On Improving In Mafia

[RadiantCowbells]

i think towns are a little worse than maybe like 2018 but not like ~significantly~ so

i am not sure of the truth of this claim because i didnt see nearly as many total town blowouts in 2018 nor across as diverse a selection of players
but shoshin irrelephant nsg etc all had insane years in 2018 and played a lot of games
what sets 2019 apart to me is the skill floor is a lot higher than it used to be, even lobbies that arent the absolute strongest usually are a lot more cohesive than their counterparts in the past

[Alyssa The Lamb]

My problem with the winrate argument as well as the claim that towns are getting better has several parts:

1. The

peak

of townplay has gotten better, but this has also coincided with a shift of what is deemed to be 'acceptable' in games, for lack of better ways to put it. It means that generally, people are better at reading or playing around weaker players and/or players that don't put more than the barebones amount of work into the game. This directly helps reinforce that players that do not wish to put in that effort do not have to put in that effort, and so they do not improve at the game in ways that they would have before. Since they aren't willing to put in that effort, they are more likely to follow those that are. This increases the likelihood of those with the highest impact and 'skill' having control over the game, but it also doesn't reflect on a town's skill more than it does a single player's skill.

2. The general 'balance' of games has gotten increasingly complex over the past several years with factors such as combined roles, more whitelisted roles, more complex game types, etc. A lot of this has resulted in a meta shift for balance in that towns have a lot of wiggle room to work with and don't necessarily need to be that strong at the game in order to have a good result thanks to the amount of power they have, while I've seen a pretty drastic uptick of games (and quite a few of the normals I've played and seen fall into this) where the scumteam have increased pressure to make very good use of their roles and setup spec in order to have a decent chance of winning a game. It takes a lot more work than it used to in order to be adept at mechanics from the scum side than it was several years ago where I was seeing scumteams be able to manipulate games into a state where they would be able to either outright conftown themselves via mechanics or be easily advantaged as such.

3. The average amount of players per game has generally gone lower over time. This is pretty simple to explain; the playerbase has definitely gotten smaller over the years, and most of the games firing are either micro or mini (9-13p) games. I suspect the breakpoint is exactly where you go from 2 scum to 3 scum, but there is a point where if you go below the line, towns need to do nothing except for kill one scum and they are likely to get a game-winning advantage from that alone, while that's not necessarily the case when you go above that line. The generally higher % of games that go below that line means that there's that much more incentive to just directly scumhunt, which favors read% over everything else.

Those three factors in some way have directly contributed to the metagame shifting to what it is now; normal games have tried to balance this meta in a very clunky way, basically by brute forcing the scumteam to focus on mechanics or just get screwed over by not doing so. This makes for really bad feeling setups to play as scum (hi jazz mafia!) where you are basically railroaded into a path that you might not necessarily have any choice in taking to have a chance of even winning.

I'm frankly not surprised that scumplay has gotten worse over the past several years. Playing scum simply is not fun in those conditions, but the way towns function now require it to be that way. I want to fix this by encouraging people to work on improving their own townplay, which would have a cascade effect from fixing my first point into eventually fixing the second point.

I really just want games that feel more fun to play. Playing either alignment within the past year and a half or so has been a huge chore and I'd really like it if more players treated it the same way I did; as a tool to help look inside my own mind, to understand how I think, and to train critical thinking and ingenuity skills. The site is healthier for it.

[Alyssa The Lamb]

I'd like to reiterate that I specifically mentioned the metagame going downhill.

This does not correlate to town winrates. I am saying that the

quality

of games is going downhill, not that town is losing more of them.

[OkaPoka]

what do you miss about the quality of past games vs today

[skitter30]

to understand how I think, and to train critical thinking and ingenuity skills.

for me at least this is a large reason why i like playing; mafia, at least in some ways, fills a critical thinking need that i have that doesn't get filled in other parts of my life.

that's why i like town more than scum; i like the process of ~solving~, of looking at all the puzzle pieces and figuring out how to make them all click into place

playing scum hones a different skillset for me, it enhances my rhetorical/manipulative skills. i feel like i'm more ~crafty~ in arguments than i used to be. i'm not sure if that's a good thing or not.

[RadiantCowbells]

1. The peak of townplay has gotten better, but this has also coincided with a shift of what is deemed to be 'acceptable' in games, for lack of better ways to put it. It means that generally, people are better at reading or playing around weaker players and/or players that don't put more than the barebones amount of work into the game. This directly helps reinforce that players that do not wish to put in that effort do not have to put in that effort, and so they do not improve at the game in ways that they would have before. Since they aren't willing to put in that effort, they are more likely to follow those that are. This increases the likelihood of those with the highest impact and 'skill' having control over the game, but it also doesn't reflect on a town's skill more than it does a single player's skill.

i... really feel like this is the exact opposite of true given my memories of 2016 where games were filled with people who would just shitpost until someone decided to play town leader and then followed their town leader
but i think you are possibly referring to a time before that and given comments ive heard about missing policy lynching you may be right!
my outlook on it though is that people who arent as super directed about things are the glue that keeps town together and that when people talk about *high caliber playerlists* being awful they usually mean when everyone really wants to do their thing as opposed to it being the quality of players

i definitely think that true lurkers are a real problem and they can fuck off but over time those true lurkers are more and more likely to be scum

[Alyssa The Lamb]

what do you miss about the quality of past games vs today

In andom order,

1. It's very rare for me to roll scum in a game feeling like I both have a good amount of agency in how to approach a game and end the game feeling like I wasn't trapped by the setup in a way.

2. The standards for players considered to be strong at town are much higher now than they used to me.

3. There's much more tolerance for play that has a negative effect on the game; I have some blame for this since I generally tolerate this too (and do it myself some games, admittedly), but it is something that is much harder to fix in game than it used to be.

[RadiantCowbells]

2. The standards for players considered to be strong at town are much higher now than they used to me.

is this poorly phrased or do you mean it as is

[skitter30]

2. The general 'balance' of games has gotten increasingly complex over the past several years with factors such as combined roles, more whitelisted roles, more complex game types, etc. A lot of this has resulted in a meta shift for balance in that towns have a lot of wiggle room to work with and don't necessarily need to be that strong at the game in order to have a good result thanks to the amount of power they have, while I've seen a pretty drastic uptick of games (and quite a few of the normals I've played and seen fall into this) where the scumteam have increased pressure to make very good use of their roles and setup spec in order to have a decent chance of winning a game. It takes a lot more work than it used to in order to be adept at mechanics from the scum side than it was several years ago where I was seeing scumteams be able to manipulate games into a state where they would be able to either outright conftown themselves via mechanics or be easily advantaged as such.

i also kinda agree with this point ...
mechanically, at least in normals, towns tend to have the upper hand.
towns tend to be loaded with prs
scum are given like a random pr or two that often has very little to do with the prs that town have
(i remember being involved with 3 normals at the same time where scum had a neighbor and the town setup ... had nothing to do with neighborhoods. i feel like from scum's pov if you know there's a neighborhood you'd expect town to have a pt cop, or something of that nature. When in fact town has nothing of the sort. It felt like a red herring and that we were almost being punished for assuming town prs had anything to do with the prs we had)
like unless you use the scum prs super super well they tend to backfire on you because scum usually don't have enough of the setup to know how to use them properly in context

i'm not sure if that's the point you were making but I've noticed that the mechanical balance in normals is kinda screwy when you compare how town is expected to use their prs and how scum are effectively expected to use theirs.

[Alyssa The Lamb]

1. The peak of townplay has gotten better, but this has also coincided with a shift of what is deemed to be 'acceptable' in games, for lack of better ways to put it. It means that generally, people are better at reading or playing around weaker players and/or players that don't put more than the barebones amount of work into the game. This directly helps reinforce that players that do not wish to put in that effort do not have to put in that effort, and so they do not improve at the game in ways that they would have before. Since they aren't willing to put in that effort, they are more likely to follow those that are. This increases the likelihood of those with the highest impact and 'skill' having control over the game, but it also doesn't reflect on a town's skill more than it does a single player's skill.

i... really feel like this is the exact opposite of true given my memories of 2016 where games were filled with people who would just shitpost until someone decided to play town leader and then followed their town leader
but i think you are possibly referring to a time before that and given comments ive heard about missing policy lynching you may be right!
my outlook on it though is that people who arent as super directed about things are the glue that keeps town together and that when people talk about *high caliber playerlists* being awful they usually mean when everyone really wants to do their thing as opposed to it being the quality of players

i definitely think that true lurkers are a real problem and they can fuck off but over time those true lurkers are more and more likely to be scum

I've seen more metas than most people on the site have, I'm pretty certain.

2006, 2008, 2011, 2013, 2016, 2019 all were years I played on the site and all have felt very different. It's entirely possible I'm mashing some of them together, but it's also true that the only two times I saw a major shift in a relatively small amount of time is when the hyper posting era started (I think 2014?)... and when I came back in 2019, where I felt less agency in a game as town in a way that wasn't the case in any year before then. That's the best way I can describe that feeling, at least. Something shifted around that time and I'm still trying to quantify what exactly that is.

[Alyssa The Lamb]

2. The standards for players considered to be strong at town are much higher now than they

used to be

.

is this poorly phrased or do you mean it as is

Probably autocorrect.

[Alisae]

I'll just say that I don't agree with a lot of stuff in here.

Even just the premise that "site meta is going downhill" when town are performing better and better and winning more and more games.

Do you think towns are performing better because they are actually getting better, or that towns are performing better because scum don't want to play scum? what's going on here exactly.

[Knightmare491]

I found this post helpful overall. Thank you for making this article/thread.

[Porochaz]

I havent played in a non-invitational/non-oldie/non-f2f in years but I thought the site meta was on an uptick, because the meta a few years ago utterly shockingly poor. What I find is that it kinda goes in cycle of hyper-analytical to one-liner/point at someone randomly. I skimmed your article, I think looking at it from a one size fits all category really doesn't work (which I think you are pointing out). I prefer playing mafia to town, always have. I have more great scum games than town games. (although I have a couple of shockers as well) I also used to try and be inconsistent with my playstyle. Both of which I don't think is usual within the sites meta (at the time I was playing). I also played a very meta heavy game, which was well within the sites meta in the early days and went very badly against it in my last few - the last game I remember I was lynched because of it. Which was frustrating because yep, I was scum, but it would have been something I would have done as town as well. But I guess that some of that falls down to me, in that as site meta evolves you have to evolve with it.

As I say, Im a few years out of date but when I last modded everyone was at each others throats/playing for themselves which meant a lack of useful interaction. Whilst I do partially mean "one-liners" or vote posts, I also mean pages upon pages of just the same person posting 37 times straight or 2 people viciously arguing without a real point behind it. I felt both in terms of playing and modding the aim of the game had turned into exhaust the other players to quit/replace out. I had thought the site had gotten better for that? In which case I think the site meta has improved?

Looking at win-rates doesn't seem to be helpful to me. If you want to talk about site-meta in terms of quality of play then win-rates shouldn't account for it. Having now read the whole thread I see people have hit upon these points already and as when I post in md threads I have to remember that a lot of the people who played when I started in 2007 don't play anymore and we have now gone through a few generations and iterations of the game. The people commenting here are mainly people who have joined in the last few years and are evolving their own site meta. Looking at past metas and thinking those are better or worse is distinctly unhelpful and will cause ill-feeling. (unless you are talking about the shit-talking/spam era) I think for this era you have to take into account the smaller player-base which means player meta is likely stronger. Smaller unthemed games as far as Im aware are taking precedent which Im sure effects things - themed games were more popular in my day, where new and unusual roles and mechanics allowed for scope for shenanigans/things for people to hide behind. It sounds like quite an open meta at the moment - which would contribute to a higher town win-rate.

In any case, the talk about game setups being worse etc. shouldn't even be an argument. Things like more properly defining normal roles, having a setup committee, feedback forms, more stringent modding/punishments etc should all be contributing towards better setups. (even if I feel that some of those things have caused more harm than good) However in my view thats a separate argument from the one being discussed here.

[Porochaz]

Also as someone who has obviously played under different accounts I will assume I have played with you at some point. I won't ask who you are but I presume that's correct?

[Alyssa The Lamb]

Also as someone who has obviously played under different accounts I will assume I have played with you at some point. I won't ask who you are but I presume that's correct?

If you are talking to me here, I don't remember specifically playing with you before. If we have, it was around the 2011-2014 era; I remember very little about the games I played back around that time.

[Ranmaru]

I realize practice makes perfect when it comes to improving. Which I guess can be applied to point 3. I know what helped me improve was better understanding of game theory for mafia. Which I could learn via trial of error, but talking game theory with more experienced players can be exhilarating. Finally reflecting on your own play / self critique is probably very important for improving. Watching my own splatoon 2 play can be interesting, but reading my own play can be embarrassing at times but it helps immensely to do so.

TL;DR: Practice, Learn, Reflect.

[Zachrulez]

I can go into this more later, but I think the town winrates going up can be pretty easily explained in a way that still correlates with skill level average still going down.

Titus is the closest to my own theory, but more in the general thought process more than the details.

Winrate arguments are lazily easy to make in arguing town improvement. In my early days on the site town could be tasked to win a game where they had no power roles to combat the scum. This is not something that happens anymore with any kind of regularity and the fact that it doesn't happen naturally leads to higher town winrates. (Man I really miss playing and winning these kind of games. That level of satisfaction couldn't be topped in mafia.)

There was also a pretty active movement years ago to make normal games less 'scum sided'. (Another way of saying more town sided.) Town winrates go up so town must be doing better right?

[northsidegal]

In my early days on the site town could be tasked to win a game where they had no power roles to combat the scum. This is not something that happens anymore with any kind of regularity and the fact that it doesn't happen naturally leads to higher town winrates.

I have a few things to say in response to this and some other points brought up earlier. First, is it really true that mountainous and other near-vanilla setups were won more often in the past?

I don't mean to call you a liar, it's just that this conflicts with what I thought I knew about those setups' history, as well as what information I can find on the subject. I can't find a single instance of Town winning a mountainous setup in the past: the wiki lists two losses, and searching reveals 1 2 3 losses that I could find. There is only one instance I know of personally of Town winning a 2:11 mountainous setup, and it happened relatively recently in the site's history. White Flag (Open Setup)'s history is similarly bad, as a look at the wiki's history shows.

Secondly, I believe there is a contradiction in your second sentence. If I understand correctly, you're saying that Town skill has dropped such that they now need power roles to win, and thus given these power roles it is natural that Town winrates would be higher. But, aren't you also saying that Town skill has gone down? Given both of the premises – Town skill dropping and the introduction of power roles, it doesn't seem natural to me that winrates would

necessarily

go up. Of course, you may say that the power roles offset the drop of skill in such a way that winrates still go up, which seems logically sound (even if I might disagree).



On the subject of winrates, it's useful to first establish what exactly the claim being made is. Do I believe that winrate necessarily correlates with skill? No. I do, however, believe that, on a general rather than individual level, there are few better indicators. I also believe that when it comes to setup balance, there is basically no other metric to attempt to optimize for other than winrate. That is, I think that whether or not people believe "towns are only winning because of setups" and whether or not that's even true, if on the whole winrates are balanced between town and scum, that is the best place for setup balance to remain. I think that, put this way, few people would disagree with this? Certainly I might expect a not-unjustified amount of frustration towards this in certain scenarios, but I don't think there's anything better to do than to balance for the skill of the people actually playing the setups.


Finally, I thought I'd share some Normal game statistics which I thought would be interesting. Going by start date, there were 35 normal games across the year of 2018. Of these 35, I counted 19 Town victories, giving an overall town winrate of 54%. In 2019, there were 32, of which 17 were Town wins, giving a winrate of 53%.

I've been noticing a sentiment that people dislike too much complexity in Normal games for a long time now. I don't disagree – having recently updated the Mini Normal archives I saw no shortage of games that seemed very non-normal and some setups that I think should never have been run, yes. That being said, it seems to me like Normal game balance is in about as good a state as it could be.

[Alyssa The Lamb]

Finally, I thought I'd share some Normal game statistics which I thought would be interesting. Going by start date, there were 35 normal games across the year of 2018. Of these 35, I counted 19 Town victories, giving an overall town winrate of 54%. In 2019, there were 32, of which 17 were Town wins, giving a winrate of 53%.

I've been noticing a sentiment that people dislike too much complexity in Normal games for a long time now. I don't disagree – having recently updated the Mini Normal archives I saw no shortage of games that seemed very non-normal and some setups that I think should never have been run, yes. That being said, it seems to me like Normal game balance is in about as good a state as it could be.

This really saddens me : /

[Zachrulez]

In my early days on the site town could be tasked to win a game where they had no power roles to combat the scum. This is not something that happens anymore with any kind of regularity and the fact that it doesn't happen naturally leads to higher town winrates.

I have a few things to say in response to this and some other points brought up earlier. First, is it really true that mountainous and other near-vanilla setups were won more often in the past?

I don't mean to call you a liar, it's just that this conflicts with what I thought I knew about those setups' history, as well as what information I can find on the subject. I can't find a single instance of Town winning a mountainous setup in the past: the wiki lists two losses, and searching reveals 1 2 3 losses that I could find. There is only one instance I know of personally of Town winning a 2:11 mountainous setup, and it happened relatively recently in the site's history. White Flag (Open Setup)'s history is similarly bad, as a look at the wiki's history shows.

Secondly, I believe there is a contradiction in your second sentence. If I understand correctly, you're saying that Town skill has dropped such that they now need power roles to win, and thus given these power roles it is natural that Town winrates would be higher. But, aren't you also saying that Town skill has gone down? Given both of the premises – Town skill dropping and the introduction of power roles, it doesn't seem natural to me that winrates would

necessarily

go up. Of course, you may say that the power roles offset the drop of skill in such a way that winrates still go up, which seems logically sound (even if I might disagree).

The F11 newbie setup routinely had no power roles vs the scum. Town wins exist in that format. I was never arguing that town winrates were ever anywhere near reasonable for these kind of setups. Naturally town win rates will rise once those setups and scenarios no longer exist. I'm operating under the assumption these setups are way less common now. F11 was being run all the time when it was active for example. (Has it even been run at all since it was deactivated?)

I also offer no commentary on the current site's skill level on townplay. I actually have no experience in the current meta as I haven't played in years. I just know from knowledge of where the site meta has been moving back when I was playing that the movement in setups is definitively conducive to town win rates getting higher and that it's lazy to use town winrates as any kind of argument toward the state of the quality of town play vs say a mafia generation or two ago. (From around maybe 2014 and before?) Because back then you were playing with much weaker town teams power role wise than you are today.

Finally, I thought I'd share some Normal game statistics which I thought would be interesting. Going by start date, there were 35 normal games across the year of 2018. Of these 35, I counted 19 Town victories, giving an overall town winrate of 54%. In 2019, there were 32, of which 17 were Town wins, giving a winrate of 53%.

I've been noticing a sentiment that people dislike too much complexity in Normal games for a long time now. I don't disagree – having recently updated the Mini Normal archives I saw no shortage of games that seemed very non-normal and some setups that I think should never have been run, yes. That being said, it seems to me like Normal game balance is in about as good a state as it could be.

This really saddens me : /

I personally get a lot more out of winning a difficult town game than an easy one but it's definitely clear the site has largely moved on from that idea.

[northsidegal]

The F11 newbie setup routinely had no power roles vs the scum. Town wins exist in that format. I was never arguing that town winrates were ever anywhere near reasonable for these kind of setups. Naturally town win rates will rise once those setups and scenarios no longer exist. I'm operating under the assumption these setups are way less common now. F11 was being run all the time when it was active for example. (Has it even been run at all since it was deactivated?)

I also offer no commentary on the current site's skill level on townplay. I actually have no experience in the current meta as I haven't played in years. I just know from knowledge of where the site meta has been moving back when I was playing that the movement in setups is definitively conducive to town win rates getting higher and that it's lazy to use town winrates as any kind of argument toward the state of the quality of town play vs say a mafia generation or two ago. (From around maybe 2014 and before?) Because back then you were playing with much weaker town teams power role wise than you are today.

I think I misunderstood some of what you were saying. I agree that a sweeping "how many games did town win this year across all queues" analysis would be confounded by so many factors as to practically be useless. I do think that there are more specific things that could be looked at that would be more useful, however—in such a way as to avoid the problem with winrates that you bring up.

I personally get a lot more out of winning a difficult town game than an easy one but it's definitely clear the site has largely moved on from that idea.

I don't think this is necessarily true. At the very least, it's not as clear to me as it seems to be to you.

As a side note, I wonder if what I was trying to say about Normal games is the same as how it comes across to others. Is roughly 50/50 winrates

not

the goal? Should it be something else?

[Alyssa The Lamb]

I vastly prefer games that are fun to play than balanced. As long as they don't feel overpowering for one alignment for the other, I don't care about balance.

[northsidegal]

I don't think the two are mutually exclusive. And, if you'll excuse my tone, if you care about a game feeling overpowering for one alignment or the other, it sounds like you do care about balance.

In fact, it seems to me like the mindset of "as long as this is fun the balance matters less" is what leads to very complex Normal setups, the kind of which I was under the impression that you disliked.


I think there's a dangerously easy tendency for conversations like these to lose track of what was originally being said, so let me just restate what exactly those things are. I think that Normal game balance is in about as good a state as it possibly could be. That being said, I think that some overly complex Normal games were passed last year, and I personally think that Normal complexity should be minimized.

I also think that the idea that Towns were significantly better in the past than they are now seems to be unfounded, although this is primarily based on my not seeing any evidence towards that being true rather than a strong belief that it is not true.

[Alyssa The Lamb]

Here's where our mindsets diverge.

I think that game feel and game balance are inherently different things. A game can be balanced by numbers and still feel really shitty to play, which the open setup listing should hopefully provide enough examples to show what I mean.

My problem is that in the normal games I've played, the scumteam either have unrealistic expectations of them compared to town, or the game is incredibly swingy to balance a specific role or player amount. This is functionally the problem I have with normal games. The fact that they're technically balanced means very little to me, because my idea of an optimal normal queue is relatively simple games that focus on day play while theme games can focus more on mechanical play. I've gotten this feeling in years past, but this has been lost in recent ones.

Fundamentally, I think this is probably a sizable chunk of why a lot of scum players can't get motivated to try in them anymore, which means that when competent towns appear, they ultimately stomp the game in my experience. I can describe what I mean by competent towns when I have time to sit down.

I call old towns better because they simply had to be. There was less of a cushion in PRs in a lot of games and there were more strong scum players. It's rather disturbing to have this reputation for strong scum play when I only ever got the impression that I got the advantage in games because towns destroyed themselves and i just happen to be half decent at amplifying that. That's not strong scum play, that's abusing weak town play.