Bad Guy %

[bobbyplump]

One time in a game someone was talking about the ideal good guy-bad guy ratio, but I can't find it.

Obviously it depends on specials and such, but what's the rule of thumb? 1 in 4? 1 in 3?

[Thesp]

Traditionally, it varies within that range - 25%-33%. Abilities have a significant impact on this number.

[Seol]

Also,

it depends on the size of the game

. The bigger the game, the more impact a large single group will have (ie, a three-man Mafia in a 12 person game with a light smattering of roles is about right, whereas a six-man Mafia in a 24-person game with similar role distribution is very, very strong). If there is a rule of thumb, it's not a simple proportion, but some geometric formula which reduces the appropriate proportion with the size of the game.

[mikeburnfire]

Agreed.

Most minis are 3:9, but I've never seen a mini that's 2:10 or 4:8. If the 4:8 one did happen, the town would have to have a lot of extra powers to comphensate.

[Mr Stoofer]

2:10 is pretty much agreed to be right if the Town has no power roles at all. With 3:9 the Town definitely needs powers. With 4:8 the Town will need LOTS of power. But as Seol says, you cannot extrapolate that linearly.

All of this assumes 1 scum group and no SK.

[Seol]

With 4:8 the Town will need LOTS of power. But as Seol says, you cannot extrapolate that linearly.

Well, that depends on the structure of that 4: if it's 3:1:8 (Mafia, SK, town) or 2:2:8 (two Mafias, town) then that's probably fine in many setups. Total number of scum isn't the only consideration. But yes, a 4:8 with a single Mafia will require power to the point of being unbalancing,

or

a rules change (eg Nightless).

[Thestatusquo]

2:10 is pretty much agreed to be right if the Town has no power roles at all. With 3:9 the Town definitely needs powers. With 4:8 the Town will need LOTS of power. But as Seol says, you cannot extrapolate that linearly.

All of this assumes 1 scum group and no SK.

I agree with

Yosarian2

Mr. Stoofer.

I have always argued that the "25-33%" thing was a stupid non actionable way to advise how many scum should be in a game. Instead, for balance you should start out at no power roles and a ratio of 1:5 scum to town, then you should increase each factions power proportionally.

[Seol]

Instead, for balance you should start out at no power roles and a ratio of 1:5 scum to town, then you should increase each factions power proportionally.

That's also a bad rule of thumb for inexperienced mods. It works for you, because you know roughly what's balanced and what isn't, but how many power roles match one additional scum? Before you mention one, I will pre-emptively state that pointing systems are horrible and work only to the extent that you're only going small distances from balanced setups anyway.

[Mr. Flay]

if it's 3:1:8 (Mafia, SK, town) or 2:2:8 (two Mafias, town) then that's probably fine in many setups.

I agree with Seol.

[Mr Stoofer]

Points systems fail due to, amongst other things, the fact that power roles interact. And Stoofer's 4th Law.

[Shanba]

The funny thing is that things that have nothing to do with how much power the scum or town have can also affect balance. For example, hidden mountainous is balanced further against town than regular mountainous.

[Thestatusquo]

Instead, for balance you should start out at no power roles and a ratio of 1:5 scum to town, then you should increase each factions power proportionally.

That's also a bad rule of thumb for inexperienced mods. It works for you, because you know roughly what's balanced and what isn't, but how many power roles match one additional scum? Before you mention one, I will pre-emptively state that pointing systems are horrible and work only to the extent that you're only going small distances from balanced setups anyway.

This is correct, but I would argue that experience is a prerequisite to modding in general, and is thus not a problem with my advocacy itself.

By experience I don't necessarily mean "modding experience" but at least a large amount of playing experience. Note my suggestion to make first time mods have to complete 20 games played.

[Seol]

By experience I don't necessarily mean "modding experience" but at least a large amount of playing experience. Note my suggestion to make first time mods have to complete 20 games played.

When I said inexperienced there, I meant in respect of modding - or at least without sufficient experience to recognise a balanced setup. Of course, once you can recognise setup issues, you don't need a system any more.

Also, a 20 game requirement for a small 12-man normal game with experienced sign-off and normal roles is excessive. For an Open, even more so.

[PokerFace]

Points systems fail due to, amongst other things, the fact that power roles interact. And Stoofer's 4th Law.

Vote: Stoofer

LAL there is no third or fourth law in the wiki!

[Mr Stoofer]

That's because the wiki crashed and I couldn't be bothered to re-write it. I actually meant Stoofer's 3rd law: http://www.mafiascum.net/forum/viewtopic.php?t=4875

[Adel]

remind me to post in this thread when a certain game is over.

[Yosarian2]

Also,

it depends on the size of the game

. The bigger the game, the more impact a large single group will have (ie, a three-man Mafia in a 12 person game with a light smattering of roles is about right, whereas a six-man Mafia in a 24-person game with similar role distribution is very, very strong). If there is a rule of thumb, it's not a simple proportion, but some geometric formula which reduces the appropriate proportion with the size of the game.

I would tend to disagree with that. If there are the same % of power roles in the 12 person game and the 24 person game and the average power role is as strong, and the same % of scum, I would think the town would actually be better off in the 24 person game with 6 scum. More time for the town to figure out stuff, more chances for the scum to make mistakes, more links to look for whenever a scum is caught, and more day info in general. In a 12 person game the town has less room for error, and it's quite common for scum to win one of those games without ever losing a single member.

[Xylthixlm]

Instead, for balance you should start out at no power roles and a ratio of 1:5 scum to town, then you should increase each factions power proportionally.

That's also a bad rule of thumb for inexperienced mods. It works for you, because you know roughly what's balanced and what isn't, but how many power roles match one additional scum? Before you mention one, I will pre-emptively state that pointing systems are horrible and work only to the extent that you're only going small distances from balanced setups anyway.

As the owner of a bot which has to generate a

lot

of setups and does so using a pointing system, I will say that they can work, but only within a certain range and not all of the time. XylBot's system is based on a lot of number crunching on historical data (about 1,500 games), and seems to work fairly well within the range of game sizes that data covers (roughly 4-8 players) and the type of play that's generally seen on IRC.

[Xylthixlm]

Well... aside from an occasional tendency to go bonkers and make 2/3 of the town Vigilantes.

[Mr. Flay]

But is the game balanced even with that many Vigs? Did the Mafia get corresponding power?

I'm not sure I get why 6:18 is any worse than 3:9, but I'm willing to be convinced. I'm assuming here that the town gets double the power roles too, of course...

[Nemesis]

I'm not sure I get why 6:18 is any worse than 3:9, but I'm willing to be convinced. I'm assuming here that the town gets double the power roles too, of course...

Well, Stoofer's third law says the doubling of power-roles makes the game swing to the scum more.

As for the 6 man group, that's one hell of a massive group. There's a massive bunch of power there and on first glance it seems very powerful.

That said, a night start with no additional kills or doc protects would leave 3:8 or 6:17 which ratiowise isn't so bad.

[Mr. Flay]

Stoofer's 3rd is controversial, and doubling the number of Cops or Vigilantes or Doctors in a game hardly gives

scum

more power. Adding a NK-immune Redirecting Insane Quack is more what he's talking about, IMO...

Also, a single 6-man group has to kill at least

12

of the town to prevail, which means they have to survive unharmed/unindicted twice as long as a 3-man group in a mini (which just has to kill 6 town to reach endgame); I'm not seeing Seol's point, yet...

[Nemesis]

Stoofer's 3rd is controversial, and doubling the number of Cops or Vigilantes or Doctors in a game hardly gives scum more power. Adding a NK-immune Redirecting Insane Quack is more what he's talking about, IMO...

Controversial yes, but it makes a lot of sense.

Yeah, extreme weird complex stuff is probably what he is talking about...

But, player A claims cop, player b counterclaims... They both get lynched. Town is down 2 cops (which are part of the majority too, and possibly 2 more townies because of kills.) That's Lylo at best, isn't it (excluding doc protects)? [n0 6:18, d1 6:17, n1 6:16, d2: 6:15 n2 6:14... Ending 6:13 with a night start.]

Obviously that scenario won't happen all the time, even if the two cops did somehow counterclaim each other day 1, the town might figure out what was going on or at least figure one of them has dodgy sanity. But the scenario isn't even complex and adds a certain amount of confusion.

I've been lynched as a role cop despite claiming and showing 3 or 4 confirmable results... There were "too many cops". Granted a role cop and an actual cop are different things, and I wasn't lynched on the basis of my claim alone, but still. I expected having accurate role results would save me. So it does have

some

merit even when dealing with conventional roles.

Also, a single 6-man group has to kill at least 12 of the town to prevail, which means they have to survive unharmed/unindicted twice as long as a 3-man group in a mini

Yes, but if they want, they have 6 votes to put on someone. Assuming the day 1 lynch is based on feelings... They don't have a random feeling, they have a 6 strong feeling and it is for the same person if they want it to be. 6 people can make a bandwagon sound fairly convincing sometimes.

If you consider the number of people alive and the spread out suspicion among them, early on the scum have a mega advantage in bandwagon derails and bandwagons in general as well as having a 6 person general consensus if they want.

Later on, their percentage of players grows and the results are the universal advantages of having a high percentage of scum left.


I'm not that great with game mechanics, but a 6 person group looks insanely powerful to me. But maybe I just think that because I've never seen a 6 man group. Experience really isn't on my side here.

[Xylthixlm]

My intuition on the math is that going from 3:9 to 6:18

without

adding extra power roles doesn't change the balance significantly. The mafia get twice as many kills to kill twice as many townies, the town gets twice as many lynches to hang twice as many mafia, the cop gets twice as many inspects to search twice as many suspects, the doc gets twice as many protects to stop twice as many kills, etc.

[Rogueben]

Yes but if a 6 man scum group teams up early on people are going to look back in their analysis for scum linking. Six scum on a wagon is going to be sniffed out pretty easily I would think.