Mafia Game Design Theory

This topic is for discussing theory related to moderating or game design.

What kind of setups do you like? I like Upicks

So I created this topic because I had a few questions. I forgot one though so here are the other two, maybe I’ll remember the last one later.

  1. How do you decide whether a mechanic or role is fun to play with/as? In my opinion, you should try to include only roles that are fun for the person playing them, and don’t cause unfun effects for other players (i.e. cult leader is plenty fun to play, but getting recruited into a cult… maybe not so much). How do you decide what is fun to play, and what is unfun to play against?
  2. As far as balance goes, I think it’s simple to say that we should try to give games a balanced EV. But that isn’t always simple. Sometimes you have roles that aren’t easy to calculate EVs for. And even if you can easily calculate the effect a role has on EV, the role might be more or less powerful depending on the site meta. For example, Vigilante has an easily-calculated effect on EV, but its actual power can vary depending on how common serial killers are in the site meta. My question is how much should I account for meta when designing a setup? If designing a game here, should I account for the fact that town has an overwhelming winrate on mafia451, and include fewer power roles than I might include on other sites?

uPicks are fun to design but require so much work over so little time! I think I’ll wait until I have a week off before I host another one.
Open games are fun to design because of the unique challenge they pose. You have to consider all the different ways the players might try to break your game. Opens also give you the opportunity to try mechanics that aren’t suited for closed games.

Marking. Have thoughts on this to add later.

I want Urist to review games I design from now on.

1 Like

So, I broadly agree with Urist. The most important element of game design for me is whether a game is fun or not—there are massively swingy games that aren’t balanced in the traditional sense of the word, that are still very fun to play. This has led me towards designing generally smaller games in the 6-9 player range, for a few reasons. It’s easier to keep track of and account for role synergy and make sure that no one role is going to make someone else’s role completely useless. It’s also easier to make sure the game is hard to break mechanically, because that kind of game just isn’t much fun to me.

On the other hand, mechanics should be interesting enough that players want to make use of them. So it’s a fine line between “give incentive to think about mechanics” and “make sure neither side can win solely based on mechanics.”

In my opinion, one of the most interesting and difficult to get right things in game design is the risk/reward for reveals. There should almost always be some benefit to revealing, but it obviously can’t be as straightforward as “reveal so we win the game.” On the other hand, I also dislike when reveals hand scum an easy victory! So when I’m designing a game, I always try to make their be some serious pros and cons to reveals, to make it so that town mass revealing D1 doesn’t break the game for either side, but does change it in an intriguing way.

So, for an example of what I’m talking about, take the most recent game I ran here, generic superheroes. The goal of that game design was to give town sizable incentive to reveal (an IC, a watcher, a doc, a rolecop, etc), but to also punish those reveals by making it easier for scum to make successful NKs and role-blocks with reveals. I didn’t quite succeed this time, but I think I was very close and with a small tweak can make it that much more interesting a mechanic.

So that’s the sort of stuff I think about during game design—given equal play, do both sides have a fair chance to win? Is this game easily broken mechanically? Are the mechanics interesting enough and integrated into the game that they’re helpful, but not so necessary that you have to use them to win? And above all—will players enjoy playing it?

I’m also really coming around to the benefits of smaller sized open or semi-open set ups.

I think I agree with you on reveals. You’re saying that the info gained by a town reveal should be balanced by scum mechanics? A lot of open setups achieve this by including a mafia roleblocker, which your “weakness” mechanic emulated to a certain extent, but in a more interesting way.

I disagree with the idea that a mass reveal on d1 shouldn’t hand scum an easy victory. Town should always have serious disincentive to claim, simply due to how good claiming is for the town, especially in an open setup. I feel like a massclaim on d1 of your superhero game would have certainly handed scum an easy victory (had we mislynched at all), and I am totally ok with that.

Have you ever run a smalltown game? I feel like that might be right up your alley. It sets up the scenario of an open setup with a d1 massclaim, so you can play around with mechanics regarding that.

I actually have a question for you specifically Nanook: how do you balance games with random mechanics? What is your thought process behind designing roles with random elements?
For example, how do you come up with the % chance that X will happen when an ability is used?

I’ll do the easy one now, and the harder one later.

Any mechanic with randomness is always going to be difficult to truly balance. Pretty much impossible to, in fact. So including a mechanic with randomness is basically saying in big letters “THIS IS A BASTARD GAME.” So you have to accept you’re not gonna get it perfect. But, that’s no excuse to not get it close!

So, basically what I do with a mechanic with randomness attached to it is make it player triggered—nothing sucks more than something bad happening to you randomly that you had no control over. Make it clear that there’s a randomness element to it. Attach it to an otherwise powerful mechanic (rewinding the game, for example). Try to weight it so that the player in control has to think really carefully before activating it. And then put some common sense parameters on it so it isn’t game breaking.

I’ll pull up my notes when I’m home, but I think I made using the speedsters time travel role something like a 5% chance per player to die and 10% per player to change alignment for each phase that was rewound. With the parameter that it couldn’t cause an insta-win for either side. So that’s something that’ll give the player in charge of it a LOT of pause about actually using it.

Alright, now to the other one.

So, I definitely think that games shouldn’t be breakable for town by mass claiming D1. However, I also don’t think it should be an auto-win or close for scum. To me, the ideal result from D1 reveals is that it’s a balanced but different game. So town can’t autowin by coordinating their night actions, and scum can’t walk to victory by killing all the useful powers.

I can see the side of the argument that says town massclaiming should be punished severely, and recognize that I’m in the minority of thinking otherwise. But, in my opinion, it makes for a much more fun and interesting game when there’s a tension between “do I reveal, or do I not reveal?” rather than just an easy choice to keep it hidden. It’s hard to do, but yeah, I think that it’s much more interesting when games have that tension between revealing and not revealing.

I haven’t run a smalltown game, but googling it, it looks pretty interesting! I might run a game like that some times. Actually, maybe the sequel to Superheroes will be a smalltown setup, if I can figure out a way to rework the NK.

I agree games can be swingy and be fun but imo the mod needs to state it. There was a game I played where because the role design made no sense it tanked town’s entire ability to play the game.

Imo you should never make assumptions about a closed setup design, maybe unless you’re familiar with the mod’s design tendencies, but even then it’s dicey.

1 Like

Give everyone a role.

Hand pick roles and affliations for specific people to maximize conflict.

Put a few bussers, role blockers, watchers in.

Watch fucking insanity insue.

We have a bunch of tags to help with this.

I think it can be assumed that #nonstandard and #role-heavy will be swingy.

I think the things that probably need to be explicitly restated in a game topic are

  1. Potential for moderator lies (bastardness)
  2. Potential for alignment changes
  3. Potential for fliplessness
  4. Potential for post-death mechanics

Because they’re probably the things that are potential dealbreakers for some players that would otherwise play a role-heavy nonstandard game.


Take number of players
Make that many roles
balance the roles between scum and town with a 25% mafia, 33% total non town aligned figure. that tends to be pretty balanced.

Obviously you scale down for smaller games

With how our automation system is going to be designed I don’t think it’ll be that much extra (but still nontrivial) work to make an engine that can monte-carlo closed setups to calculate balance. Might be interesting.

Probably would have to select from a pre-programmed role base

Eh, i dunno. It’s the mod’s responsibility to try his best to make a game that’ll be enjoyable for everyone who signs up. To that end, wild crazy bastard games should 100% carry a disclaimer.

1 Like