Theoretical Game Setup Discussion

[Seol]

Was this crafty, or bastard-mod? I consider it crafty, but it's arguable that I was lying to the nonpowered players on death revelation. I don't consider it lying (the cop

was

a cop, but that was flavour not functionality and I didn't specify) but it's pretty borderline.

edit: also, doesn't open discussion of closed setups render them unusable?

[Seol]

I'd say that missing out crucial info in the role reveal in an active attempt to mislead players is probly bastard modding. As I say, though, I don't really consider bastard modding an inherently bad thing.

The rule of thumb I've always applied is that if the setup allows a reasonable chance of the town uncovering and understanding the true situation, as opposed to the immediately apparent one, it's fair. There's a difference between the fair/unfair divide and the crafty/bastard divide, but for example an insane (reverse) cop is IMO unfair in a mini but fair (albeit still inadvisable) in a 30-man game, due to the latter allowing enough results that there's a decent chance the cop can derive their own sanity.

Personally I think the fair/unfair divide is the important one and anything unfair is unreasonable in the vast majority of contexts.

[Seol]

Everyone shoots the person below them in the first post N1/N0. One vigilante should live, for a town auto-win.

Not if it's closed setup, they won't.

[Seol]

RE: Crafty V. Bastard.

Calvin and Hobbes mafia was an inadvertent mountainous. I would say that this is a crafty setup. However, there was a no claim rule, and limited reveal. I'd say this pushes the bar to the town having little chance to figure it out, leaving the game in

bastard mod

territory.

Well, one thing about something being a bastard setup is that if you don't guess what the setup is, it has to hurt you some way. How does not realising the game is mountainous hurt you?

[Seol]

Well, my point being that you shouldn't be making assumptions like that about the setup - there's no implied contract there are power roles in a game not advertised as being mountainous. It'd be interesting to have more closed mountainous games or even games where the town have nothing but drawback roles being run, but the metagame supports the assumptions that the town will have at least one investigative role, at least one protective role, and no duplicated roles as overwhelmingly safe assumptions in minis.

[Seol]

It'd be interesting to have more closed mountainous games or even games where the town have nothing but drawback roles being run, but the metagame supports the assumptions that the town will have at least one investigative role, at least one protective role, and no duplicated roles as overwhelmingly safe assumptions in minis.

And the metagame gets worse if only

certain

mods are bucking the system...

Do you think those

should

be safe assumptions?

I quite agree however that if the intention is to decrease the predictability of setups, then if only a subset of mods act it doesn't address the issue on a global basis. That doesn't mean that subset of mods are acting inappropriately in running a greater variance of setups, especially if they have a reputation for doing so, as long as they can run them well. For them to run setups which are predictable in their variance (eg running lots of stealth vanilla setups) is possibly counterproductive though.

Trouble is, nonstandard setups are harder to do right than standard setups (obviously, the standard setups are proven) and thus run a greater risk of being broken. So there's two sides to the story there.

[Seol]

I think you misunderstand me. I

am

one of the mods who doesn't subscribe to the "at least one investigative role, at least one protective role, and no duplicated roles" model - but once that gets out, players begin to metagame you as moderator, and some of the advantage (in cancelling assumptions) gets taken away...

Well, I'm one too - but yes, if you're looking to

punish

people for assumptions, then you're undermined. If you're trying to make people open-minded about the setup, it works well. The latter is the objective - the former is only desirable as a step towards achieving the latter in future.

The trick then is keeping the games fresh and unpredictable and not just running, say, three copless games in a row.

[Seol]

I do agree that coptype roles need to be less ubiquitous than they currently are, but I think simply mixing up what coptypes you get helps to break the meta of follow the cop to some extent. However, I do think coptypes have a role in mafia.

It's not just about coptypes, it's about being able to rely on setups in terms of assessing truth of claims. If there's always

a

coptype, then you get into claim, wait for counterclaim, if no counterclaim safe to believe it, if a counterclaim that means one is lying. Investigative roles are just one way of many to add subtlety and information to the game of Mafia, and not every game should need one. I suspect that most people include one because they think it's expected, perhaps even required - and that's not the case.

Mountainous closed games probably do therefore have a role in breaking the meta.

It doesn't have to be mountainous, either - maybe a game where the only power roles are a nontalking mason loop of 3 (A knows B is innocent, B knows C is innocent, C knows A is innocent) and a vigilante, or something like that. I ran LO2 with no investigative roles (apart from a coroner, who only works on dead guys).

[Seol]

OK, here's as good a place as any to stress-test my coming game - Heroes Smalltown.

It's a Smalltown, so it's open, and roles and alignment are separate. That means the alignments in the game are:

8 Town: You are Town. You win when there is nobody alive who is not Town and at least one Town player is left alive, even if you are dead.

3 Mafia: You are Mafia with X and Y. You may talk with them during night and night only. Each night one of you (and only one of you) may target a player and kill them. If you do, you may not use any other night or day choice your role would normally allow you to use until the following night. You win when there is nobody alive who is not Mafia and at least one Mafia player is left alive, even if you are dead.

1 SK: Despite appearances, you are Sylar. You've already eaten the brains of Candice and [X], and now you've assumed the form of [X]. Each night you may target a player and kill them and eat their brains. Doing so will mean you gain their Powers¹. Furthermore, each night cycle you can use one of your abilities as well - either that of [X] or any abilites you have gained through your kills. Note that using a nightchoice (other than your kill, which is separate) will leave you unable to use any daychoices the following day.

The roles:

Peter Petrelli, Role Absorber
Whenever you are targeted by a Power¹, you will gain that Power and will be able to use it on subsequent nights. I will inform you of what powers you have gained as the following day starts. Each night, you can use

one

power which you have learned.

Nathan Petrelli, Motivator.
You can fly, which means you can get from A to B very quickly. You can carry someone whilst you fly, which means you can help someone else get to two places in a given night. Each night, you may target a player, and they will be able to use their Power¹ twice on that night.

HRG, Jailkeeper.
You might not have a Power¹, but that doesn't mean you can't do anything. The rooms below can keep someone locked up for a night. For their own good, of course - protective custody. Each nigt you can target a player, and they will be both untargetable by other players' actions and unable to use any abilities that night (note: does not restrict players from using day abilities the following day. One-shot or sacrificial abilities are not used up).

The Haitian - Dampener
You're a walking field of pure, distilled anti-consequences-of-evolution. You can repress the gene that lets people fly, bend time or turn invisible just by being there. As such, you can't be targetted by Powers¹. Furthermore, each night you can target a player and escort them, which will make them both immune to being targetted by Powers for the night and incapable of using their own Power.

Ted Sprague, Walking Bomb
You're radioactive, man. You're a loose cannon and you know it. You've finally got your power under control, but you've also got a bit of an explosive temper. At any point during the game - day or night - you may go nuclear, target a player and kill them in the blast. Only downside is that this will also kill you.

Hiro, Timebender
You have control over the realms of time and space. Sort of. Well, you can influence them, but you're kind of... unreliable at best. It's still really impressive though! Seriously. You have two abilities - a one-shot nightskip and a one-shot dayskip. The dayskip is a nightchoice, and will result in the night it is used on being followed immediately by another night with no day in between. The nightskip is a day choice (I must be PM'd with a request to use this prior to posting the lynch scene) which will result in the day being followed immediately by another day with no night in between. Each of these abilities can be used only once during the game.

Ando, Enthusiastic Sidekick
You're Ando, the Luigi to Hiro's Mario! And you're not particularly good at anything! Yay! You have no abilities whatsoever.

Mohinder Suresh - Networker
You've been following up your late father's research, and whilst you don't have anything solid yet, you do have an irrepressible urge to discuss it with people. Each night you may target a player and invite them for discussion. From that point onwards, they will be able to talk to you at night (but not to each other).

Eden - Subliminal Influencer
You've got a seductive glance and hypnotic genes. You can make anyone do anything you want - within reason. Each night, you may target a player and name another player. If the player you targetted chose to use a Power¹, they will use that Power on the named player instead. You cannot make a player target themselves.

Matt Parkman - Tracker
You can read minds. Well, you can hear thoughts - bits of them, fractured and incomprehensible. You can't get at all the details but you can certainly get

something

. Each night you can target a player, and you should be able to at least make out who, if anyone, they are targetting that night. Note - if you target Sylar, you will get there too late to pick up anything about the kill, and only hear about who, if anyone, he targetted with his regular ability.

Claude - Watcher
You're invisible, when you want to be. A useful skill. The things people do when they don't think anyone's watching are the sort of things that you need to know, so it's definitely come in handy once or twice. Each night you can target a player and just, you know, hang around them a bit. You should be able to see who comes to visit them, even if you don't get a decent look at

what

they're doing.

DL Hawkins - Insubstantiator
You can phase out. Or become spectral. Or do something, but whatever it is, it allows you to walk through walls. Which means when you want to not be somewhere, you can not be there. At night, you may choose to not be there - this will mean you cannot be targetted by any abilities. However, you can't do this indefinitely - if you phase out one night, you cannot phase out the following night.

Linderman - Doctor
You have a unique gift, a healing touch. You can undo all sorts of harm by the merest laying on of hands. With all the carnage going on around, this will surely be of invaluable assistance. Each night, you may target a player, and if they have suffered a fatal injury, you can return them back to the land of the living. That is, as long as there's something to return - the damage that Sylar does to a corpse is too great for your ability to undo.

Micah - Doublevoter
You have a knack for technology. You can make computers do things they weren't meant to do. Which is particularly useful when computers are being used to track who's next to be executed via this bizarre mob rule. Each day, once per day only, you may PM me with an additional vote. Please note that once cast, this additional vote may not be altered or revoked.

1: Powers. Various roles refer to Powers. Not all night choices are Powers - specifically, the Mafia kill, Mohinder's recruit, HRG's jailkeeping and (for balance rather than flavour reasons) Sylar's kill. Certain roles interact with Powers as opposed to nightchoices - specifically, Sylar's brain-eating, Peter's absorption, Nathan's motivation and the Haitian's dampening.

So that means Sylar can't steal Mohinder's ability by killing him - however, if Mohinder (or HRG)

is

Sylar, then he

does

keep the ability (flavourwise - because the ability is dependent on them actually

being

Mohinder or HRG, and Sylar has already convinced the other people in the game he is them, but once it starts the dead body would be a dead giveaway it wasn't actually the real Mohinder/HRG).

More importantly, it means Sylar can't get an additional kill by NKing a Mafioso, and a motivator hitting scum doesn't mean the scum get an extra kill.

This is a 12-person game, and that's a list of 14 characters. I'm not intending to allocate roles randomly, but rather allow people to draft their roles (in a random order). This means two of the roles won't be in the game. Alignments are distributed

after

roles are drafted, so at that point it's simply who the player most wants to play as.

For prominent characters being omitted, Claire is being omitted because I feel NK-immune is too powerful and nothing else really works for her, whereas Nikki/Jessica and Isaac are being omitted for the lack of a good role for them. Ideally if I could find suitable roles for them I'd have the draft list 15 roles long.

[Seol]

Can a Motivator/SK kill twice by motivating himself? Can a Motivated Hiro skip two days (or two nights) at once?

No and no - I'm going to amend the rules to say that Sylar's power isn't considered a Power, and Hiro's abilities are both one-shot (as is Ted's).

[Seol]

So, with the drafting, everyone knows what role everyone else is right from the start? And the catch is just that any role could be scum because scum/town get determined after drafting?

I would think the Mafia would walk away with it with complete knowledge of all town roles from the beginning of the game.

Yes, that's a given - this is Smalltown after all.

I might not have been clear on that - Smalltown isn't just open setup,

everyone knows everyone else's role

from the beginning of the game. It's part of what Smalltown means. It's, like, Super-open.

[Seol]

I tend to think that every, or at least most, mini games should have some mechinism to stop the "person A claims cop, is sane cop, person B is doc and protects person A, town wins unless scum manages to find and kill person B almost instantly" madness; I do NOT think that "1 cop, 1 doc" SHOULD be the default normal mini setup, especally when there are so many easy ways of preventing that from happening.

The question is, why don't people do this? Is it because it doesn't occur to them that it's a problem, is it because they see it as a problem but aren't sure how to fix it, or is it because they've tried alternatives and prefer their games like this? Because at the end of the day, we might be in a minority here. Personally I'd like nothing more than a meta where any claim of power is met with skepticism as to whether it's a bid for survival and we gave people a chance for last words before dying instead of clamouring for a claim.

Also, some solutions other than giving the scum a roleblocker or a daykiller should be used, like eg not giving the town a cop at all. I'd be very amused to see someone try to mini a game with doc, vigilante, two masons - and tell the scum there's no cop.

[Seol]

And, yeah, I think that game you're suggesting sounds balanced. Personally, I'd say having 2-3 masons is about equal in power to what a cop should be in a mini game; if a cop gets off more then 2-3 investigations in a mini game and shares them and they're all accurate the scum basically auto-lose.

I don't think it's unbalanced in principle, I'm just unsure as to whether trying to come up with a setting designed to punish a prevailing strategy is fair on the players (obviously, it's designed to encourage the scum to claim cop, safe in the knowledge there will be no counterclaim).

[Seol]

Maybe I'm missing something, Seol, but if people get to draft their roles, what motivation is there for anyone to choose Ando? I mean, unless you're specifically trying to deny Sylar or ScumPeter a power (which doesn't seem like it would be worth doing), being vanilla doesn't really seem preferable to any of the roles.

Some people prefer being vanilla. Ando gives them that option.

Hiro is too dangerous to be allowed in the game, on any side. I'd pick Ando over Hiro.

You think so? If he's too powerful, he needs to be dealt with in the setup tweaking stage. What's the worst that he can do?

As town, he can give the town an additional lynch - dayskipping is only very rarely good for the town and generally only when they've pretty much won already. As scum, he can function as a one-shot extra kill that risks being messed with by other abilities that are out there - it is only very rarely good for the scum to skip a night. Both of those are good but not devastating, and everyone knows who he is so he can be killed if necessary. If he uses either early-game he's answerable to the town for it, if he waits until lategame he may well miss out on using them altogether. Am I missing something?

fine then, auto /in for Seol's small town, and I have dibs on Ando.

Nobody gets dibs, the drafting order will be randomised.

Also, I'm fairly sure Nikki/Jessica should get a double vote.

Or Mica gets a one-shot "rig the voting machines" Power.

Good call. Flavourwise I prefer Micah for the invokable double-vote. Generally I prefer night actions but as long as it's not passive (so people are answerable for how they use their powers) job's a good 'un.

Soel- Have you considered allowing Powers to stack and combo in unexpected ways unknown to the players?

No, the whole point of Smalltown is the setup is totally open. No random factors, no "mod notes", everything there is to know about the setup is known by the players. Hence my reluctance to use Isaac as some sort of "difficult to interpret investigator".

btw it's Seol, not Soel.

[Seol]

My consideration was Sylar, who can just steal the ability and use it to jump himself into a win/decent endgame without anyone even knowing who actually used it. Though that's probably not as bad as I think it is.

Well, the town will know Sylar has that ability (kill MOs are revealed) so they'll be taking into account the possibility of him getting an extra kill in. Of course, if there are any Mafia still about, this is a hella risky tactic.

[Seol]

looking forward to this! Just checking - you've got me as /auto-in too, right? Getting to actually play a smalltown's something I've been looking forward to - and will be useful in designing more in the future.

I'm being very restrictive with auto-/ins, and not taking them at the moment except for special cases (eg you, obv, since you've modded all prior Smalltowns, and a couple of others I won't mention). The game's probably a couple of months off anyway.

[Seol]

Mmm. A lot of people want pre-/ins. Maybe I should point out that as a completely open game there's no reason why the game can't be run more than once?

PARALLEL UNIVERSE SMALLTOWN HEROES perhaps?

[Seol]

It was a joke...don't worry. I can wait.

Also, if you don't mind I want to nom it for the Open queue, once you've run it of course.

Feel free - although I'm sure the game could use tweaking after the first run, I'm pretty sure the core concepts are repeatable. Although, for the Open queue, we might be better off designing a "vanilla" Smalltown game (in the style of the original).

EDIT: Inspired by the "Why start games at night?" thread, I give you - Poisoner Mafia.

Scum headstart, but scum kills are poisonings. Poisoned players learn they are poisoned as day starts, and if they are not doc-protected the following night they will die.

[Seol]

NOOOO! The trick to smalltown (and this is why it is so damn hard) is to make it balancable without making the roles suck. The first smalltown was well balanced (IIRC) but the roles were boring Watcher, Tracker, etc.

The later smalltowns had some FANTASTIC roles, but unfortunatly allowed an unlucky roll to unbalance them (e.g. SK/Mafia gets vig as well. )

And lessons have been learned from that, for example in Heroes there is no vig and certain abilities (eg Peter's role-copying and Nathan's motivation) can't lead to doubling-up of kills. When I say we should come up with a vanilla Smalltown setup, then I mean one where the lessons learned are implemented.

[Seol]

You could use Niki/Jessica, who has a 50% chance of being retaliator and 50% chance of being vanilla town at the beginning of the game. Could be a nice exception to the Smalltown rule.

You could just make her claim D1. Besides, I like it being so "pure" open. It's my favourite thing about the setup.

Agreed. No random factors, no hidden information. Again thanks for the idea, but it's not quite what I'm looking for, and retaliator (I presume you mean bomb?) has issues in Smalltown.

Some season two characters could be use to make the list a little longer.

Yeah, but then I'll be lost, confused, and sad.

Why do you want to make me sad?

Season one only, partly for spoiler reasons and partly so I can draw a line and make the roles. I don't want to be continually adding and editing the setup as new Heroes are revealed. Besides I think the optimal list length is 15 anyway, so I'm only one off.

That's

exactly

the trick to it; to come up with 12 different medium-strength powers (which ideally should all be choice-powers, not automatic, and ideally all used at night). It's hard.
Btw, Smalltown 2 (scrubs) was actually better balanced than the original - at least, I think so. Only 3 players survived the game (I've always thought, barring particularly bad play on either side, the fewer players survive a game, the better balanced it probably was).

Kind of a narrow way of judging it, to be honest - there was potential for the game to be massively swingy. I don't agree they should

all

be optional night abilities - too much night analysis leaves you prone to redux situations - but they should at least mostly be to a greater or lesser extent under the control of the player.

[Seol]

Besides I think the optimal list length is 15 anyway, so I'm only one off.

You could use Wireless as message sender, but I'm not sure that'd be a good Smalltown role.

Which has occurred to me, but it's too similar to (and significantly less powerful than) Mohinder.

If there was a natural fit, I'd like to slip an inventor in there, but I can't see how or where.

[Seol]

How does dayless

work?

Eg, how does the town kill? How do cop results feed back?

[Seol]

Don't know. You'd have to think up clever new roles; give everyone something useful to do.

I'm pretty sure I could get

lynchless

Mafia working neatly, but dayless is asking a lot.

[Seol]

Okay, I have a question about bastardity.

Lets say you make a "movie Mafia mini". The players are the names of characters, except for 3-5 which are the names of movies. This will cause a "characters = town, movies = scum" idea to form

Only if the town are stupid. If they mass-name-claimed off the bat, then yes, there's a distinction, but not necessarily implying scum. There should never be a mass nameclaim off the bat though, and you'll have dead town movies and/or dead scum characters by the time it's appropriate.

and they will probably lynch wrong on it.

The moment someone says "hang on, movies aren't scum, I'm a town movie" that idea should be kiboshed. Any town that comes to such an arbitrary conclusion and then follows it through despite objections is idiotic. Period.

[Seol]

Bridge Roles Mafia.

Central conceit: there are three alignments. One's Town, one's Mafia A, one's Mafia B. 24 players, 3 are Mafia A, 3 are Mafia B. 13 are town. The rest are bridge roles.

Bridge roles share two win conditions. For example, there are two players aligned with

both

Mafia A and the town. The town doesn't need to kill them to win, nor does the Mafia A. The bridge roles don't know who's in Mafia A, but they have abilities. They might be "masons", for example. Mafia B has similar roles. They aren't survivors, they're simply dual aligned.

The "serial killer" here is aligned with both Mafia A and Mafia B - as long as all townies are dead (

not

including bridge roles) he wins. He

is

a survivor role - has to be alive to win - but has a win and remove condition.

[Seol]

Mafia: Endgame

Exactly what it sounds like. A 7-player game constructed to look like the end of a 20-player game, with various roles having four nights worth of results and a list of nightkills - and 13 dead - to give clues.

[Seol]

I personally am trying to make my mini's unpredictable. However, my 1st normal was an attempt to break the meta. Unfortunatley people knew I was doing that and wen't "inadvertent mountainous...yeah sure"

Heh, if people go into a game knowing it's an attempt to break the meta, I'd run a 2 goon, gf, cop, doc, roleblocker, 6 townies game, and watch them drive themselves nuts looking for the twist.

[Seol]

Why?

[Seol]

I personally am trying to make my mini's unpredictable. However, my 1st normal was an attempt to break the meta. Unfortunatley people knew I was doing that and wen't "inadvertent mountainous...yeah sure"

Heh, if people go into a game knowing it's an attempt to break the meta, I'd run a 2 goon, gf, cop, doc, roleblocker, 6 townies game, and watch them drive themselves nuts looking for the twist.

Indeed. If you're trying to make your games unpredictable, then

consistently

breaking the meta is counter-productive.

No... if you're trying to make your games unpredictable, you have to do so

in different, unpredicatable ways

(inadvertent mountainous is one of the great cliches of cliche-breaking). If you're looking to break the meta, though (which is subtly different), then you're no going to do that with a reputation as someone who acts outside the meta.

Also, games shouldn't be announced as "meta-breaking", they should just

be

meta-breaking.

Depends on your purpose. I like being known as someone who has twists in their games. It appeals to the sort of players who are more likely to appreciate it.

[Seol]

Ah, I was thinking in relation to the long term "breaking the meta", rather than just "making games unpredictable".

Changing your own games isn't enough, you need to influence others to do the same. Partly by arguing the cause in MD, and partly by running really interesting and fun games which are outside the meta - but making them interesting and fun is more important than being "meta-breaking", otherwise it's hardly a persuasive argument.

Most specifically, what's

not

a good idea is running mechanics which are designed to punish people for expecting the meta. It's not fun to lose a game because the mod was trying to prove a point.

[Seol]

Points Mafia.

Vanilla game, daystart. Each side starts off having 20 points, split unevenly between their players. So the roles are something like:

Mafia worth 9, 7, and 4.
Town worth 6, 4, 3, 2, 1, 1, 1, 1, 1.

If, in the morning, one side is 9 points ahead of the other, that side wins then and there. A points win is considered a superior win to an exhaustion win.

Not done any work on balance yet, it's pure concept, so those numbers may well be busted.

[Seol]

Alternative idea: Give power roles the least points.

That's the obvious extension of the idea, yeah.

[Seol]

Good idea but I object to one win being superior to another unless you can put numbers on how much superior; if I don't know what expected value to maximize I don't know what game I'm playing.

What's the difference between a victory by KO and a victory by points in boxing?

[Seol]

I don't know anything about boxing but if it has badly-defined goals that's bad, but maybe not as bad as in mafia which is more analysis-intensive.

Yeah, I agree with you that if there's a distinction, that needs to be quantified. Before I quantify that, I'd want to get a better idea of balancing the setup (which I'd probably approach with a bunch of random simulations with different points values), but for when I get that far, boxing seems like a good starting point. The question was neither rhetorical nor aimed specifically at you.