town randomly winning in a "base game" is far lower than a typical game with active players.
|
Blazinghand
United States25550 Posts
On March 18 2015 00:14 plotspot wrote:
I think you understood me partly. But I was not talking about rules to forbid posting. I meant a minimalistic game out of free will, where no one chooses to talk at any time. The game with the smallest amount of posts possible. What would it look like? I just want to grasp the lowest possible boundary for this game, regardless of whether it makes sense or is the norm in the first place.
It seems pretty obvious what it would look like: each Day, you'd have a day post by the host, then you'd have everyone vote without discussing, then silence for 72 hours until the next day started, with maybe a post or two saying "I'm still here" so that people meet minimum activity requirements. It'd be boring as heck and would not actually exist since people want to play this game instead of not playing this game.
|
On March 18 2015 03:00 Hapahauli wrote:
Add this all up and we get a ~20.778% chance of a town victory in this setup (assuming random lynches).
Well mafia would obviously not vote randomly, did you take that into account?
|
On March 18 2015 03:19 raynpelikoneet wrote:Show nested quote +On March 18 2015 03:00 Hapahauli wrote:
Add this all up and we get a ~20.778% chance of a town victory in this setup (assuming random lynches).
Well mafia would obviously not vote randomly, did you take that into account?
Nope. There's also noooooo way I'm going down that road. There's a lot of strategic-choice/game-theory behind how mafia chooses to vote, and how a town (knowing about this mafia strategy) would choose to act. It would take me days to figure that out.
I'm more trying to illustrate what a "base game" can look like, and the general point that a "base game" is pretty unfavorable to town. Add coordinated voting for mafia and you'll probably end up lower than the ~20-21% mark I solved for. But whether we're at 5%, 10%, or 20%, the general point is the same.
|
Blazinghand
United States25550 Posts
On March 18 2015 03:45 Hapahauli wrote:Show nested quote +On March 18 2015 03:19 raynpelikoneet wrote:On March 18 2015 03:00 Hapahauli wrote:
Add this all up and we get a ~20.778% chance of a town victory in this setup (assuming random lynches).
Well mafia would obviously not vote randomly, did you take that into account? Nope. There's also noooooo way I'm going down that road. There's a lot of strategic-choice/game-theory behind how mafia chooses to vote, and how a town (knowing about this mafia strategy) would choose to act. It would take me days to figure that out. I'm more trying to illustrate what a "base game" can look like, and the general point that a "base game" is pretty unfavorable to town. Add coordinated voting for mafia and you'll probably end up lower than the ~20-21% mark I solved for. But whether we're at 5%, 10%, or 20%, the general point is the same.
I think the better way of putting this is that if town decided that votes would be random, instead of each individual voting randomly, we'd use my RNG system and have everyone vote for whoever was RNGed. This way, no amount of scum voting one way or another could impact things, it would be pure RNG.
Yet another way in which my RNG system is better than most people's RNG system.
|
Your RNG system has hit mafia 22% of the time 0% of the time. 
|
Ah ok, interesting. 20% (the actual number doesn't matter so much). So balance is indeed reached when both parties try to win. Does it mean that if the mafia is acting randomly instead of directionally, it also becomes a game of luck with the townsfolk winning 20% of the time? The problem I see for mafia to act randomly is that there will be cases where they have to explain themselves, and this will be difficult, unless they can really fit their apparent inconsistencies into a neat system. But they can randomly kill people, that is a valid strategy, right? So are there any comparison of success for random kill vs target kill strategies? Being naive to the game, I can imagine while it's good to prioritize targets, it also paves a way for others to recognize your motives.
|
Blazinghand
United States25550 Posts
On March 18 2015 05:45 plotspot wrote:
Ah ok, interesting. 20% (the actual number doesn't matter so much). So balance is indeed reached when both parties try to win. Does it mean that if the mafia is acting randomly instead of directionally, it also becomes a game of luck with the townsfolk winning 20% of the time? The problem I see for mafia to act randomly is that there will be cases where they have to explain themselves, and this will be difficult, unless they can really fit their apparent inconsistencies into a neat system. But they can randomly kill people, that is a valid strategy, right? So are there any comparison of success for random kill vs target kill strategies? Being naive to the game, I can imagine while it's good to prioritize targets, it also paves a way for others to recognize your motives.
So town has two things that increase their winrate:
1. town can identify people who are playing like scum, and vote them instead of voting randomly (this is the important one)
2. town has power roles like the cop or the vigilante or the doctor who can trip up scum quite a bit (less important, but valuable)
If mafia shoots randomly, this is much worse than shooting people who are trying to kill the mafia. For example, let's say the following players are alive in a game:
Palmar, the Vanilla Town
Jackal58, the Vanilla Town
Ver, the Vanilla Town
Blazinghand, the Mafia Goon
It's currently night time and I'm deciding who to shoot. Now, I could shoot randomly if I wanted to. But, Jackal58 is 100% sure I'm town, Ver is on the fence, and Palmar is 100% sure I'm scum. Jackal and Ver are both mildly suspicous of Palmar, and Ver is suspicious of Jackal.
I could shoot randomly, or I could do something awesome like shooting Palmar. If I shoot Palmar, I won't get lynched tomorrow since the player who wants to lynch me is dead. So, this is why shooting and picking your shots is way better than shooting randomly. The 20% winrate is "mafia shoots randomly and town lynches randomly" both of which are strategies of limited use.
The other time to shoot is if someone says "I'm the Doctor" or something. Definitely shoot that guy, he's the Doctor. You don't want him to be alive.
There are no formal comparisons of success between random and non-random kill strategies, mostly because scum is ALWAYS picking a target. It's generally considered super obvious that random NK is not good. "misleading NK", where you shoot someone who is defending you, happens, but random NK has not in my experience happened.
Edit: Also, you might be overestimating how clever town is. In the example I gave, I shoot Palmar because Palmar wants to lynch me, right? Now you think that a clever Town player will look at who the scumteam shoots and doesn't shoot and uses that to catch them. Well, that's a good strategy, and one that I personally use, but it's a lot harder to use than you think. Most town players, unlike me, pretty much ignore who was shot (don't read their filters, etc) because they've got a lot on their hands in any given day, which is totally understandable. The "who died" info is just a little piece of signal in the massive amount of data that is the game. Singling out could help you, maybe.
Also, look who I left alive. The two players beside me alive-- one of them thinks I'm definitely town, and the other one thinks that guy is scum. By shooting people who suspect them, Scum silences them! Over the course of the game, the number of people who think the scumteam is scum will go down as scum shoots them, making it harder to convince scum. Scum likes to leave around quiet, unuseful players, and players who defend the scumteam.
|
On March 18 2015 05:45 plotspot wrote:
Ah ok, interesting. 20% (the actual number doesn't matter so much). So balance is indeed reached when both parties try to win. Does it mean that if the mafia is acting randomly instead of directionally, it also becomes a game of luck with the townsfolk winning 20% of the time? The problem I see for mafia to act randomly is that there will be cases where they have to explain themselves, and this will be difficult, unless they can really fit their apparent inconsistencies into a neat system. But they can randomly kill people, that is a valid strategy, right? So are there any comparison of success for random kill vs target kill strategies? Being naive to the game, I can imagine while it's good to prioritize targets, it also paves a way for others to recognize your motives.
Maybe my English is bad at the moment, if I understand you right, I think you are misunderstanding balance. The hosts balance the game with typically 10 town and 3 scum in a game and give some sort of appropriate roles/abilities for certain people on each side.
And regarding luck, and that's what the student games with coaches are for. If you RNG (randomly assigned) scum, you will have a scumteam coach that will teach you how to mislead town, your game relies on misconception/misdirection, that's your win condition. It's up to town to be able to detect that (you will hear a term called scumhunting in this forum) and there are many ways of doing so and when they do, they vote you down. There's filter reading and trying to trace inconsistency, voting analysis, process of elimination, etc. Just a few of many methods.
There's no randomness to this. It's all subjective. At least not in a normal game. In more complicated themed games, there might be some more factor of randomness but for the purposes of this student game, no.
Everyone has a defined wincon that is determined by their faction's collective strategy.
Kills (nightkills) are never RNGed. If you are town vigilante at night you are asked to submit a target's name. If you are scumteam, you do the same. If you wish to yolo random your target, I mean, that's on you but it's obviously discouraged.
|
Damn it BH, you ninjaed me. Sort of. <3
|
On March 18 2015 05:45 plotspot wrote:
Ah ok, interesting. 20% (the actual number doesn't matter so much). So balance is indeed reached when both parties try to win.
Yep. Town winning % is positively correlated with things like activity and overall player-skill.
Does it mean that if the mafia is acting randomly instead of directionally, it also becomes a game of luck with the townsfolk winning 20% of the time? The problem I see for mafia to act randomly is that there will be cases where they have to explain themselves, and this will be difficult, unless they can really fit their apparent inconsistencies into a neat system. But they can randomly kill people, that is a valid strategy, right? So are there any comparison of success for random kill vs target kill strategies? Being naive to the game, I can imagine while it's good to prioritize targets, it also paves a way for others to recognize your motives.
I think you are too focused on the "concrete" actions of the game. You are correct that interpreting night-kills, lynches, and other "certain" things in the thread is a part of the game. However, a great deal of the game is reading a player's posts, deciding if he's being "genuine"... etc. Those "soft" reading skills are a great deal of the "skill" in this game. These skills have logical foundations, but they aren't the "concrete" types of evidence you are referring to.
|
Please remove me from the coach pool
I'm having IRL stuff going on and I'll have to be away from Mafia for a long while
Thanks
|
Blazinghand
United States25550 Posts
On March 18 2015 06:32 GlowingBear wrote:
Please remove me from the coach pool
I'm having IRL stuff going on and I'll have to be away from Mafia for a long while
Thanks
Saint Patrick's Day can hit pretty hard, I feel ya
|
I see, it's kinda hard to act as if nothing happens, when stubborn people starts to converge on you.I'm just trying to understand why randomness cannot be enforced by the mafia, since theoretically this is something good. If a game remains totally random, as the mafia you have a 80% win chance, which I think every mafia would take instead of the real 55% win chance estimated a page earlier given millions of mafia games.
The question is how can the mafia enforce as much randomness or (as the guides calls it) "confusion among the townspeople" as possible, obviously not by "standing out", which includes acting randomly, when everybody else is not. Lol I bet no townspeople will act randomly risking to be lynched. So if you're mafia you're actually playing as if you're townspeople, and the inevitable imperfection of that performance will eventually come back to lynch you?
|
Blazinghand
United States25550 Posts
Yeah, basically. If town decided to all act random for no reason, then sure, that would help mafia, but there's no way for mafia to MAKE town do that. Your strategy as mafia is usually "pretend to be a town player who is acting normally" which is fairly difficult.
|
On March 18 2015 06:37 plotspot wrote:
I see, it's kinda hard to act as if nothing happens, when stubborn people starts to converge on you.I'm just trying to understand why randomness cannot be enforced by the mafia, since theoretically this is something good. If a game remains totally random, as the mafia you have a 80% win chance, which I think every mafia would take instead of the real 55% win chance estimated a page earlier given millions of mafia games.
The question is how can the mafia enforce as much randomness or (as the guides calls it) "confusion among the townspeople" as possible, obviously not by "standing out", which includes acting randomly, when everybody else is not. Lol I bet no townspeople will act randomly risking to be lynched. So if you're mafia you're actually playing as if you're townspeople, and the inevitable imperfection of that performance will eventually come back to lynch you?
Because "random" play has nothing to do with the "soft" portions of the game.
Detecting lies from a post (or a string of post) is largely about how someone is pushing an objective, and not what they are pushing. Doing something random involves the former and not the latter.
And yes, a strong way to play mafia is to make the game seem "random" by creating chaos, derailing discussion... etc. That is the skill of mafia play, going against the skill of the town.
|
lol
most of the most effective players use both hard (votes, nks, etc.) and soft (posting, tone) evidence ^^ but getting the balance right can be tough
|
|
|
|
|
|
EPT
