Normal Game Balance

[Mizzytastic]

If you consider that example game emblematic it sounds like your problem may be more to do with swinginess of power roles instead of balance. It's not how much town is winning but how town is winning that feels bad to you.

This is important because game review cannot review games to make day play better. But if this is a wide spread attitude maybe we can find the problem roles and limit how many there are and see how that works. It may be that the NRG needs to be more picky on how town roles are combined without abandoning their (presumably statistically based) approach to balance.

Town being worse than EV is a pretty common result of studies into mafia. If balanced game designs feel bad to play from one side it would also make me wonder if the default game sizes are wrong for site meta and people would enjoy town having more VTs instead of power roles

[Mizzytastic]

I'm not voting cos I'm not experienced enough with the site to have an opinion. Just a game design nerd in general

[Something_Smart]

I think the biggest problem role in the game Menalque linked was actually the bulletproof. The scum checked the bulletproof, found out he was a PR, and wasted a shot on him which also led town to think that the person protected by the doctor that night was conftown. They then proceeded to leave the bulletproof, the vig, and the person who was supposedly doc-cleared alive till lylo (which sounds like bad play to me, or at least unlucky play), and the bulletproof happened to also be vengeful. Pretty big defiance of expectations. Bulletproof is a very swingy role, possibly on par with cop and vig. There's probably an argument there about how vengeful shouldn't work in lylo.

The point is that I see Menalque's grievances. The setup had the potential to be pretty gross, and by chance it happened to end up realizing that potential. But none of this has to do with

balance

; it all has to do with

swing

. Had that unlikely confluence of factors not occurred, the gross situation wouldn't have happened, and scum would have probably just won via faking a guilty in lylo like they were planning to.

So I think this is a case of addressing the wrong problem, at least with regard to that particular game.

[skitter30]

While we're at it something that bugs me that is actually kinda common is when scum prs have very little relationship to town prs. So when you're scum playing such a setup you really have no idea what you're going up against whatsoever. So you'll make an educated guess what the town prs are based on what you have, but it turns out you were barking up a totally wrong tree (and may have made a fatal misstep taking that educated guess ... )

Things like scum getting a ninja when there arent any tracker/watcher type roles in the setup, that sort of thing

This is one of my biggest pet peeves wrt setup-spec and balance ^

[Mizzytastic]

Also yeah, an unbalanced fun game is better than a balanced one people find a chore to play. But knowing its unbalanced can change how people perceive the fun of a game - same way feeling its unbalanced is doing for Mena now

Pedit: setup WIFOM like that bugs me too, I'd be far happier to see players pick up in the hints in my design and be rewarded for it

[implosion]

For what it's worth, there are plenty of normal game reviews where the reviewers will describe things as swingy but balanced. Designing to minimize swing in general is, imo, more challenging than people might realize. Part of this is because it can conflict with other design principles like simplicity; increasing simplicity would have you consolidate town power into fewer roles while reducing swing might imply you should spread it out. It often can conflict with balance, because we've found that for balance the town generally needs some very tangible power, but most tangible power introduces swing. The good majority of town power roles are swingy simply by nature, and those that aren't swingy are mostly not swingy because they don't impact the game all that much. Any investigative that gets tangible info will either find or not find what is best for them; any disruptive or protective role might just have no impact whatsoever if they target wrong; etc.

Another conflict is that between players and mods - mods will generally design the setups they want to run, not that players necessarily want to play, and it can sometimes be an uphill battle to convince a mod that their setup is a bad idea beyond simply conceptions of balance. Pre-designed setups can perhaps be held to a higher standard here.

[Mizzytastic]

Knowing that I could get feedback in the context of what players want would make me want to pass that threshold too. But I can't think of way of having that as an option to be reviewed for without it sounding like a value judgement. Player me is less likely to sign up for a game not reviewed for "what players want" if I know it's an option the mod had and refused unless I already like the mod.

[implosion]

With regards to balance it's important to mention, for those who are new to this discussion, the historical context. One old point that might be new information to many of you, I'm not sure, is that mini games used to have a 12 player limit, not 13 - this thread is the reason that changed. The other is that a few years ago (I think this is when I became listmod but I might be horribly misremembering tbh) there was a paradigm shift in reviews, because games were *still* scumsided. The change that resulted from that is that, in essence, every NRG member had to pretend that setups were a little bit more scumsided than their instincts thought... because they were, in fact, more scumsided than peoples' instincts thought. I think the NRG has pretty much internalized this at this point, which is reflected in generally fairly even winrates as of late, apparently, which I'm happy with.

If people as a whole want to shift the paradigm again to aim for something that is not a 50% win rate for each side, then I'm listening... but if that's the question, then the way to answer that question isn't a poll about whether people feel like normal games are balanced. Because,

empirically

, how people feel about balance is always more townsided than actual balance on this site, so people feeling like games are townsided doesn't intrinsically mean things need to change.

For what it's worth I suspect there are a lot of factors at play in why people find setups more townsided than they are. There's a cultural element where people harp on towns for playing bad more than they harp on scum for playing bad, I think. There's probably a Dunning-Krueger effect where people think that they can estimate balance because they've played lots of games, when those are really two separate skill sets. I think there's also an effect that people are more likely to notice a town stomp as townsided than they are to think of a scum stomp as being scumsided because the town is the team that ultimately has more agency in the game; ergo, after a scum stomp people will blame the town because why didn't they just catch the scum, but after a town stomp people will blame the

setup

because well the scum couldn't necessarily have played better, the town is the side that's ultimately making the decisions.

[Isis]

While we're at it something that bugs me that is actually kinda common is when scum prs have very little relationship to town prs.

I occasionally float the idea of a universe where the scum always know the entire setup (not who the cop is, but that there's a cop) while town has to figure it out. For balance purposes if you adopted that scum would need less power, their massclaim claim quality would improve a lot. Symbolically having the scum play "informed minority" in an even broader sense of the idea seems attractive to me, but it might be a poison pill in a lot of ways I don't realize. I think it gets played with a bit, due to Mafia informed and some mafia informed variants getting more and more informative.

[Something_Smart]

The other is that a few years ago (I think this is when I became listmod but I might be horribly misremembering tbh) there was a paradigm shift in reviews, because games were *still* scumsided. The change that resulted from that is that, in essence, every NRG member had to pretend that setups were a little bit more scumsided than their instincts thought... because they were, in fact, more scumsided than peoples' instincts thought. I think the NRG has pretty much internalized this at this point, which is reflected in generally fairly even winrates as of late, apparently, which I'm happy with.

This is interesting; it must have been before my time (and I joined the NRG a bit after implosion took over). I think my design sense has been edging on the townsided side of things since the very beginning (my first modded Normal was Rolestopper/Neighborizer/Neapolitan/Roleblocker/6 VT's vs Rolecop/1-shot Encryptor/Goon which seems townsided in retrospect but was nevertheless won by scum), and so I never really had to go through this shift.

But I think it's a pretty natural thing to do since it's easy to romanticize good townplay as easier than it actually is.

[skitter30]

I occasionally float the idea of a universe where the scum always know the entire setup (not who the cop is, but that there's a cop) while town has to figure it out.

I am intrigued

[implosion]

y'know, off the top of my head I can't really think of any or many normal reviews that had that (maybe there are some), but I do agree it's an idea with merit. It's perfectly implementable within the current rules of course, just tack on an informed modifier to a scum and then if desired give public setup info that such a role exists. Of course it does fundamentally change how certain roles work or would be balanced (some examples being watchers/trackers/roleblockers/any role that cares which scum makes the kill, because the scum would know if they want a scummy scum or a not scummy scum to make the kill).

In particular, if it is something that people would be interested in, it'd be quite easy to e.g. ask the NRG to make some pre-designed setups with it.

[skitter30]

I feel like town-me would want to know that scum had such knowledge
(Like i wouldnt want to /in for what i believed to be a regular normal game and actually play against scum who knew the whole setup. If i signed up for a game knowing that that would be the case i'd be fine with it)

And scum-me would enjoy the heck out of abusing that knowledge

I like this notion of letting scum be as informed as possible - i think what i dont like about the sort of setups i described above is that i think scum should fundamentally be at an information-advantage compared to town, and 'red-herring' scum roles handicap and hobble that advantage

[Isis]

I think it's something that, even though it could be a role, maybe fits better as a "setting" anyway?
Like Daytalk y/n Inherently Multitasking scum y/n, scum have Informed: "setup" y/n .
It gives you slightly more design space for vanilla cops hitting goons, that way.

I know the entire setup has rarely been used with informed, but I could swear I've seen a couple normal setups where an informed modifier told the mafia the -really important- aspects of a setup, so that comes kind of close.

But yeah, it's something that's always intrigued me too!

pedit: pedit makes my post relevant I guess.

[Datisi]

ego.

[TheGoldenParadox]

ego; i played in the game being referenced, and i'll put my thoughts here sometime, just not now.

[NoPowerOverMe]

Mafia is a swingy game. A lot can happen depending on the experience of the players on each side, luck, and what Not_Mafia had for breakfast. I can't think of any mod that has a vested interest in either side winning and their ultimate goal is for their game to be as even as possible. That being said when I modded my first game I initially thought it was townsided, I deferred to the reviewers, but mafia ended up winning pretty handily. I don't think most mini normal setups are going to be played enough times for there to be an adequate sample size.

[Nero Cain]

y was swingy not an option?

sometimes I do like they are townsided and sometimes I feel like they are scumsided.

[shellyc]

In my opinion, normal games (specifically mini-normals) have become overly townsided, often as part of the review process.

If this is a commonly held sentiment (i.e. not just me and a handful of others) then this seems to me like something that should be redressed for the sake of the normal queue. Consider this an info-gathering exercise to see if (1) there actually is a widespread perception that mini-normals are townsided and (2) as a sampler for interest for those who would be interested in playing in mini-normals that abide by all the rules of normalcy but may not be approved for balance (due to the bias towards townsided setups).

In the case that (1) is true, I think the NRG needs to take that feedback on board and work towards countering the townsided bias present. In the case that (2) has significant or moderate interest, I may run a game/a series of games in the mini theme queue that will be normal for all intents and purposes, but will be balanced by my standards rather than the current NRG standards for balance.

I almost exclusively play normals on this site so far and I agree with this, there are a lot of really townsided setups in normal queue

I'll be interested in mini themes that are normal but more scumsided than the current normal site meta

[shellyc]

If you consider that example game emblematic it sounds like your problem may be more to do with swinginess of power roles instead of balance. It's not how much town is winning but how town is winning that feels bad to you.

I played in that game and imo, the vengeful/vig gave town win on the brink of defeat

scum not knowing that they were in a bad position doesn't make for great gameplay
the setup had a load of PRs vig/vengefulBP/doctor/role watcher/Vanilla Cop had the potential of multiple clears/town killing power which together could result in town stomp

[iamausername]

Regarding Mini 2169: you said that scum killed PR's as though that's supposed to be indicative of good play. What happened was, scum killed all the useless PR's while leaving all the important ones alive. Presumably those town PR's played well in order to misdirect the scum kills; I don't know when exactly all the roles were claimed.

pretty sure the vig killed more town PRs than scum did, and the important one they actually did try to kill was bulletproof, so...

like, i don't know if that game was emblematic of normal games in general these days, but i don't think arguing from the position of "scum should've played better" is a reasonable response to what happened there.

bulletproof vengeful is not a fair role and it shouldn't have made it through the review process.

[Flavor Leaf]

There's a cultural element where people harp on towns for playing bad more than they harp on scum for playing bad, I think.

Piggy backing on this, I also feel like a lot of people unfairly push the idea town's are playing poorly from an outside perspective in dead thread or spectating without giving scum actions credit. Town start all tangled up, and it's their job to become untangled while scum want to act like they're helping untangle while really tangling it even more, and this is kind of a behind the scenes, invisible effect done by scum.

The game in question, if I am correct on what happened, didn't multiple town PR's pull off successful gambits as well? It got to the end, town just seemed to pull a fast one and scum didn't see it coming because it was a successful town gambit? I might be missing something, however. I didn't look at the game that deeply.

[iamausername]

yes, the vanilla cop pulled off the extremely successful gambit of claiming to be a full cop with a guilty on a townie

[Datisi]

i may be biased in this discussion, but i don't see a way there was to see the vengeful coming. we had a (basically) vanilla cop, we knew the player was not a VT. shot him once, then figured out he was bulletproof. with a crapton of town power already dead or claimed, what reasonable person goes on to say "tbh this VT-claiming bulletproof is probably also a vengeful"?

also, can i just point out what a

rude

combination a mafia vanilla cop and a town bulletproof is?

[northsidegal]

a lot of people have already made the points that i wanted to make with regards to balance as a desirable metric, so i won't repeat them.

i will, however, link this massive diatribe i went on in a thread very similar to this one, because i think that it addresses a lot of the same points being brought up here e.g. 28 "red herrings" in setup design and skew as a potential reason for it feeling like town has more agency

i will say though that i still dislike overcomplicated setups, and that if it were overwhelmingly shown that people preferred a meta where the NRG balanced such that towns won less often but people felt like games were more fair, i might agree that that metric would be more important to strive towards than balance. absent that though, i still think roughly 50/50 winrates is a fairly good goal