Conquer Club

Epitaph1 wrote:I just learned that Ultimate Werewolf has a system for creating balance. Each role is assigned a number (positives for townies, negatives for werewolves) and you use those to get as close to zero as possible.

https://boardgamegeek.com/thread/641590 ... calculator

I haven't played UW yet (I will tomorrow) but my understanding is that most of the roles correlate to what we have in mafia. Perhaps we can adopt the number system to mafia roles and try a few setups based on that.

I don't think this will work very well for mafia. The combinations of roles make for very non-linear algorithms for determining balance. Even if it could work, you'd have to empirically calibrate the numbers by examining past games. Perhaps an enterprising individual with more time than me could do it on the mafiascum dataset.

Metsfanmax wrote:

Epitaph1 wrote:I just learned that Ultimate Werewolf has a system for creating balance. Each role is assigned a number (positives for townies, negatives for werewolves) and you use those to get as close to zero as possible.

https://boardgamegeek.com/thread/641590 ... calculator

I haven't played UW yet (I will tomorrow) but my understanding is that most of the roles correlate to what we have in mafia. Perhaps we can adopt the number system to mafia roles and try a few setups based on that.

I don't think this will work very well for mafia. The combinations of roles make for very non-linear algorithms for determining balance. Even if it could work, you'd have to empirically calibrate the numbers by examining past games. Perhaps an enterprising individual with more time than me could do it on the mafiascum dataset.

I also don't really think this is how you should calibrate most setup's here. Most of our setup's are much more complicated than that of Werewolf where there is a very limited number of roles, mixed with known interactions. Because many of our mafia roles are much more complex/varied rather than create a balanced game it would be restricting instead.

So in other words, a great idea on theory, but not practical?

subtleknifewield wrote:So in other words, a great idea on theory, but not practical?

Thats the very simple conclusion.

It's clear that Epi wants a system to evaluate game balance, most importantly when it comes to the Official Games.
And if not by the point system, then how determine if any game is balanced?

What determines 'balance'? Is a game balanced when all parties involved have equal chance of winning? How determine this?
How would you even calculate a chance of winning %?

You could look at it from another angle, and ask yourself: when does a game become unbalanced?

Let's take HP mafia, again, as an example. Mafia won hands down, but the game was not unbalanced. Town was loaded on powers, but mafia played very well.
The fact that mafia did not lose powers with deaths was maybe an issue, and would have improved balance. It did not make the game unbalanced per say.

Personally, I consider mafia game balance something very subjective. And we aren't even including what is probably the most important factor in balance: the players. Put the 5 best players here in 1 team and the rest doesn't stand a chance, apart from maybe a serial killer allowed to perform 10 kills a day.

Probably what we need is a checklist of sorts

-Are power roles divided equally among parties?
-Will mass claiming break the game?
-Are there massively OP'd power roles in the game that need rebalancing?

And who would be doing the checking? That person won't be able to play...

Streaker wrote:It's clear that Epi wants a system to evaluate game balance, most importantly when it comes to the Official Games.
And if not by the point system, then how determine if any game is balanced?

What determines 'balance'? Is a game balanced when all parties involved have equal chance of winning? How determine this?
How would you even calculate a chance of winning %?

You could look at it from another angle, and ask yourself: when does a game become unbalanced?

Let's take HP mafia, again, as an example. Mafia won hands down, but the game was not unbalanced. Town was loaded on powers, but mafia played very well.
The fact that mafia did not lose powers with deaths was maybe an issue, and would have improved balance. It did not make the game unbalanced per say.

Personally, I consider mafia game balance something very subjective. And we aren't even including what is probably the most important factor in balance: the players. Put the 5 best players here in 1 team and the rest doesn't stand a chance, apart from maybe a serial killer allowed to perform 10 kills a day.

Probably what we need is a checklist of sorts

-Are power roles divided equally among parties?
-Will mass claiming break the game?
-Are there massively OP'd power roles in the game that need rebalancing?

And who would be doing the checking? That person won't be able to play...

I'd agree with that assessment Streaker.

There was a system that Mr. Squirrel had some time ago for determining how balanced a game was. Basically the idea was to simulate all possible actions randomly multiple times. From there you could see what probable outcomes were and try to adjust accordingly. It sounds tedious, but effective. Applying it to an open setup would likely yield interesting and (hopefully) balanced results.

That is easy to program but I think it is not helpful. There is no reason to suspect that the actions are random; that is generally not how the game progresses, and I don't think it averages to random over many games. So even if 80% of the random iterations show mafia winning, that doesn't mean the game is unbalanced, if the 20% where town wins are much more likely to occur together.

One approach is to determine what is the shortest amount of time it would take to complete a game. For example, if mafia, SK, and all other killing roles get their kills in each day/night. If a medium to large game could conceivably run its course in 3 days, even 4, then perhaps some adjustments could be made. This is more helpful when you're calculating how many killing or recruiting roles to include in a game.

To Streaker's point, balancing a game seems like more of an art than a science. It's just too difficult to run all the scenarios and get a reliable data set to nail down something that is purely balanced.

I think the best we could do is have a few sets of eyes looking at each set up before it goes live. fwiw, I had two other people review HP mafia before the game started. I guess multiple eyes isn't a silver bullet, but it doesn't hurt.