Calling All Mafia Players

[TinCow]

I'll limit my inquires to the basic balance issue to start with. I'll start with the random assumption, based off of Beefy's post, that 1:9 is a good mafia:town ratio in a 30+ player game. So, in a 30 player game there will be 3 mafioso and 27 townies, with the mafia getting one kill per person. Assuming that is balanced, what then happens if we add on a single detective role? Does that skew the balance enough towards the town to require compensation on the mafia side? How much compensation? How about a doctor?

Essentially, I'd like to try and figure out a basic blueprint for a standard, balanced large game. Working from a balanced mafia and townie-only game with no roles, we could then 'weight' all other common roles (i.e. detective, doctor, serial killer, survivor, etc.) according to how much they shift the balance towards the town or the mafia. For an example, I will randomly assign the following values to roles, with + numbers benefiting the town and - numbers benefiting the mafia.

Detective +2
Doctor +1
Mafioso -4
Serial Killer -1

If we assume that the 'balanced' ideal has a starting value of 0, then adding in a single doctor would unbalance the game by 1 point in the town's favor. That could then be corrected by adding in a SK to restore the score to 0. If 2 detectives were added in, a bonus mafioso could be added to balance out the detectives.

Again, the above are just randomly assigned numbers with no basis in reality... I just made them to illustrate the system I'm thinking of. If we could arrive at some kind of consensus about the proper base mafia:town ratio, and the proper weighting of basic roles, that could provide a solid foundation for people to build their games off of. Of course it is pointless to try and account for every possible role and rule foible, but at least with a system like this people would know that if they added something on one side, they would need to then add something else back to the other side.

Thoughts?