Site Meta Thread

[Ellibereth]

People blame people other than themselves for losses and think they did most to everything right so they don't improve.

[Ellibereth]

I mean I understand the qualms of the whole geriatric crew and am an absolute supporter of that sort of game being a consistent thing.

But blaming losses hyperposters and "bullshit" in a nongeriatric game because you consequently can't keep up is silly. Someone could argue just as easily the blame lies on the person that's unable to keep up. It doesn't just go in one direction.

[Ellibereth]

(and ftr, making one random read and yelling a lot that this person is obvscum and therefore should be lynched is not, by any reasonable definition, "scumhunting")

I mean, it is if you take away the word "random".

It's just that if you want to do that succesfully in thread consistently there's probably a decent amount of hidden work you have to do that never makes it to the thread.

[Ellibereth]

I mean I think recognizing when someone else's reads are more likely to be right than your own is an important skill to develop. I'm plenty comfortable with being a sheep a whole game and I think there's nothing wrong with that.

The usual result of the scenario you're describing is the most charismatic person in the fight wins for better or worse.

[Ellibereth]

If the goal is improvement, I feel like most people (guilty myself) focus on improving other people over themselves.

Quoting myself from elsewhere:

Honestly - after most losses - there's probably some set of things any individual player could have done differently that would have flipped the result to a win. If the interest is in getting better that's probably the best mindset to have, and I think that the statement is generally true.

If everyone said "town lost this game I fucked up. I think my fuckups were BLANK, BLANK, BLANK" instead of "YOU GUYS SUCKED AAA THROW" I'm sure the play level would go up.

[Ellibereth]

Eh, they can be useful.

Almost everyone tool is useful in some context.

[Ellibereth]

No. All tonereads are bad. If you have tonereads you have a way to provide proof in order to solve a slot.

When I toneread I nitpick on word choice or if the stream of consciousness looks genuine, or scripted. I can go from there and see if I have anything to be used to solve the slot.

I don't faceplant in the thread saying "we can't lynch X because I tr their tone". Lazy ambigous posting like that hurts town.

Providing proof isn't necessarily good.
Ambiguous posting isn't necessarily bad.

"If you have tonereads you have a way to provide proof in order to solve a slot. " <- and this in particular isn't true anyway.

[Ellibereth]

The reason towns struggle:

1) Overconfidence in reads

2) People think non scummy things are scummy

This is literally it

Who disagrees with this

Meh.
For number 2 it goes both ways.
If you're scumread when you're town by another town player both players have room for improvement.

[Ellibereth]

Or you could be better at reading said type of posting?
And maybe there's something your missing when it comes to tonereading?
And maybe some of your beliefs on what good and bad play are, aren't necessarily true?

Like this absolute confidence you're displaying that your viewpoint is right is exactly what the problem is.

[Ellibereth]

What's so hard about having the attitude

"Damn, I misread them, maybe there's something I missed that could have helped me read them better. Maybe the metrics I'm using for scum and town are off somehow?"

vs.

"It's not my fault I misread them, they were playing like scum."

The latter is how things don't progress and it's why a lot of people (again, I'm guilty of this too) have applied the same faulty "tells" over and over and over.

[Ellibereth]

I mean I understand the qualms of the whole geriatric crew and am an absolute supporter of that sort of game being a consistent thing.

But blaming losses hyperposters and "bullshit" in a nongeriatric game because you consequently can't keep up is silly. Someone could argue just as easily the blame lies on the person that's unable to keep up. It doesn't just go in one direction.

It doesn't fully work both ways though.

If half of the game is making twenty posts a day (good content or nothing), and the other half cant reach that (work, school, etc) the people posting have a decision to make

1) Post less in order to allow the rest of the game to contribute
2) Continue posting while knowing that they are impacting the ability of players to contribute to the game.

You can argue that by intentionally hyper posting knowing that a portion of the game is unable to be effective due to it is in itself an anti-town action as you immediately cripple the amount of information that you get from a few players by your own choice. There is a difference between a lurker and a player who only has a few hours a day to play. Its a bit of hyperbole, but it would be like having a PR which allows you to limit the amount of posts someone could make to a detrimental point, and then arguing that its something that helps the town to use it on a random player.

I don't buy that in order to play a game on this site you have to be nearly constantly online. It should be a site where you can spend a few hours after work, or between classes, or things like that and be able to keep up with the game.

Yeah you're right.

I mean the argument a hyperposter could make there is "the way I'm playing is going to turn out to be so positive that it cancels out the negative for you being unable to read everything" just like the argument from your side is "it's worth it for you to change how you post in a way that potentially is more uncomfortable to you so I can read the game and be more effective".

Either argument can be true depending on the specific case. I think we intuitively think the hyper-poster lowering their volume is the more polite and expedeint thing to do because we feel like it wouldn't signifigantly impact their happiness level and that its fairly easy to do(we're not asking them to chop off an arm!). But who knows, maybe posting less is hard for certain people and would actually ruin their day. In the end of the day if they don't feel like posting less there's nothing you can really do once the game starts.

I agree with you that you should not have to be constantly online. There's the garuanteed geriatric set for that if you want to read everything. And even for "spam heavy games" I think there are techniques to be effective while only using an hour or so a day. Reading everything isn't absolutely required. It's just an adjustment. (One I had to make coming back to this site!)

[Ellibereth]

Biggest thing for me is ego and “me first attitude”.

Don’t get the sense that most people are really interested in playing mafia here so much as having shouting circles insisting that they're "more right”.

and the (unfortunate?) adjustment a player could potentially have to make is if they think all the loud people's reads are wrong and that there own are correct would be to shout even LOUDER. And a lot of us are bad at knowing when we're more likely to be wrong than someone else so the cycle goes on.

[Ellibereth]

I should have confidence on a subject I know well.

It's called taking a stand on something. Opposed to being unclear about my motives.

My viewpoint is "this concept can lead you to better evidence but offer nothing worthwhile to yourself or town". If you dispute that by genuinely believing that solving slots isn't in your interest, or that vague reads on players is okay, then keep these posts in mind the next time you find yourself in a situation like this in a game.

I've usually felt that the more I learn about something, the more I become aware that I still don't know a lot about that thing. That's applied to mafia as well.

My dispute is that I could feel that it could offer something worthwhile and that convincing evidence sometimes isn't the most important thing in the world. I'm also not sure what game situation you're describing.

[Ellibereth]

Oh and don't use a technique until you know how to use it.

I disagree with this. I think people should feel comfortable experimenting with things even if they're not sure if it works or not. There's no mafia "sandbox" or "practice" mode to do that in so the only way to test ideas is in actual games.

[Ellibereth]

My main usage of meta is that I think that people have a bunch of unconcious hard-data tells scattered across their games. I use some primitive programs to help me find them.

[Ellibereth]

Lycan, would it blow your mind if I thought that pushes with weak/few to no accusations/reasons that the author believes is untrue are often as or more effective as ones with a case and real reasons?
If I'm able to get someone lynched with either one (and thats very often the case), I would choose the former option almost every time.

[Ellibereth]

Weak cases are usually symptomatic of a TvT or SvS. Succeeding with lynching bait isn't anything to be proud of.

I think you're mixing up two things.

Having good reads

and

choosing whether to state reasons in thread at all and choosing whether to be honest with ones reasoning

are two different things.

It sounds like me that you're used to people that have bad reads and also choose not to state their reasons or lie in thread.

My claim is that people with good reads may (and often do) choose not to state their reasons or lie in thread OVER writing out a strong case and discussing their real reasons, because they believe the former is more effective in a lot of scenarios.

[Ellibereth]

I'm copying this over from the other thread.

I actually talked to some people about this recently.

A pretty big problem with mafia as whole is that there's very little consensus on what "good" or "bad" play is. Like for a lot of things there are people objectively better at some things than other but the lack of metric for it makes it hard to get a sense of whats actually "good" and "bad".

See Chess, Starcraft, League, Sports, (any other game) Players- there are Heros to emulate and almost everyone agrees they're good and that they're good at particular things. We don't have anything close to that. When I first started I found a few people I thought was good and read a shitton of their games and tried to emulate them. Some stuff stuck and some didn't. Some of things I was copying I now think aren't that good, others I think are but I'm just not good at.

But yeah - it would ideally be easier to have visible improvement lines, goalposts, and models but that would require some sort of consensus on a bunch of stuff that would be tough.

And yeah, the various disagreements floating around in this thread on some very fundamental stuff shows why its hard to get that consensus.

[Ellibereth]

In terms of good teamwork.

I think

everyone working together to solve a game

vs.

everyone finding someone to sheep, and that someone solves the game

are both equally good cases of teamwork.

The latter option is often more realstic and easier to achieve so I think it's a better starting point.

[Ellibereth]

I think you'll find a lot of arguments that concludes there's only "one method" is likely wrong. "Motivation" based reading is actually my last resort.

There are people on this site that will analyse their posts in excruciating detail to see what words, sentence length, post length, spread of topic, and lines of questioning to see what they do naturally as either alignment. And then they will draft their posts multiple times until it perfectly represents their town meta specifically to fuck with meta tells.

1) Most people DON'T do this and thus don't know their own set of unconcious tells. There's also a treasure trove of samples on this site and other mafia sites across the internet. Very few people bother to analyze them all thoroughly because of the work required. That doesn't mean the method is bad if it's done in a good way.

2) A competently written program (with varying amounts of human guidance depending on how it's coded) still might be able to find some unconcious tells that a player doing your cited "muffin" process might have missed. Things would be different if both sides put in the same amount of effort, then things would be more interesting. Then I would have to switch to other methods or try to code something even better. This arms race is actually easier for town though since I have to identify unconcious tells while the scum player has to identify which unconcious tells I've detected and try to correc them. However that's not the case right now since most people DON'T put in that type of effort.

[Ellibereth]

I mean regardless - you can't say there is only one "reliable" way to read people when there are players have demonstrated signifigantly higher accuracy over the norm consistently in terms of read accuracy and they use different techniques.

[Ellibereth]

I input a bunch of peoples posts they made with a known alignment into a program and it returns possible patterns. How is that motivation reading.

I attempt to build an ai that teaches itself to read alignment based on people's completed games. How is that motivation reading.

[Ellibereth]

If a competent group of people with enough time, motivation, and resources got together to try and build an ANN for forum mafia it's very likely the result would do a much better job than most humans. Said ANN wouldn't care about abstract human concepts like "motivation" at all.

[Ellibereth]

Tech has helped use realize our original "human" assumptions in many fields were not necessarily correct and that there were other ways of vieweing and analyzing things. Mafia should be no different.

The point isn't that everyone should use tech or some other tool- the point is that its very likely wrong to think there's only one way of doing things or some set of rules that have to be true because its what makes sense to us logically as humans. There are plenty of true things that we can't prove yet. There are tons of methods for various things that may work that we just don't understand well enough. That doesn't mean we operate under the assumption they're false or don't work until then.

[Ellibereth]

Do you think there are actions that humans do that have no motivation?

[Ellibereth]

Yes.

But I don't think that any of them exist in a game of mafia.

Therein lies our disagreement!

[Ellibereth]

most of my actions in mafia games tend to be motivated by my own personal enjoyment, not based on me trying to achieve win com so I think it’s hard to determine alignment based on something like that

Eh, I think just knowing your wincon can subconciously make changes to how you post.

[Ellibereth]

Mafia is a single player game with a buncha poorly coded NPC's

[Ellibereth]

Content is this:

Revolutionizing a game with a data-driven approach.

[Ellibereth]

DGB's gone though.

[Ellibereth]

how often does doing this motivation stuff get you a wrong townread.

[Ellibereth]

whats "pretty good" for town reads.

like 77% accuracy is just around random.

[Ellibereth]

you should actually measure the "more often than not" to see how much better than random you are .

I think most people are kinda vague and just judge their read accuracy by feel after games.

[Ellibereth]

We could also start a rating system for each alignment.
Using something like msoft's true skill.
Can x10 or x100 it to make the numbers prettier.

Everyone starts at same amount now and after every game just take avg rating of each team before game with the result and do + teamratingchange/teamsize to everyone.

and just not count multiball and bastard and 3p games and stuff like that for simplicity.

I'm also too lazy to do this (mhmsmith0).

edit: actually if we're compiling stats from the last year might as well count the last year in the above too...

[Ellibereth]

i lied and meant 75

assuming 10-3 is standard and you're town and randomly pick someone not you d1 and call them town you're right 9/12 so.

obviously when you're measuring this stuff you have to adjust for how many ppl alive and also ic's and confirmed roles/innocents/guilties since latter two aren't really reads anymore.

[Ellibereth]

you realize 2/3 right on a readslist for everyone isn't necessarily good right

for 12 player with 9/3 distribution and assuming you assign everyone T or S, the possibilites are then that you're right on 1 and wrong on 2 (if declaring 3 scum), that you have all the scum on your 'scumlist' but you're calling 7 people scum, that you have 2 scum right out of a list of 5 and have the last one as town, or you have 1 scumread and it's wrong.

[Ellibereth]

1 and 4 are just bad.
2 usually comes in practice as a decent PoE and not as I "think these 7 ppl are scum because they're scummy."
Same with 3 except its less conservative and has the hard misread.

[Ellibereth]

you realize 2/3 right on a readslist for everyone isn't necessarily good right

for 12 player with 9/3 distribution and assuming you assign everyone T or S, the possibilites are then that you're right on 1 and wrong on 2 (if declaring 3 scum), that you have all the scum on your 'scumlist' but you're calling 7 people scum, that you have 2 scum right out of a list of 5 and have the last one as town, or you have 1 scumread and it's wrong.

Right, but is it bad? If every Town member get's 2/3 wrong, what are Town's chances of winning if they all decide to lynch the people they agree are Scum?

it's going to be a mislynch because if there's an intersection (which isn't that likely), it's on the 1 person that's being scumread incorrectly in case 4.