By Kraj
This article is intended for designers who have a firm grasp of the basics, perhaps even have run a game or two, and are interested in trying more complicated setups and innovative roles. I’ve split this article into four sections: Core Elements of Mafia, Evaluating Individual Roles, Balance vs. Swinginess, and General Tips and Guidelines. I also follow up with some advice for how to be a good game moderator. If you haven’t already, I strongly suggest reading Puzzle’s basics of game design. I also suggest reading Mark Rosewater’s column on WotC’s website, as many of the underlying concepts that make good Magic design also apply to Mafia.
I’d like to start by reiterating and expanding on Puzzle’s #
1 point:
This is something designers often lose sight of when creating a complex game, not because they’re not trying to make the game fun for the players but because unfun mechanics are not always obvious. The first step to understanding what is fun for players is understanding the core elements of a Mafia game, i.e., what makes the game what it is and how that translates into fun.You are making a setup for your players to have fun, not for you. The lesser your presence is felt, the best the game will be for the players.
Core Elements of Mafia
A. Informed Minority vs. Uninformed Majority
B. Voting, and the Majority Lynch
C. Night Period When the Mafia Gets to Kill
These three mechanical elements are the fundamental aspects that make the game what it is. It’s very difficult to change one of these elements and still have it feel and play like a game of Mafia. Some setups have successfully done so, but most that try end up failing in
some way. As a rule of thumb, if you are going to alter the basic mechanics of the game, change only one and keep the rest intact. Take, for example, the Kingmaker variant. This variant is successful at changing the way voting and lynching is handled, but rarely if ever will you see a Kingmaker with untraditional alignment groups, daykill, lack of night period, etc.
It is important to understand, however, that these mechanical elements aren’t inherently fun, it’s the gameplay that emerges from them that’s fun. This distinction is important because when thinking about innovative mechanics to try, you need to evaluate whether the fundamental gameplay is preserved. The fundamental mechanics with corresponding gameplay looks like this:
A. Informed Minority vs. Uninformed Majority (Behavioral Analysis)
B. Voting, and the Majority Lynch (Persuasive argument)
C. Night Period When the Mafia Gets to Kill (Strategic choices)
So the point is, if you’re trying a setup that alters the
night period somehow, you should make sure that strategizing – particularly for the mafia – is still relevant to the game. If you’re going to alter how lynching works, you need to make sure persuasive argument is still important. And so on. For example, the variant of having two mafia groups in competition with each other in addition to the town can be reasonably well-balanced and seem fun. However, this effectively destroys the town’s ability to analyze behavior because the two mafia groups are much more similar to the town in terms of information and motivation. Therefore if you implement this variant you need to find a way to reintroduce behavior difference for the town to analyze.
This concept gets especially tricky, though, when you also consider that adjusting these fundamental mechanics not only affect their corresponding gameplay but also the gameplay of other mechanics. For example, night abilities add variety to the game and ensure that no two games will have all the same strategic choices.
However, too many night abilities or too powerful night abilities diminishes the focus on behavior and the relevance of persuasive argument. To illustrate my point, imagine a setup where every townie is a cop and every mafia is a roleblocker. The result of this game would depend 100% on night actions and have nothing whatsoever to do with argument, votes, analysis, or persuasion. Obviously this is an unrealistic extreme, but the relationship holds true: the more powerful and the higher the number of abilities, the less relevant behavior analysis and persuasive argument becomes. So while just about any setup can and should be enhanced by interesting night abilities, too much of a good thing can kill the fun.
Understanding the core mechanics and why they make the game fun is the most important thing you can learn about designing Mafia setups.
Evaluating Individual Roles
The most important technique for creating a new mechanic or role, or for deciding how well an
established role fits your setup, is considering how the role might be played. This means stop thinking about the role as the moderator (in terms of the entire setup) and think about it from the player’s perspective (who only has that information, or that and his or her mafia partners’). Remind yourself that as a townie you won’t know for sure whether you can trust a certain player; an obvious interaction between players might be less obvious when you can’t be sure of the other’s alignment, etc. Think about what is a logical action under normal circumstances. What other roles might this mechanic interact with in a significant way? What information might change how the role is played? Thinking about how you might play a role can help you realize when a role isn’t fun.
However, there is a pitfall here to watch out for: while it’s important to have an idea of how a role might play out, it is dangerous to expect a role to play out a certain way, especially if your expectation affects the game balance. While
it’s out of your control if a player does something completely off the hook – like a cop who just doesn’t bother to investigate anyone – thinking along the lines of “if this player claims the cop is sure to investigate him” is almost sure to be proven wrong. Do not balance roles on the expectation of it being played a certain way.
Some other general things to watch out for when designing roles (some of these are reiterated from Puzzle’s aforementioned article):
- Too much certainty kills the game. Roles that pump a lot of reliable information into the game should be minimized. Mafia commonly devolves into a process-of-elimination game, which is the antithesis of the core mechanics, due to too much ability-based information in the game. Roles that provide potentially useful information depending on how it is used are a good compromise. Roles that allow the mafia room to lie, misdirect, and disrupt the town are often essential.
- On the flip side, the wrong kind of uncertainty can ruin a role. A player should have the relevant information needed to play the role correctly. You can hide things from a player, but they shouldn’t be things that not knowing will potentially screw that player over. An ambiguous win condition is the perfect example of bad uncertainty.
- Risk/reward tension is a good tool for roles and abilities, particularly in townies. When an ability has the potential to both help or hurt your faction, or when it could be very helpful but only in the right situation, etc., it creates interesting decisions. Sometimes, though, tension is pushed into the territory of being forced to play directly contradictory to your main win condition, such as having a win condition that a member of your faction dies. This is almost always a bad thing.
- [
*]Passiveness. Whenever you design a role with a passive (i.e., always in effect) ability, consider whether it could be done as an activated ability, or passive but only under certain circumstances. Letting players make decisions about when and how to use an ability is more interesting than simply having an ability and waiting for it to do something. Many abilities do make the most sense as passive, though, so I’m not suggesting no ability should ever be passive. Along those same lines, one-shot abilities are often more fun than repeatable abilities because there is a decision of when to use it instead of automatically using it every night. - Randomness. When applied correctly, randomness can be an interesting tool, but blunt mechanics like a random chance of success for an ability are generally unfun. All-or-nothing randomness tends to leave players never knowing when or whether they should use an ability. “Good” randomness should add some uncertainty to a role without
affecting how the role should be played too much. For example, a role that gains one of three pre-determined abilities at random at the beginning of the night would be interesting. But a role that asks a player to choose a target and afterwards the ability is selected is probably bad. - Narrowness. When an ability has “one right way” to play it, it’s less fun. Consider this role: “Each night you may either prevent all kill attempts made on target player, or learn target player’s role name. You win if Billy survives to the end of the game.” This role presents the player with limited decisions because they essentially have to name cop until they find Billy, then protect him or her once they do. There’s no strategy involved, no real decisions to make. You should instead design roles that give the player tools that can be used in a variety of ways.
- Misdirection. Red herring abilities (ones that trick players into thinking they are relevant) should be
rarely and carefully used. It’s incredibly frustrating for players to discover at the end of the game that they put thought and effort into decisions that actually didn’t matter. - Extraneous abilities or clauses. Abilities so weak in comparison to a player’s other abilities that they’d never logically use them are pointless clutter. Special clauses that interact with other roles in a very narrow set of circumstances add complication to the game without adding play value.
Balance vs. Swinginess
When balancing your setup there are two main factors to consider: balance, and swinginess.
Balance, put simply, is the chance each faction has of winning the game. The goal is to give
the mafia and town an equal chance to win, with neutrals having a lesser but reasonable chance. The first step to balancing a setup is to start with Axelrod’s pointing method to assign each role points and make sure both factions have roughly equal totals. Having an idea of how a new role might be played is crucial to determining an accurate point assignment. From there you need to consider other factors that can affect the outcome, like special information a faction may have, the likely effect of a mass claim on both the town and mafia, how much time the town is expected to have until lylo, and interactions between roles. Every setup should strive to be as balanced as possible.
Swinginess is the degree to which individual roles affect the outcome of the game. In its most basic form, Mafia is one lynch each day and one kill each night. The higher potential the game has to deviate from that, the swingier it is. This means that a certain degree of swinginess is desirable, since the game would quickly become
dull and repetitive without cops to find scum, vigs to shoot players, SKs to thin the field, and so on. Swingy roles can also benefit games by buffering against a landslide victory by a faction, the Serial Killer being an excellent example. The trick is to make a setup swingy enough to be interesting but not so swingy that behavior analysis and night kill strategy become unimportant compared to abilities. There’s no clear line the sand to be drawn that says “too swingy” or “not swingy enough” since different players have different preferences. Some players prefer games with minimal swinginess, while other players enjoy setups with extremely powerful roles. However, the more swingy elements are in a game the more difficult it tends to be to balance. Therefore less-experienced designers should strive to minimize swinginess, and as you gain more experience then fiddle with swingier design space.
Some methods to avoid swinginess include:
- An overall lower power level/ higher ratio
of vanilla roles to power roles. - Spreading abilities out among as many roles as possible, rather than having fewer but more powerful roles. For example, two one-shot vigs in a game are generally less swingy than one full vig.
- Backup roles act as buffers by keeping relevant abilities in play when the original owner dies. Along the same lines, sometimes a mafia faction will have group abilities any one of them can use, such as a roleblock, rather than risk an important ability being lost too early in the game.
- 1-Shot abilities are a good way to have powerful abilities like cop investigations, vig shots, mass roleblocks, resurrection, etc., in the game while being less swingy than if they could be used every night.
the term “swingy” is used to describe a role or setup that is undesirably swingy.
General Tips and Guidelines
There are various lessons I’ve learned about setup design that don’t really fit into a larger category. Here they are:
Every setup should have a goal in mind before you start designing, a definition of what you want the setup to do. It can be exploring a certain design space, implementing a particular mechanic idea, crafting a game around a certain flavor, fixing a problem you’ve identified with other setups, and so on. If you have a goal it gives the setup focus and direction, which makes the design process easier and usually results in a stronger end result.
Often a setup will feature a gimmick or have a special theme. You want to make sure every role or almost every role connects to that in a fun, meaningful way. You don’t want only a handful of roles to feature it and the rest don’t at all, or only loosely relate to it. When the moderator presents a
mechanical theme, players expect the role to have that theme and will be disappointed and confused if they don’t. This is especially true for a game selected by the FTQ or PCQ. A good example of this mistake is Tales of the Fantastic.
Given the long times moderators have to wait before getting to run their games, there’s a strong temptation to fit as many cool ideas into one setup that you can. This usually results in chaotic, confusing setups that don’t play very well. Pick one idea, or maybe one major idea and one minor idea, and do it well. Focus on roles that interact with that idea in interesting ways, explore the space, etc. Don’t try to do too much at once. Also, don’t feel compelled to have as large a number of players as possible. The larger the game the longer it goes, the harder it is on players, the harder it is to find replacements, etc. Rather than start with 24 slots to fill as the default, try starting with 20 and then only add more slots if the balance or mechanics warrant it.
Design
more roles than you have room for. If you’ve got a 20 player game and you designed 20 roles, your game has bad roles. By the time you’re done, you should have at least a couple roles that you decided had some flaw or that you just don’t like as much as some others that you’ve cut.
Complexity vs. Depth –The more mechanics you add to a game the more complex it becomes. There’s a fuzzy line between complex enough to be interesting and so complex it’s confusing and frustrating, but you want to avoid crossing that line. Depth refers to the impact a mechanic has on the game in terms of how interesting it is to play, what interactions it has, what decisions it creates, etc. There is a concept in game design called elegance that, in a nutshell, means maximizing depth while minimizing complexity. Elegance is difficult to achieve but is an excellent goal. When you add a mechanic to the game or to a role, consider what it adds to the gameplay. Does it create meaningful decisions? Does it encourage strategizing? Does
it have cool interactions? Is it just plain fun? If you’re adding something to the setup that doesn’t really add much to the gameplay experience then take it out. For example, let’s say your setup’s flavor has lots of robots and ninjas. You decide that every role has a “type” of robot or ninja. Then you have one role in the game that repairs robots, but that’s the only mechanic that actually cares about the “type”. You’ve added a layer of complexity to the game without really adding any gameplay elements. Either remove the “type”, or commit to it by making it a central theme and creating more roles and interactions that care about robots and ninjas.
Along that same line, when designing a game around a certain flavor there’s a tendency to design roles that mechanically interpret the flavor as literally as possible. These type of designs tend to get clunky and have complexity without depth. For these roles, try to get to the core of what you want the role to “do” in the game and make that mechanic happen.
Trim off anything that doesn’t add to the role.
How to Be a Good Moderator
So you’ve finished your game design and your players are going to love it. Well, you’re not done yet. Now you actually have to run it, and that requires and entirely different set of skills. These are some tips and suggestions I’ve picked up along the way:
Be attentive. If you want players to be active and engaged in your game, then you should be too. Provide regular votecounts, prod players who haven’t posted, be available to answer questions, etc.
Be clear. When you write up your role PMs, pay attention to the language you use and try to anticipate where players might be confused, or identify any ambiguity on how an ability functions.
Be organized. Spreadsheets are a great way to keep track of vote counts and abilities used. Mistakes will happen and that’s ok, but you should try to minimize them.
Know before the game starts in what order abilities resolve, whether roleblocked players
will be informed, whether roleblocking a 1-shot ability uses up the shot or not, etc. Identify strange interactions between abilities and corner cases and know ahead of time how to handle the situation rather than figuring it out at the time it comes up.
Avoid game-relevant flavor in night actions. Everyone likes a flavorful PM instead of “You saw that Bobby was targeted by Sally last night”, but be careful about including details that hint towards game-relevant information. A doctor, for example, shouldn’t get flavorful clues as to the identity of the killer if s/he blocks a kill. Try to make flavor interesting but as ambiguous as possible. Rule of thumb: if you think players will interpret your flavor in a certain way, they won’t.
This last point is controversial and many mods will disagree, but I suggest you do not be afraid to modkill players for inactivity. Modkills over inactivity are generally avoided because the moderator’s judgment is influencing the outcome of the game, but in my experience
roles that have multiple replacements during the game hurt the gameplay and drag the overall experience down for everyone. Usually replacing is the right thing to do, but sometimes the best and most fair way to move the game forward is to modkill.