Heya ... wanted to post some of my thoughts about "mafia" style games and what they could ... become ... given a little work and a little thought.
My issue is largely that it seems there's often little strategy involved; at least half of the votes, often more, have nothing to do with clues but instead have to do with forum-related and/or posting-related choices. For example, the first vote in mafia 32 was as far as I could tell entirely random, and the second was based on the desire for one of the players to abstain. The third looks as if it's based on clues at least in part, but perhaps only for a few players.
The issue, to me, is that there's a strong disincentive to actively vote for someone you think is the mafia. The first player or two basically draw attention to themselves, and likely are killed by the mafia themselves, while the rest of the players have a strong incentive to docpile, as it were, to avoid attracting attention of either voters or mafia. Once three or four people have voted for one player, it's safe to jump on, and most people who are playing to win, do.
Also, there's really no way to tell who the mafia might be for several rounds. The GM can't give 'good' clues in the first few rounds, or it too quickly becomes obvious - even if you give red herrings also, unless you give dozens of them (in which case you might as well not clue at all) the red herrings will get killed first, leaving your clues to be obvious a few turns down the road. Although that resolves itself later in the game, as fewer players are present and clues can be made more obvious (if still devious).
So I will suggest the following game, presented in the following few posts.
Clues
Players: 8+ (16+ for a great game)
Description:
Players find themselves locked in a deserted castle of some sort, with a murderer on the loose. Their job: to find out who killed Mr. Corpsey, before they are killed themselves
Setup:
Players choose one of four classes for their character. Players may (optional, GM's choice) either choose teams of up to 4 players, be assigned teams by the GM, or start out solo and join up with whomever they like (up to four players in a group) during the game. All players start out in the Study, and may explore and traverse the castle at their whim. This may serve as a meet-up location, and is a 'safe' room (no kills may occur in this room, and a sniper may not attack from this room out; although he may attack through the room). However, players may spend no more than two consecutive turns in the Study without leaving. If a player spends two consecutive turns and the third does not leave (counting the first turn as the turn in which he enters the room, and the second being a turn in which he chooses not to move), the GM moves him to a room at random. Other rooms may be stayed in indefinitely, but the killer of course can always find you ...
Map:
The GM must make a map of the castle, and it must be entirely made before the game begins. It must consist of a series of rooms, each of which has a number of entrances which pass to other rooms. Rooms may have 1 to 4 entrances (N, S, E, W) and must be geographically logical (ie a room that is east of one room and west of another must have those two rooms as next to it on the correct side, and travelling N, E, S, then W must return you to your original room regardless of the location in the castle, assuming these entrances exist). Although no specific regulations exist as to the size of the castle or the specific layout, there should be an adequate number of rooms in a 'row' to allow for the skills and actions that gain advantage from seeing or moving two consecutive rooms in a row, and should have enough rooms to allow for interesting play without being so enormous as to delay the game unreasonably. At least one room per player is recommended.
Exploring:
Players start out not knowing anything about the castle except for the rooms adjoining the study. Each turn, they may make one move from one room to another, and upon entering a room they discover what rooms adjoin that room. Players organize in parties (voluntarily or mandatorily, up to the GM) and go out; however they do not have to stick with their parties. It is assumed that they do as long as they choose identical sets of moves. Players who perform a different action but the same move are still in the party but do not gain any bonus from the actions others take (still sharing clues and whatnot, but not gaining bonuses directly related to actions others take).
Action:
Then they may take an action. They may use one of their abilities, which are limited to three uses each per game (six total). In addition, one of the following actions may be used:
Camp: Player makes a camp (with other party members who also choose camp) that is well defended and watched in the middle of the room. Bonus: +50% chance of spotting attacker if a camped party member is attacked; if spotted, 35/50/15 (35% chance to kill attacker, 15% chance killer kills victim and escapes, 50% chance killer escapes w/o killing victim; regardless, killer's identity is known). (For example, an attack with no other modifiers will have a 75% chance to be caught by an encamped party if attacking an encamped party member.) Two or more players must camp together to gain any bonus.
Search Room: Player looks about room, searching for clues. Bonus: Double the chance compared to normal of spotting physical clues in the room (before other modifiers).
Hide: Player hides behind whatever is convenient in the room. Some rooms may be harder than others (GM discretion, and it does not have to be revealed to the players directly), affecting the hiding bonus. Bonus: 50% base chance of not being seen; +25% chance of seeing attacker if he acts in the room (on someone else). Bonuses are halved (after modifiers) if the player moved that turn.
Run: Player continues moving in the same direction of his first move. He is not seen by players in the in-between room. He has a 0% chance (unmodifiable) to see clues in either room.
Follow: Player chooses to follow another player. He must not have moved that turn (this counts as his move and as his action). The player has a 50% chance of being spotted by the player he follows; that chance is decreased to 25% if he was hiding last turn. If the player is spotted, a 10% penalty is applied to any further action they take that turn, and the other player is notified they are being followed; there is a 50% chance they will be notified of whom is following them. (The 10% penalty is taken, for example, if the killer is following someone and noticed, then if he chose to kill that person this round, not only is a 25/50/25 chance calculated just as if he were spotted in any other situation, but it is now a 35/50/15 chance; conversely if a player follows the killer, and then is targetted and/or the killer targets someone else, they suffer a 10% penalty if they are noticed, making it a 15/50/35 chance.)
Clues:
Upon entering a room, players have a chance to discover clues in that room. The base chance is 25% chance for either a physical clue OR a situational clue, if either or both are present. Obvious clues are also occasionally present, in which case they are automatically discovered; obvious clues include dead bodies and other things that would be seen by any observer. (It is up to the GM to determine clues left and which are obvious and which are not. Each murder must leave at least one obvious clue - the dead body - and at least one hidden clue.)
The clues may be more or less clear - this is entirely in the GM's hand.
A Killer Strikes!
Each Killer must kill at least one person every three turns. Trap kills count, as long as they occur within the time frame, as a kill by the setter of the trap. If someone is not killed for two consecutive turns by one of the killers, the third turn the Killer must actively attempt to kill someone. (The GM may modify this rule as fits the number of players and killers.)
The killer may simply directly kill someone in the same room as himself, or may kill someone through the use of one of his skills (Snipe or Trap). "Kill" is an action that may be combined with other actions, except that it may not be combined with both 'move' and 'hide'.
The Innate Ability of the Killer is to have +50% actions available to him, and they are 'pooled'; no more than 3 actions of each type may be attempted, but 9 actions are allocated to the Killer for all four abilities he has. He has a normal class, in addition to being the killer - for example, a Soldier Killer or a Gentleman Killer.
The killer may kill before OR after moving - other players must move first before acting; however any other action follows the same rule as other players.
There is always a 25% chance the Killer will be noticed, on a normal kill. If he is noticed, then a 25/50/25 chance is calculated (plus other modifiers); a 25% chance of the killer being killed, a 50% chance of him escaping, and a 25% chance of the killer killing the victim and then escaping.
The Killer may choose to kill other killers (and does not know any other killers' identities); in fact, there is only one victor in this game (or team of victors if a conversion occurs). If the game is down to only killers, the two killers or teams of killers attempt to kill each other just as they were killing detectives; and it can be to the Killers' advantages to do some detecting in addition to killing people, as they are able to detect just as any detective is.
Conversion!
A killer may choose to reveal his identity to one or more players, and attempt to 'convert' them to being a killer. No more than three players may be converted by any one killer. (The GM may limit this further in smaller games.) Groups of killers still may not kill more than one player per original killer (the actual killer designates either who dies that turn or which member of his group may kill someone). If the killer converts the team, it is a 100% conversion - a player who tells the GM that he has converted may not return to being a detective, and wins or loses based on the killer's fate. However, there is an inherent danger in that the killer must reveal his or her identity to the to-be-converted player before receiving an answer. The converted player must be in the same room as the Killer, although a conversion attempt does not take an action. The identity of the Killer does not have to be revealed prior to the actual conversion, although the discussion may not begin unless they are in the same room and may only continue during turns that they are cohabiting the same location, so it will probably be obvious who it is quickly. If the GM believes that coordination has occured outside of the game, they may assess any reasonable penalty, including disallowing or forcing conversion of players, or 'killing' them.
The converted killers may then follow the killer that converted them, or follow another player, or go wherever they want, and may communicate with the killer regardless of location inside the castle. They gain all of the abilities of the killer, but do not gain the 50% bonus to 'actions'.
If a killer accidentally meets another killer, they may choose to join forces, but one of them relinquishes the title of "Killer" and becomes a groupie, with only one kill allowed for the group (just as any other). It may not occur after the point in which killers and groupies outnumber detectives (the GM will inform the killers if this occurs only in this instance).
End of the game:
The game ends when either the last killer or groupie is killed, or when only one killer or killer groupie remains. If the lead killer in a group is killed, then the GM chooses another, unless the lead Killer has informed the GM of his successor's identity prior to the turn he is killed during. Allied wins are not allowed except among a group of killers, or by the entire team of non-killers.
What do you think? It's pretty detailed, the product of several days of thought, particularly today. I suspect I left some details out, but I'm looking more for general comments as to whether people would find this interesting. If enough people would, I'd like to try a test game with these rules (with modifications as needed by people spotting issues). Balancing issues undoubtedly would come up as well ... particularly with the skills.
Following is a list of the class-specifics in a jpg.
My issue is largely that it seems there's often little strategy involved; at least half of the votes, often more, have nothing to do with clues but instead have to do with forum-related and/or posting-related choices. For example, the first vote in mafia 32 was as far as I could tell entirely random, and the second was based on the desire for one of the players to abstain. The third looks as if it's based on clues at least in part, but perhaps only for a few players.
The issue, to me, is that there's a strong disincentive to actively vote for someone you think is the mafia. The first player or two basically draw attention to themselves, and likely are killed by the mafia themselves, while the rest of the players have a strong incentive to docpile, as it were, to avoid attracting attention of either voters or mafia. Once three or four people have voted for one player, it's safe to jump on, and most people who are playing to win, do.
Also, there's really no way to tell who the mafia might be for several rounds. The GM can't give 'good' clues in the first few rounds, or it too quickly becomes obvious - even if you give red herrings also, unless you give dozens of them (in which case you might as well not clue at all) the red herrings will get killed first, leaving your clues to be obvious a few turns down the road. Although that resolves itself later in the game, as fewer players are present and clues can be made more obvious (if still devious).
So I will suggest the following game, presented in the following few posts.
Clues
Players: 8+ (16+ for a great game)
Description:
Players find themselves locked in a deserted castle of some sort, with a murderer on the loose. Their job: to find out who killed Mr. Corpsey, before they are killed themselves
Setup:
Players choose one of four classes for their character. Players may (optional, GM's choice) either choose teams of up to 4 players, be assigned teams by the GM, or start out solo and join up with whomever they like (up to four players in a group) during the game. All players start out in the Study, and may explore and traverse the castle at their whim. This may serve as a meet-up location, and is a 'safe' room (no kills may occur in this room, and a sniper may not attack from this room out; although he may attack through the room). However, players may spend no more than two consecutive turns in the Study without leaving. If a player spends two consecutive turns and the third does not leave (counting the first turn as the turn in which he enters the room, and the second being a turn in which he chooses not to move), the GM moves him to a room at random. Other rooms may be stayed in indefinitely, but the killer of course can always find you ...
Map:
The GM must make a map of the castle, and it must be entirely made before the game begins. It must consist of a series of rooms, each of which has a number of entrances which pass to other rooms. Rooms may have 1 to 4 entrances (N, S, E, W) and must be geographically logical (ie a room that is east of one room and west of another must have those two rooms as next to it on the correct side, and travelling N, E, S, then W must return you to your original room regardless of the location in the castle, assuming these entrances exist). Although no specific regulations exist as to the size of the castle or the specific layout, there should be an adequate number of rooms in a 'row' to allow for the skills and actions that gain advantage from seeing or moving two consecutive rooms in a row, and should have enough rooms to allow for interesting play without being so enormous as to delay the game unreasonably. At least one room per player is recommended.
Exploring:
Players start out not knowing anything about the castle except for the rooms adjoining the study. Each turn, they may make one move from one room to another, and upon entering a room they discover what rooms adjoin that room. Players organize in parties (voluntarily or mandatorily, up to the GM) and go out; however they do not have to stick with their parties. It is assumed that they do as long as they choose identical sets of moves. Players who perform a different action but the same move are still in the party but do not gain any bonus from the actions others take (still sharing clues and whatnot, but not gaining bonuses directly related to actions others take).
Action:
Then they may take an action. They may use one of their abilities, which are limited to three uses each per game (six total). In addition, one of the following actions may be used:
Camp: Player makes a camp (with other party members who also choose camp) that is well defended and watched in the middle of the room. Bonus: +50% chance of spotting attacker if a camped party member is attacked; if spotted, 35/50/15 (35% chance to kill attacker, 15% chance killer kills victim and escapes, 50% chance killer escapes w/o killing victim; regardless, killer's identity is known). (For example, an attack with no other modifiers will have a 75% chance to be caught by an encamped party if attacking an encamped party member.) Two or more players must camp together to gain any bonus.
Search Room: Player looks about room, searching for clues. Bonus: Double the chance compared to normal of spotting physical clues in the room (before other modifiers).
Hide: Player hides behind whatever is convenient in the room. Some rooms may be harder than others (GM discretion, and it does not have to be revealed to the players directly), affecting the hiding bonus. Bonus: 50% base chance of not being seen; +25% chance of seeing attacker if he acts in the room (on someone else). Bonuses are halved (after modifiers) if the player moved that turn.
Run: Player continues moving in the same direction of his first move. He is not seen by players in the in-between room. He has a 0% chance (unmodifiable) to see clues in either room.
Follow: Player chooses to follow another player. He must not have moved that turn (this counts as his move and as his action). The player has a 50% chance of being spotted by the player he follows; that chance is decreased to 25% if he was hiding last turn. If the player is spotted, a 10% penalty is applied to any further action they take that turn, and the other player is notified they are being followed; there is a 50% chance they will be notified of whom is following them. (The 10% penalty is taken, for example, if the killer is following someone and noticed, then if he chose to kill that person this round, not only is a 25/50/25 chance calculated just as if he were spotted in any other situation, but it is now a 35/50/15 chance; conversely if a player follows the killer, and then is targetted and/or the killer targets someone else, they suffer a 10% penalty if they are noticed, making it a 15/50/35 chance.)
Clues:
Upon entering a room, players have a chance to discover clues in that room. The base chance is 25% chance for either a physical clue OR a situational clue, if either or both are present. Obvious clues are also occasionally present, in which case they are automatically discovered; obvious clues include dead bodies and other things that would be seen by any observer. (It is up to the GM to determine clues left and which are obvious and which are not. Each murder must leave at least one obvious clue - the dead body - and at least one hidden clue.)
The clues may be more or less clear - this is entirely in the GM's hand.
A Killer Strikes!
Each Killer must kill at least one person every three turns. Trap kills count, as long as they occur within the time frame, as a kill by the setter of the trap. If someone is not killed for two consecutive turns by one of the killers, the third turn the Killer must actively attempt to kill someone. (The GM may modify this rule as fits the number of players and killers.)
The killer may simply directly kill someone in the same room as himself, or may kill someone through the use of one of his skills (Snipe or Trap). "Kill" is an action that may be combined with other actions, except that it may not be combined with both 'move' and 'hide'.
The Innate Ability of the Killer is to have +50% actions available to him, and they are 'pooled'; no more than 3 actions of each type may be attempted, but 9 actions are allocated to the Killer for all four abilities he has. He has a normal class, in addition to being the killer - for example, a Soldier Killer or a Gentleman Killer.
The killer may kill before OR after moving - other players must move first before acting; however any other action follows the same rule as other players.
There is always a 25% chance the Killer will be noticed, on a normal kill. If he is noticed, then a 25/50/25 chance is calculated (plus other modifiers); a 25% chance of the killer being killed, a 50% chance of him escaping, and a 25% chance of the killer killing the victim and then escaping.
The Killer may choose to kill other killers (and does not know any other killers' identities); in fact, there is only one victor in this game (or team of victors if a conversion occurs). If the game is down to only killers, the two killers or teams of killers attempt to kill each other just as they were killing detectives; and it can be to the Killers' advantages to do some detecting in addition to killing people, as they are able to detect just as any detective is.
Conversion!
A killer may choose to reveal his identity to one or more players, and attempt to 'convert' them to being a killer. No more than three players may be converted by any one killer. (The GM may limit this further in smaller games.) Groups of killers still may not kill more than one player per original killer (the actual killer designates either who dies that turn or which member of his group may kill someone). If the killer converts the team, it is a 100% conversion - a player who tells the GM that he has converted may not return to being a detective, and wins or loses based on the killer's fate. However, there is an inherent danger in that the killer must reveal his or her identity to the to-be-converted player before receiving an answer. The converted player must be in the same room as the Killer, although a conversion attempt does not take an action. The identity of the Killer does not have to be revealed prior to the actual conversion, although the discussion may not begin unless they are in the same room and may only continue during turns that they are cohabiting the same location, so it will probably be obvious who it is quickly. If the GM believes that coordination has occured outside of the game, they may assess any reasonable penalty, including disallowing or forcing conversion of players, or 'killing' them.
The converted killers may then follow the killer that converted them, or follow another player, or go wherever they want, and may communicate with the killer regardless of location inside the castle. They gain all of the abilities of the killer, but do not gain the 50% bonus to 'actions'.
If a killer accidentally meets another killer, they may choose to join forces, but one of them relinquishes the title of "Killer" and becomes a groupie, with only one kill allowed for the group (just as any other). It may not occur after the point in which killers and groupies outnumber detectives (the GM will inform the killers if this occurs only in this instance).
End of the game:
The game ends when either the last killer or groupie is killed, or when only one killer or killer groupie remains. If the lead killer in a group is killed, then the GM chooses another, unless the lead Killer has informed the GM of his successor's identity prior to the turn he is killed during. Allied wins are not allowed except among a group of killers, or by the entire team of non-killers.
What do you think? It's pretty detailed, the product of several days of thought, particularly today. I suspect I left some details out, but I'm looking more for general comments as to whether people would find this interesting. If enough people would, I'd like to try a test game with these rules (with modifications as needed by people spotting issues). Balancing issues undoubtedly would come up as well ... particularly with the skills.
Following is a list of the class-specifics in a jpg.
Comment