[OLD] Open Setup Ideas and Discussion

[T-Bone]

The rolecop and watcher essentially preform the same function because both can be used to find the players with power roles. Since its an open set-up, the watcher can't fakeclaim. I think a framer makes the set-up more interesting, and there is a lack of set-ups featuring a framer. Maybe...maybe you give the town one more weak role for balance.

[Plessiez]

Hmm, remember when Fire & Ice had no crosskills? And how that was ridiculously scum sided? Well, what if you combined Fire & Ice with Double Day?

[...]

Balanced, y/n?

I think this is still significantly scum-sided? Obviously less so than if it weren't double day, but I don't think that that change is enough to make it balanced.

If the doctor were just another VT then I make it that the random lynches/kills EV is just under 19% [1]. It's hard to believe that bringing in a doctor would help improve this figure all

that

much -- and even if it turns out it does, this just means the game is now really swingy, no?

--
[1] I can post some lines of Scheme to back this up if you really want, but you probably don't...

[Plessiez]

I need thoughts on whether scum should have daytalk in We Need A Fifth. I gave them daytalk when I ran it, but I don't know what would be more balanced.

I'm not sure it has that much of an impact? The only time I can see a lack of daytalk being significant is if a goon is lynched day 1 and can't work out which of the two townies his partner would rather have as confirmed town. But I'd guess that in most actual games you'd just want to confirm the town player who was least suspected by the other, which is something which should be obvious to both mafia players without the need for communication.

(By the way, the mafia role PM in the wiki needs to be fixed, whatever you decide: right now it says the goons may communicate "during the night phase".)

Edit: Is giving additional shots overkill?

I like the fact that losing a PR gets you additional shots? Seems to compensate for potential swing quite nicely.

Rather than introducing a godfather/framer, could you try increasing the initial Y-value (or decreasing the initial X-value)? Or does that not boost the scum enough?

[Alduskkel]

If the doctor were just another VT then I make it that the random lynches/kills EV is just under 19%

Shit, really? Damn.

[Kaiveran]

Not going to be here for the rest of the night, but before I leave I'm gonna drop a crazy idea I came up with yesterday.

C18++

[REDACTED]

Pretty much exactly what it says on the tin. I wanted there to be a generative Multiball setup but was convinced it was impossible until I had a late-night epiphany after waking up from a Mafia-related dream.

Unlike C9++, it is a Modular Open setup, particularly with respect to the town roles.

I can identify two main issues at present:

1. No considerations for inter-scum balance. There could be the issue of little to no power rolled to one team, and lots rolled to the other. While both scumteams will be balanced with the town since the power roles only work against one faction at a time, it's possible that one scumteam will be much more powerful and curbstomp the other. I have a couple preliminary ideas. One is creating power tiers for the scum, where if one scumteam ends up on a higher tier than the other, the less powerful team gets (a) helpful role module(s) depending on the difference. Another is biasing the selection of Traitors somehow.

2. Multiball games are not known for their stability, but this setup seems particularly swingy. Just look at that second game and its absolutely loaded town (particularly note the OP Mason pair of a Tracker and Seer). One of the more important roles or combo roles dying can be a huge hit to the hunting efforts against a particular scumteam. It seems that the best way to alleviate this is to add more townies and expand the setup, but I'm not exactly sure how many to add and how to balance the game for the increased number of townies. I basically copied most things whole sale from other generative setups, so any input on that front from people more experienced designing these kinds of setups would be appreciated.

So leave comments on the doc regarding any suggestions, concerns, or confusions you may have and I'll address them in-thread in the coming days.

(EDIT: Back and going to sleep...just realized on a check that

I forgot to allow comments on this document, that's fixed now.

)

[pieguyn]

Bulletproof's Mafia

2 - Mafia Goons
1 - 1-shot Bulletproof SK
5 - 1-shot Bulletproof Townies
1 - Town Doctor (protects from 1 NK)

BPs get notified when hit

i stole the idea from another site. i have no idea if this is balanced

[JasonWazza]

Seems townsided from a glance.

If the SK is lynched mafia is fucked.

If the Mafia is lynched then killed by the SK then the SK is likely fucked.

Regardless of both of the above, it tends to have no kills constantly (unless mafia and SK both hit the same person they won't die not including the doctor) and looks like a mountainous.

Given randomness, probably town sided, given claim-ability Probably extremely town-sided.

Lynch BT day 1
Doctor does nothing Night 1
Both kills hit a different BT (if one hits the SK he is fucked, the only chance he has is being avoided till late game)
Doctor claims with BT's Day 2
3 conf. town, 8 alive

Yeah probably really townsided, fact is unless something really screwbally happens, and each night the Mafia and SK manage to target the same BT (without the doctor targeting this same BT), it's townsided, if SK and Mafia do manage to hit the same target each night, then it basically becomes a 6v3 and town is fucked.

(sorry for how stupid this probably looks, i did it as i thought it up.)

[GuyInFreezer]

Bulletproof's Mafia

2 - Mafia Goons
1 - 1-shot Bulletproof SK
5 - 1-shot Bulletproof Townies
1 - Town Doctor (protects from 1 NK)

BPs get notified when hit

i stole the idea from another site. i have no idea if this is balanced

There's a similar setup in EpicMafia.
1 goon, 4 1SBP, 1 SK, 1 doc, night start. Has 50% Town win, 25% Mafia win, and 25% SK win if you want stats.

[Plessiez]

If the doctor were just another VT then I make it that the random lynches/kills EV is just under 19%

Shit, really? Damn.

Actually, no, that's wrong (although only -- I think -- by a couple of percentage points).

In particular, I was assuming town always used both lynches, which isn't actually always best (or at least, is not always best if there is only one scum team left). I've attempted to produce some more accurate numbers, which are spoiler-tagged below together with some brief comments. Although I'd note that these numbers are probably off too, because I'm not thinking about ways the scum can be clever.

Spoiler:

3P: 2 - 1 - 0
Town wins 33.3% of games

(there's only one chance to lynch)

4P: 3 - 1 - 0
Town wins 50% of games

(there are now two lynches before nightfall)

4P: 2 - 1 - 1
Town wins 16.7% of games

(there are six possible pairs of players to lynch, only one of these pairs results in a town win)

5P: 4 - 1 - 0
Town wins 50% of games

(the best strategy is to no-lynch and go to 3-1-0 -- without this trick the chance of town winning would only be 40%)

5P: 3 - 2 - 0
Town wins 13.3% of games

(town have to lynch right the first time - 2/5 chance of this. Then it's best to not-lynch, to get to 2-1-0 which you'll win 1/3 of the time)

5P: 3 - 1 - 1
Town wins 13.3% of games

(from the town's perspective, this is exactly the same as 3-2-0 once you assume random lynching)

6P: 5 - 1 - 0
Town wins 58.3% of games

6P: 4 - 2 - 0
Town wins 25.6% of games

6P: 4 - 1 - 1
Town wins 27.0% of games

6P: 3 - 2 - 1
Town wins 8.1% of games

6P: 2 - 2 - 2
Town wins 1.0% of games

7P: 6 - 1 - 0
Town wins 64.3% of games

7P: 5 - 2 - 0
Town wins 28.6% of games

7P: 5 - 1 - 1
Town wins 35.3% of games

7P: 4 - 2 - 1
Town wins 12.5% of gaes

7P: 3 - 2 - 2
Town wins 2.7% of games

8p: 7 - 1 - 0
Town wins 64.3% of games

8P: 6 - 2 - 0
Town wins 33.8% of games

8P: 6 - 1 - 1
Town wins 40.4% of games

8P: 5 - 2 - 1
Town wins 20.2% of games

8P: 4 - 2 - 2
Town wins 8.6% of games

9P: 8 - 1 - 0
Town wins 68.3% of games

9P: 7 - 2 - 0
Town wins 40.5% of games

9P: 7 - 1 - 1
Town wins 40.1% of games

(this is slightly worse than 6-1-1, but I don't think there's a clever way for town to no-lynch into that set-up)

9P: 6 - 2 - 1
Town wins 20.1% of games

9P: 5 - 2 - 2
Town wins 8.0% of games

10P: 9 - 1 - 0
Town wins 71.4% of games

10P: 8 - 2 - 0
Town wins 42.9% of games

10P: 8 - 1 - 1
Town wins 47.0% of games

10P: 7 - 2 - 1
Town wins 27.0% of games

10P: 6 - 2 - 2
Town wins 14.5% of games

11P: 9 - 2 - 0
Town wins 45.6% of games

11P: 9 - 1 - 1
Town wins 50.3% of games

11P: 8 - 2 - 1
Town wins 30.6% of games

11P: 7 - 2 - 2
Town wins 17.0% of games

12P: 9 - 2 - 1
Town wins 35.0% of games

12P: 8 - 2 - 2
Town wins 22.1% of games

13P: 9 - 2 - 2
Town wins 21.2% of games

(and this is the set-up we care about)

(These numbers are still mostly computer-generated, so I can't give full EV project style details.)

Anyway, the upshot is that I

think

that a more accurate EV here is (a very slightly better) 21.2%. Again, that's with 9 VTs and no doctor.

But I'm pretty sure the general claim I made is still right -- if these numbers are even close to accurate, the set-up with a doctor seems unlikely to be balanced. (That's almost certainly true if you'd like town to be winning this about half the time, and is probably true even if you only want town to win a third of the time. I'm not sure which of these targets is what you'd want to meet to call the game 'balanced', actually. Possibly somewhere between the two?)

[Plessiez]

Lynch BT day 1
Doctor does nothing Night 1
Both kills hit a different BT (if one hits the SK he is fucked, the only chance he has is being avoided till late game)
Doctor claims with BT's Day 2

Given that you've lynched a BT on day 1, there's a pretty decent chance that the doctor is going to die on night 1, no? [There's a 5/6 chance the mafia don't shoot the doctor and a 6/7 chance the SK doesn't shoot the doctor, which means there's a 2/7 chance that at least one of them does shoot the doctor.] Town's chances must (presumably?) drop quite a bit when this happens [you go from three people claiming on day 2 to just one claiming].

Of course, that said, I agree this set-up looks town-sided. First instinct is to suggest that you don't inform the BPs when they've been hit, maybe?

[Alduskkel]

Anyway, the upshot is that I

think

that a more accurate EV here is (a very slightly better) 21.2%. Again, that's with 9 VTs and no doctor.

But I'm pretty sure the general claim I made is still right -- if these numbers are even close to accurate, the set-up with a doctor seems unlikely to be balanced. (That's almost certainly true if you'd like town to be winning this about half the time, and is probably true even if you only want town to win a third of the time. I'm not sure which of these targets is what you'd want to meet to call the game 'balanced', actually. Possibly somewhere between the two?)

I'm okay with the Town winning only 30% of all games... but at this point the only solution is to add more PRs which will increase swing. Not really sure how to salvage the setup. Faction Cops are the obvious but shitty solution.

[Plessiez]

... I should

really

stop trying to think today. I'm pretty sure those last numbers I gave are

also

wrong ... this time because I forgot that fire and ice aren't meant to have cross-kills . (Technically, I didn't allow them to both kill each other at night, but I was allowing scnearios where one faction killed a townie and the other faction cross-killed.)

Unsurprisingly the numbers once you fix that are much worse for town (I think we're down to a mere 15% chance of a town win in the initial 9-2-2 set-up). But rather than make another long post which turns out to be wrong, I think I'll come back to this in a day or two when I can hopefully think straight.

Not really sure how to salvage the setup. Faction Cops are the obvious but shitty solution.

One idea I had (and in fact, this is what I was going to check when I noticed the embarrassing error above) might be to say that, although the scum can't cross-kill, any attempt by either scum faction to shoot scum causes all scum kills that night to fail? Not sure how much that would help though -- suspect it's not enough.

[JasonWazza]

Bulletproof's Mafia

2 - Mafia Goons
1 - 1-shot Bulletproof SK
5 - 1-shot Bulletproof Townies
1 - Town Doctor (protects from 1 NK)

BPs get notified when hit

i stole the idea from another site. i have no idea if this is balanced

There's a similar setup in EpicMafia.
1 goon, 4 1SBP, 1 SK, 1 doc, night start. Has 50% Town win, 25% Mafia win, and 25% SK win if you want stats.

Epic Mafia =/= here

They don't play optimally.

Also Mafia and SK can't win together on proper forums in most cases.

Lynch BT day 1
Doctor does nothing Night 1
Both kills hit a different BT (if one hits the SK he is fucked, the only chance he has is being avoided till late game)
Doctor claims with BT's Day 2

Given that you've lynched a BT on day 1, there's a pretty decent chance that the doctor is going to die on night 1, no? [There's a 5/6 chance the mafia don't shoot the doctor and a 6/7 chance the SK doesn't shoot the doctor, which means there's a 2/7 chance that at least one of them does shoot the doctor.] Town's chances must (presumably?) drop quite a bit when this happens [you go from three people claiming on day 2 to just one claiming].

Of course, that said, I agree this set-up looks town-sided. First instinct is to suggest that you don't inform the BPs when they've been hit, maybe?

2/7 ain't that good a chance, i agree win rate drops slightly when that happens, Mostly because you then get Max 1 Conf. Town and that conf. town always dies.

Basically we come to a setup that is completely town-sided unless one thing happens;

Mafia makes a big ass gambit, that is likely to make or break the game.

[pieguyn]

There's a similar setup in EpicMafia.
1 goon, 4 1SBP, 1 SK, 1 doc, night start. Has 50% Town win, 25% Mafia win, and 25% SK win if you want stats.

WTF
that's the exact same setup as the one I stole the idea from. i guess that's where the guy i stole it from got it > <

not notifying BPs when hit would greatly reduce claimability. sure it's swingy as fuck but that's true about any multiball setup (and in this case a 9p setup with an SK will probably run into similar problems no matter what)

[JasonWazza]

I like how you took out the win rate that i was about to comment on

I think dropping claimability is very important to make it less swingy, when deaths are going to be avoided so often.

[pieguyn]

yeah sry for confusion, I think not notifying when hit is a good idea. it prevents any way of clearing people via being hit. although mafia could also do a no kill gambit and claim they got hit

yeah I got really bored so I wrote a simulator for the setup but I don't trust I did it right 0.0 I'd rather do an actual EV calculation before putting nay results up

[JasonWazza]

yeah sry for confusion, I think not notifying when hit is a good idea. it prevents any way of clearing people via being hit. although mafia could also do a no kill gambit and claim they got hit

yeah I got really bored so I wrote a simulator for the setup but I don't trust I did it right 0.0 I'd rather do an actual EV calculation before putting nay results up

No kill gambit means they have to kill themselves and is a terrible gambit.

[Loranthaceae]

Spectator Duel Mafia

Normal setup + 2 spectators dueling in the art of alignment-guessing the players. The spectator who has the most correct reads for the longest amount of time wins. Every 48 hours they submit a list of players they have an alignment-read on. Reads are evaluated in a stochastic manner based on the alignment probability of the setup (if there are 12 players with a 3 player scumteam the spectator will get 0,25 points (3/12) for a correct townread and 0,75 for a correct scumread. 0,75 for an incorrect townread and 0,25 points for an incorrect scumread will be subtracted)
The read lists are posted in the thread by the mod.

Probably needs 1 more mafia than normal to balance it out.

[Loranthaceae]

Or the scumteam get to rig a read or two from the list each time, which would require daytalk. That would be a more elegant solution. And more fun for the spectator since he gets hints based on what reads have been pre-dictated (not taken into account when evaluating points)

[TheButtonmen]

That's not an open setup, that's just two people reading along.

[Loranthaceae]

..posting their reads every 2 days in an unbiased fashion. It should influence the gameplay in favor of town therefore I'd like some feedback.

[Leafsnail]

Re: Popcorn mafia:
http://wiki.mafiascum.net/index.php?title=Popcorn_Mafia

You win when all threats to the town have been eliminated.

You win when all members of the town have been eliminated or nothing can prevent this from occurring.

Under a strict reading of these win conditions the mafia cannot possibly win because they have no way to kill the last townie (barring a modkill). This should probably be fixed.

Also, I was thinking of running a version where the mafia get to choose the next gunbearer after they're shot (kindof like a vengekill). Has anyone ever run a setup like that, or alternatively, does anyone have any tips for calculating win probabilities more easily for such a game? My initial instinct is that something like 8 townies vs 3 goons would be roughly balanced, but I can't say I'm entirely sure and I'm not sure if there's some shortcut I could use rather than crunching every scenario.

[LlamaFluff]

Giving mafia gun would just end up with them giving it right back to whoever shot unless you are also going to modify rules to allow scum to be gun bearer

[JasonWazza]

Also, I was thinking of running a version where the mafia get to choose the next gunbearer after they're shot (kindof like a vengekill).

Why would they give it to anyone but who shot them?

Fact is giving someone the gun give's them conf. town status, so whoever had the gun is conf. town, and now whoever has the gun is also conf. town.

[Leafsnail]

Giving mafia gun would just end up with them giving it right back to whoever shot unless you are also going to modify rules to allow scum to be gun bearer

Oh, I guess I didn't clarify: the gunbearer always dies after they shoot. So townies and mafia are effectively in the same exact boat (when scum have the gun I'd probably close the thread though, since there's no real point in pretending they're town when they're already doomed). That's why the setup would need some additional townies.