Julius isn't saying "scum do scummy things", he's saying "scum never do silly things". He is saying that scum actions are always linear and entirely predictable, and anything that seems even mildly implausible should be treated as not possible. I'm quite familiar with the reasoning, since it was the entire basis for lynching me over the three scummier options in Elderscrolls Mafia. "The mafia had better options than to block you; therefore, you are scum. The cult had better options than to recruit Scipio; therefore, he and CK are town. Mole defended the town cop from lynch yesterday; therefore, he is town."

Also, Bulba brought up the guilty several times, as did a few others. However, point stands. At the end of Day 1, it should have been so blatantly obvious who the scum were that the town couldn't possibly lose.

Quoted from Rotaretilbo:

Also, Bulba brought up the guilty several times, as did a few others.

Didn't even notice this over Maffia's call to arms against Herr and my palms slapping my face.

Maffia's reasoning was that it was fine, because you could just kill Mole that night. Of course, you couldn't, being dead and all. What was odd was that he used your death as evidence against Herr, so it wasn't like he hadn't noticed.

The biggest problem I saw with lynching Herr was that he was confirmed to be Take Two and either town or the most helpless scum ever, what with being unable to kill a fly. Any other scum had a chance of being a killer that wasn't Take Two.

You're confusing early with ASAP.

Indeed, by early I meant ASAP exactly as you defined it. See Newt's ongoing mafia for what I was going for.

Totally irrelevant, but you should probably remove my flavor guide, Mole. That was just me being bored and it's not at all helpful or constructive.

EDIT: I can rewrite it if you want. My writing has improved considerably since then.

[This message has been edited by Ashrzr (edited 06-07-2012 @ 02:51 PM).]

Quoted from Rotaretilbo:

Julius isn't saying "scum do scummy things", he's saying "scum never do silly things". He is saying that scum actions are always linear and entirely predictable, and anything that seems even mildly implausible should be treated as not possible.

I'll say this one more time only, as you apparently don't want to try to understand.

We are considering a case much like the one Dante quoted. We suppose that a Scum Player is in a Situation where they have two options to choose from. One option is intrinsically (ie in itself) more likely to secure their objective. In the example, one road provides an intrinsically better chance of escape. We further suppose that this is so obvious that everyone realises that that option is intrinsically more advantageous for the Scum Player. We also assume that the Scum Player is (a) rational, and (b) wishes to secure their objective.

Now, our starting point is that the rational Scum Player would always choose the intrinsically more advantageous option.

The only point that saying "WIFOM" makes is that this is not always so. This is a valid and correct point. By always taking the intrinsically more advantageous (for the Scum Player's objective) option the Scum Player makes his actions predictable. By acting in the way that we would predict a Scum Player to act, the Scum Player makes their objective more obvious (and so is more likely to be caught). This provides a rational reason not to take the intrinsically more advantageous option. For example, if the policeman predicts that the criminal will choose the dark road then the policeman will follow down that road - which provides the criminal with a rational reason to take the otherwise less advantageous option. The same goes for mafia games. Take the example you gave of an apparently stupid counterclaim. The starting point is that the rational Scum Player would never do that, but we are here acknowledging that they in fact do have a rational reason to do it. Let's call this reason the "unpredictability reason".

So we now modify our starting position. It is not the case that the Scum Player will always take the intrinsically more advantageous option, because sometimes he might think that the unpredictability reason to take the intrinsically less advantageous option outweighs the difference. The Scum Player concludes that although one option might be intrinsically less advantageous, it is overall more advantageous. This is because of the unpredictability reason. This shows that the Scum Player is not in fact deliberately acting against their interests.

How often will the Scum Player come to that conclusion? We do not know. There is no logical way of working it out, as it will always depend on the particular context and the particular thought processes of those involved.

But we do know some conclusions that we have definitely not reached. We have not decided that the Scum Player will never do something apparently silly, as we realise that it's possible the Scum Player could rationally decide that it actually wasn't silly. We have also not decided that the Scum Player is entirely predictable, as we acknowledged the uncertainty in the previous paragraph. Thirdly, we have not decided that the Scum Player is incapable of making the calculation wrongly, because rationality doesn't equal infallibility.

I do claim that Scum Players are rational. This means they never deliberately do things disadvantageous to themselves. It is possible for them to make mistakes. It is possible for them to conclude that doing something apparently stupid for a Scum Player will advantage them by making it seem unlikely that they are a Scum Player (the unpredictability reason).

Lastly, I also claim that more than 50% of the time Scum Players place more weight on the intrinsically advantageous option than they do on the advantage gained by deliberately not choosing it. This doesn't always have to be true, but I think that across HG mafias it is true in practice. The exact percentage is arguable, but I think it's clear that more than 50% of the time Scum Players will choose the more intrinsically advantageous option.

This means that if the following conditions are satisfied, it indicates (but is not conclusive) that Player A is scum:
1. Player A has done a Thing.
2. That Thing is advantageous from a Scum perspective.

If you don't agree with this, then you're an idiot.

From this you can realise other things. If the following conditions are satisfied, it indicates (but is not conclusive) that Player B is scum.
1. Player B is scum unless a Fact is true.
2. If that Fact is true, then Unknown Scum did a Thing.
3. That Thing is disadvantageous from a Scum perspective.

These indications are stronger the more intrinsically disadvantageous the Thing is for Scum. That is why Scum 1 doesn't claim Cop with a guilty on Scum 2 at the start of the first day. The intrinsic disadvantage of doing this is massive, but the unpredictability reason isn't very strong.

I think this is about as bleeding obvious as I can make it.

The problem is that you're assuming that all players will agree on what is more intrinsically advantageous, and that all players are entirely rational. We have players at varying skill levels, and sometimes the "intrinsically advantageous" road is not apparent to the person calling the shots. Going back to Industry Mafia, we saw a lot of moves on both sides that were neither intrinsically nor overall advantageous.

Peter choose to use his kill on Night 1, but choose not to shoot at any of the three major suspects. From a town perspective, shooting at one of the three major suspects is more advantageous than taking a pot shot.

CK initially intended to revive Herr, which was to absolutely no gain whatsoever to the town at that stage, as Herr was incapable of harming scum. He changed his action to revive WRP, the doctor, which was better, but reviving Peter, the vigilante, would have been much better, as the mafia had already stolen town majority, and preventing kills would not take the majority back.

Barcaii's desperate counterclaim of Pulse was not wrought of the belief that it was too unpredictable to be called on. Barcaii lucked into town incompetence. It was an open and shut case. Of everyone, however, the doctor should have known that two Guardians and a Doctor made no sense, yet WRP protected Barcaii on Night 1, immediately after Pulse's lynch.

Towards the end of the game, Mole became paranoid, and instructed Gil not to use both kills that night, despite the fact that one of the kills would expire. While this didn't end up costing them the game, it delayed it without merit. Mole reasoned that Maffia's ability was an unknown, and he might be some kind of chaos role. However, no matter what Maffia's ability was, Mole had the ability to just block him and shut the action down.

Maffia's support of the lynch on Herr was literally one of the stupidest things I've ever seen a player do, and I've seen a few pretty dumb moves in my time. I thought the mafia had the game, but Barcaii was afk, and so there was still hope for the town. All Maffia had to do was unvote and the town could have won. Instead, despite previously noting that the vigilante was dead, and despite the fact that a scum group was pushing an overt speed lynch, he opted to help them out because "we can deal with them later". I literally facepalmed.

So, while it is safe to say that one thing is more likely than the other, one should never simply rule out all other possibilities simply because one is more likely. That is my point, here. Moreover, because scum are pretty much always better informed than town, we can never truly judge if an action they are making is more advantageous. We can't see the whole picture. This is why WIFOM is so important. Scum will not only sacrifice short-term gain for long-term gain, but there will be situations where scum appear to be sacrificing short-term gain when, in fact, they are playing with a loaded deck, as we saw with my Lie Detector gambit in Silmarillion.

Quoted from Rot:

The problem is that you're assuming that all players will agree on what is more intrinsically advantageous, and that all players are entirely rational. We have players at varying skill levels, and sometimes the "intrinsically advantageous" road is not apparent to the person calling the shots.

That is clearly a separate issue. I explicitly said that it was possible for players to be mistaken. WIFOM is not relevant where the argument is whether the player considered one option to be intrinsically advantageous. It is obviously a valid response to the statement "I would never do something as stupid as that" to say that "We have reason to think you didn't realise it was stupid". That is a separate and unrelated response to the WIFOM argument.

This is why WIFOM is so important. Scum will not only sacrifice short-term gain for long-term gain, but there will be situations where scum appear to be sacrificing short-term gain when, in fact, they are playing with a loaded deck

You are again conflating too issues. I explicitly allowed that it may be unclear in the thread which option was intrinsically advantageous. If it is not known then the WIFOM argument never arises. It is obviously a valid response to the statement "I would never do something as stupid as that" to say that "We have no reason to think your choice was intrinsically disadvantageous". That is a separate and unrelated response to the WIFOM argument.

WIFOM is the argument I gave in my previous post, and nothing else. Other valid arguments exist. It does not help to confuse the issue by calling everything WIFOM. At the moment you are using WIFOM as a kneejerk response that can apply whenever someone says something is likely. You might as well respond to things by saying "But maybe not!" and link to the Mafiascum page on WIFOM to show that you're right.

While they are separate arguments, it is impossible to discern which is valid at the time of making the argument. So rather than saying "It is possible that you sacrificed a short-term goal for a long-term goal (WIFOM), or that you made a mistake, or that the situation is not as clear to us as it may seem" every time someone makes the argument "I'm not stupid enough to do that," one could simple say "WIFOM" and have conveyed that entire point.

WIFOM is not inherently invalid, but it is inherently unable to support an argument on its own. It is essentially an incredibly weak argument. The problem I take with Mole's playstyle is that he basis his arguments almost wholly on WIFOM and random probability that he might as well have just pulled out of his ass. He doesn't just say "Why would I do that;" he says "I'm not stupid enough to do that, and therefore, I am confirmed town, should receive doctor protection, should not be investigated by cops, my opinion should carry more weight than everyone else's, and anyone who does not accept this is probably scum."

He doesn't just say "Why would I do that;" he says "I'm not stupid enough to do that, and therefore, I am confirmed town, should receive doctor protection, should not be investigated by cops, my opinion should carry more weight than everyone else's, and anyone who does not accept this is probably scum."

lol. but true.

WIFOM is not inherently invalid, but it is inherently unable to support an argument on its own.

Which is the point of this argument. Few things are conclusive, therefore you make assumptions and act upon them until something convincing shows its face. Ballpark your assumptions and you will usually be right. Just remember to refresh your assumptions every now and then.

The problem I take with Mole's playstyle is that he basis his arguments almost wholly on WIFOM and random probability that he might as well have just pulled out of his ass.

Hardly. I always prefer the more solid arguments, but yes, I am quite willing to go by an estimate based on rationality. WIFOM is something different. If you want to bring my theory down, that's fine. But call it something else. WIFOM has no possible way of discerning anything. Estimates based on rationality can easily be understood as logical. Just think about a situation, any situation, in which you can guess what someone will do. Such situations exist frequently. Therefore it is a reasonable short term argument to infer that someone would or would not have done something in the past if you need to make a theory.

He doesn't just say "Why would I do that;" he says "I'm not stupid enough to do that, and therefore, I am confirmed town, should receive doctor protection, should not be investigated by cops

Exaggeration and a stupid one at that. You saying this is clear evidence that you're not really addressing my ideas that have been laid out. For one, I repeatedly said that I'm not confirming anyone town. Just saying they're good enough for now. The second part about night actions sounds like you're just basing this off of Plants vs. Zombies, which is an isolated case and cannot be applied to every game I play. I rarely say things like this. PvZ is merely a rare case in which I believe that I, and the group, am extremely likely to be town relative to the other players, because we are a decently large group with a shared mechanic which inherently requires a similar numbered presence of other members, and that cannot be faked by uninformed scum. I don't think the group should be investigated. I didn't say anything about docs. But either way, it's just an isolated case. It doesn't often happen and is not representative of the general theory. If, for example, I were to be a mason, I would probably play in a similar fashion.

my opinion should carry more weight than everyone else's, and anyone who does not accept this is probably scum."

Hardly, although you can generally estimate that if you are town, you are more likely to be opposed by scum than town. This is not an argument that you can use to persuade people because it relies on information they do not have, but it is worthwhile for forming opinions. I would usually support this kind of argument from people because it creates notable relationships and gets more claims. A basic message behind this is that more claims are better. You don't need nearly as strong an argument as HG thinks you do. Stop being fussy about protecting information. It'll help in the longrun.

Tldr: WIFOM doesn't invalidate either instance, it is not a standalone argument outside of metagaming. If you believe there is a possibly of WIFOM it just means you should look for more information.

example: Rot is a high priority target for killing. It has been proven that he is less of a high priority target for protection. This is because kills often go through onto him night 0.

If Rot lives to day 1, do we assume the scum were scared off by the doctor? Okay, some information comes out. Rot becomes less of a target as a result.

If Rot lives to day 2, is he scum, or is that what scum want us to conclude from this? We can't conclude either from this alone, but we can consider which scum are playing us, because this is more than just a read on Rot. Who would have the forethought to attempt this in such a way, or is this really just a result of being scared away by the doctor?

If you ignore WIFOM, Rot being alive at day 2 means absolutely nothing to you. If you acknowledge it, you can come to several assumptions which can give you applicable leads should you lack them.

[This message has been edited by Peter Fallon (edited 06-13-2012 @ 03:26 PM).]

oh wow

if we assume failed kills are shown, rot becomes (in your scenario) more and more likely to be scum with no failed kills. However, if there are failed kills, he instantly becomes more likely town.

this changes with the setup, of course.

That's metagaming, not WIFOM. But WIFOM is a stupid concept, so I'll support that.

Thymole, I think you are employing a rather ropey idea of what "metagaming" is. It's perfectly possible to combine knowledge from outside the game (metagaming) with wondering whether someone has deliberately chosen the intrinsically disadvantageous option (WIFOM). There is a combination of both in the scenario where a Good Player remains alive when you'd expect him to be a high priority target.

By the way, the "How Not To Be Terrible" guide is supposed to be under gameplay.

Eh, I'll fix that...

Can you please remove the Flavor Guide? Seriously, it's really bad.

Well, we didn't want to say anything...

Some of the points in it are good. But you could cut some of the more bizarre spiritual mentor lines and it would be no worse off.

I quite like zen ash.

But he alternates with down-with-the-kids, street-wise, foul-mouthed Ash. And then sometimes "I'm insane, right? Guys? Right??" Ash drops by too. Whoever's narrating that guide breaks about half his own rules.

[This message has been edited by Julius999 (edited 07-29-2012 @ 09:19 PM).]

I call that epiphany ash. There is just so much brilliance that he can't contain it within a single personality.

Heh, sometimes I get bored so I come here to see ThyMole butchering the meaning of WIFOM :P

Oh also, my own guide if no one minds.

How to act dead/lynched:

A)You do not post.

B)You do not post.

C)You do not act like ThyMole did in PvZ

More to come soon.

-PKLordy

     In a short game a mafia player is fully able to remember most of anything that has occurred so far: their role, any claims, tells, what roles must be in the game (2 constant kills = two killers, someone complaining about a roleblock = scumblocker). Who pressured who and how that last bandwagon worked out. It is a trivial feat to accomplish over the course of 5-7 pages, even if the game were to take place over the course a month this information can be easily recalled.

     Through inactivity on parts of players or hosts, or high participation by these people following and recalling the game becomes difficult. If you recall 95% of the game, but your scenario is dealing with the unknown 5%, the would-be obvious solo player goes unblocked and the doctor is dead with no one the wiser.

     This is why it is important to take a step back and ensure your bases are covered. If pressure seems as if it will flip an easy lynch, is this really what you want? Someone has fumbled their claim - and yet the entire first day is a blur. Do you vote? No. You do not vote. You inform yourself - either by re-reading what you have missed, or taking notes.

     If a game is particularly active, I would recommend the note approach. Keep a text file somewhere and as the claims come in, store the post# to document this. List your first impressions, and other peoples' reactions - later in the game, these can become invaluable in finding the obvious scum.
     
     
     You want to have information if you want to find scum: scum are forced into lying. This SHOULD be balanced out by the town willing to cooperate - the extent of this will vary by game, but without a damn good reason to do so, town members don't lie. A dead townie's reactions can be superior to the argument of someone whose alignment is unknown: if the confirmed townie's reaction was reasonable, you should note this and hunt this unknown. If you are scum you need to know what to say.

     I am not able to explain every detail that must be recorded, but if the event was of value, and the discussion on topic there is always a motive behind what is said. If while uninteresting, an argument over theory between two established townies goes on for a sizable period of time, it is worthwhile to note the implications the argument were to have on the game, if applied to the game. One of these players could easily be a wolf in sheep's clothing. If scrutiny is applied to your target - while the chances are their existing reason to be perceived as town is sufficient, you at least explored an avenue where there was a chance of pressuring scum.

That's a good one. Especially on the simple note taking if you can't simply store all of the info in your head. My ability to simply remember facts from x pages ago has been very handy. I can't really tell you how to just remember things, so do what peter says and take notes. Contradictions sometimes occur and when they do they are both great fun to find and damning to expose.

How do you folks ensure that your games are balanced properly? Do you have a rule of thumb for the number of mafia in a game (eg 1:3 ratio)? Do you rely on the number of mislynches, or compare a best case and a worst case scenario?

I look at these things:

1. An absolute rule is that it should not be possible for any player to lose before there has been a single day. (Incidentally, this is among the reasons why faction-win is a must.)

2. I calculate a near-worst case scenario for the town as far as night actions go. The only favourable assumptions I make is that Vigilantes will choose not to shoot, and the like. I work out how many mislynches the town can afford in this scenario. The town absolutely always has to be able to afford at least one mislynch no matter how unfortunate the night actions are. The larger the game, the more mislynches the town should be able to get away with (although it doesn't scale linearly).

3. I calculate a worst case scenario for the scum (or each scum faction) as far as night actions go. If it's possible for all the scum (in a faction) to be dead or investigated with an incriminating result within a single night, that is almost certainly not acceptable.

4. Make sure that any individual players have a fair shot at winning the game, more or less on par with the larger factions. This will almost certainly mean giving them immunity to Cops and to night-kills.

I've left out obvious things that are more to do with creating a proper game than about balancing it, such as making sure that the scum have proper safe claims.

When thinking about balance you're thinking about two things: approximate chances of winning, and volatility. The second of these is often over-looked, but you can assess it relatively easily by calculating the worst case scenarios.

Quoted from Julius999:

1. An absolute rule is that it should not be possible for any player to lose before there has been a single day. (Incidentally, this is among the reasons why faction-win is a must.)

Such an absolute rule would preclude pretty much all third party factions. Survivors could not exist in such a system. I will say that I do support faction victory in most situations these days, because it promotes better team play and mostly prevents spite killings as the game winds down, but I wouldn't say that every player must absolutely not be able to lose on the first night. I would say that any town or anti-town player should absolutely not be able to lose on the first night, though, which means that some form of limited kill/investigation immunity is almost always appropriate on a serial role. But third parties, who I would consider to have somewhat easier win conditions, given that they are normally not in conflict with very many players, I do not generally grant such immunities to.

Quoted from Julius999:

2. I calculate a near-worst case scenario for the town as far as night actions go. The only favourable assumptions I make is that Vigilantes will choose not to shoot, and the like. I work out how many mislynches the town can afford in this scenario. The town absolutely always has to be able to afford at least one mislynch no matter how unfortunate the night actions are. The larger the game, the more mislynches the town should be able to get away with (although it doesn't scale linearly).

This is solid advice. Putting it more mathematically, I'd roughly estimate that, for every 5 total players, it should not be possible for town to reach Lynch-and-Lose for one additional day, so it should not be possible for a 10-player game to be at LyLo on Day 2, a 15-player game on Day 3, and so on.

Quoted from Julius999:

3. I calculate a worst case scenario for the scum (or each scum faction) as far as night actions go. If it's possible for all the scum (in a faction) to be dead or investigated with an incriminating result within a single night, that is almost certainly not acceptable.

This is a bit more difficult. This rule works alright for single-house games; the town should not possess enough kills and investigates on its own to accidentally nail the entire mafia on the first night, but when you have multiple anti-town killing factions, such as a mafia and serial or two mafias, it becomes a lot harder to adhere to this. I will admit, though, that multi-house games are generally a lot harder to achieve a balance as a result.

Quoted from Julius999:

4. Make sure that any individual players have a fair shot at winning the game, more or less on par with the larger factions. This will almost certainly mean giving them immunity to Cops and to night-kills.

Again, solid advice. A host should strive for a serial role to have close to if not even odds with the main factions. It is tempting for players to throw in serials and third parties to balance the other factions without considering how well they stand up on their own, and this is something that should be avoided.

-----

Probably the only thing Julius leaves out is a general rule of thumb for ratios. In a single house game, I generally shoot for 1 mafia per 3 townies, assuming both factions are of similar power saturation. If the mafia is less saturated than the town, I might go as far as 1 mafia per 2 townies, and if the town is less saturated, I might go 1 mafia per 4 townies. It's hard to really describe, because it's not just raw saturation, but consideration for how powerful each role is. I'd say that plotting worst case scenarios for each faction will probably give you a more accurate idea of where you stand.

On point 1: I'll stand by this. Survivors can perfectly easily be included in a night-start, provided that they are given at least one-shot Iron. Another option is to have a day start, so that everyone gets a chance to participate and influence whether or not they are targeted by night actions. If a host isn't willing to do either of those things, then maybe they shouldn't be throwing in Survivor roles for the hell of it. I agree with what you say towards the end of your post about avoiding the temptation to do this.

On point 3: This is a major part of the reason why multiple scum factions should be confined to games large enough to contain them while giving everyone a roughly even chance of winning. In the case of a lone Serial Killer, I think it's basically necessary to give them investigation immunity and at least one-shot Iron.

There's one more point I would like to add specifically:

5. A hidden balance concern is the overwhelming effect that a mass-claim can have in the town's favour if you've not structured the game properly. The game shouldn't really be about using night actions co-operatively in order to build a block of essentially confirmed townies and then to catch scum by process of elimination. I see that far too often and it turns the game into more of a crossword. I am of the view that it is strongly desirable to ensure that pursuing an early mass-claim strategy will be a net detriment to the town (assuming the scum competently use their safe claims).

Quoted from Julius999:

On point 1: I'll stand by this. Survivors can perfectly easily be included in a night-start, provided that they are given at least one-shot Iron. Another option is to have a day start, so that everyone gets a chance to participate and influence whether or not they are targeted by night actions. If a host isn't willing to do either of those things, then maybe they shouldn't be throwing in Survivor roles for the hell of it. I agree with what you say towards the end of your post about avoiding the temptation to do this.

Giving a Survivor a One-Shot Iron is like giving a Survivor a free win. With a win condition that let's them side with whomever, and disincentivizes either side from killing them, giving them a One-Shot Iron on the off chance they're shot on Night 0 basically means that, unless it is expended on Night 0 (which should be the only night the Survivor is not known to everyone in the current meta), the Survivor is basically unkillable.

I wouldn't mind giving some limited protection to some of the other third party roles that have harder win conditions, like say a Lyncher who only knows the character he needs to lynch, rather than the player, or something like that, but a Survivor with even limited protection is a cake walk victory for that player.

Quoted from Julius999:

On point 3: This is a major part of the reason why multiple scum factions should be confined to games large enough to contain them while giving everyone a roughly even chance of winning. In the case of a lone Serial Killer, I think it's basically necessary to give them investigation immunity and at least one-shot Iron.

Keep in mind that the thing that makes Point 3 difficult is the presence of multiple anti-town killers. Add a lone Serial to a game with a 3-man mafia, and there is now the possibility that all three members are killed or investigated on the first night. And 1-house plus SK games are fairly common these days.

Quoted from Julius999:

There's one more point I would like to add specifically:

5. A hidden balance concern is the overwhelming effect that a mass-claim can have in the town's favour if you've not structured the game properly. The game shouldn't really be about using night actions co-operatively in order to build a block of essentially confirmed townies and then to catch scum by process of elimination. I see that far too often and it turns the game into more of a crossword. I am of the view that it is strongly desirable to ensure that pursuing an early mass-claim strategy will be a net detriment to the town (assuming the scum competently use their safe claims).

This is actually a really good point, though I think you make it for the wrong reason. While it is important to ensure that town cannot create a super bloc via mass claiming, I think the bigger danger than night plans for roles is the fact that a lot of people want to host games in themes with limited characters. Consider a Fellowship of the Ring Mafia, where town are the Fellowship and scum are the forces of evil. It's really tempted to say that the town has all nine members of the Fellowship, but when it comes time to claim, it will be really obvious which players are town and which are scum. I bring this up because it really happened once. The town created a super bloc of flavor confirmed players and proceeded to shut down what would otherwise have been a tough game, with two mafia houses and two serials.

This is probably a good time to talk about safe claims. There are two trains of thought regarding safe claims. The first is to give a scum faction one less safe claim than they have members, but make all given safe claims strong. The second is to give a scum faction one more safe claim than they have members, but to make some of the safe claims weak and some strong.

I've personally been using the latter of late. When I go to build a game, I map out all the characters that could possibly be in, I decide who scum will be and how large they will be, and then I start by taking out characters for safe claims to ensure that I give a fair group of both strong and weak safe claims. Only then to I start selecting which characters will be town. For me, hosting games in a source material with a decent cast of characters, this isn't so hard, but when making a game with a limited cast, it is even more important to consider safe claims. Not just the ones you're providing, but possible fake claims altogether. If there are fewer possible fake claims than there are scum, a mass claim will break the game.

About Survivors: First and foremost, I think giving the Survivor an easy time is a lesser evil than giving them the chance of an instant, faultless loss. Second, I think the game needs to be constructed in such a way that the Survivor doesn't have an easy time; don't just add a standard Survivor for no reason. Third, I don't think towns are quite so favourable to Survivor claims as you seem to suggest; if they are, then it is poor town play.

Add a lone Serial to a game with a 3-man mafia, and there is now the possibility that all three members are killed or investigated on the first night.

Only if the town has two investigative roles (or one plus a lucky vigilante), I think. And really, considering such a scenario merely demonstrates the importance of the rule and the consequent need not to go overboard with investigative roles and killers.

This is actually a really good point, though I think you make it for the wrong reason. [Stuff about broken themes.]

I agree with all that you say - I gestured towards it in my reference to safe claims in Post 228. However, I wanted to make the point separately from the issue of broken themes. Even flavour-less combinations can cause the problem. Broken themes are especially egregious, but not the only way it can happen.

This is probably a good time to talk about safe claims.

Yes, although it's an issue distinct from balance. It's about hosting a mafia game rather than a game of flavour-claim analysis.

I think it's hard to go wrong with giving the scum too many and too strong safe claims. It's not really enough just to give them some plausible characters. In order that they have the proper flexibility to claim different role abilities, they need a variety of plausible characters so that they can pick and choose the most suitable.

[This message has been edited by Julius999 (edited 07-26-2015 @ 02:35 PM).]

http://aok.heavengames.com/cgi-bin/aokcgi/display.cgi?action=ct&f=22,41943,,10

Guys, I'm hosting a classic vanillaish mafia at AOKH. Feel free to drop by.

Quoted from Julius999:

About Survivors: First and foremost, I think giving the Survivor an easy time is a lesser evil than giving them the chance of an instant, faultless loss.

Better to have the minimal chance of faultless loss than nearly guaranteed victory.

Quoted from Julius999:

Second, I think the game needs to be constructed in such a way that the Survivor doesn't have an easy time; don't just add a standard Survivor for no reason.

Any suggestions for how to do that?

Quoted from Julius999:

Third, I don't think towns are quite so favourable to Survivor claims as you seem to suggest; if they are, then it is poor town play.

If you believe a Survivor claim, lynching one is a wasted lynch. Nothing is gained from lynching a Survivor. The only reason to lynch a Survivor is if you don't believe the claim. Generally, this means that claimed Survivors will be lynched early, generally due to an absence of better suspects, which is just as much a faultless loss as being killed early, or they aren't going to be lynched at all.

Quoted from Julius999:

Only if the town has two investigative roles (or one plus a lucky vigilante), I think. And really, considering such a scenario merely demonstrates the importance of the rule and the consequent need not to go overboard with investigative roles and killers.

What, Vigilantes don't count? Just because Vigilantes shouldn't shoot Night 0 doesn't mean they won't. If we're ensuring that a mafia can't be discovered/eliminated Night 0, Vigilantes should count just as much as any other killing role.

At the end of the day, there are going to be incredibly small chances that a faction has a really bad day. There are certain random elements to mafia, the Night 0 targets being the best example of this, that are simply unavoidable if you don't want to run a game where 9/10 of the town are boring VTs, and the mafia just all claim VT, and everyone winds up lurking because there's nothing to talk about.

Quoted from Julius999:

I agree with all that you say - I gestured towards it in my reference to safe claims in Post 228. However, I wanted to make the point separately from the issue of broken themes. Even flavour-less combinations can cause the problem. Broken themes are especially egregious, but not the only way it can happen.

It's a lot harder to create a confirmed town super bloc without flavor. Getting confirmed town under a good host should be difficult. Staying alive once confirmed town even harder.

Quoted from Julius999:

Yes, although it's an issue distinct from balance. It's about hosting a mafia game rather than a game of flavour-claim analysis.

I think it's hard to go wrong with giving the scum too many and too strong safe claims. It's not really enough just to give them some plausible characters. In order that they have the proper flexibility to claim different role abilities, they need a variety of plausible characters so that they can pick and choose the most suitable.

Claim analysis is an integral part of mafia, because it creates something to discuss in a game where there is otherwise nothing to discuss. Without initial discussion, you can't analyze behavior, because there's been no behavior to analyze.

I also dislike handholding for scum. Safe claims are important, but give too many or too strong, and scum will breeze through the game. If scum have foolproof claims, it isn't hard to act like town, because town are either going to be lynching randomly because scum aren't under any real pressure when asked to claim, or playing follow-the-cop, and shutting down follow-the-cop isn't hard in a well balanced game, because follow-the-cop otherwise is broken.

Any suggestions for how to do that?

Yes. Give someone or some faction a reason to want them dead. Give them a win condition that isn't simply to survive. Basically, don't include in your mafia game a player who is neither town nor scum unless you can think of something interesting.

What, Vigilantes don't count? Just because Vigilantes shouldn't shoot Night 0 doesn't mean they won't. If we're ensuring that a mafia can't be discovered/eliminated Night 0, Vigilantes should count just as much as any other killing role.

Yes. So don't include a 3-man mafia in the same game as a SK, Cop and Vigilante. Honestly, I think you're just illustrating ways in which my suggested rule helps avoid defaulting into setups that are mediocre or worse.

There are certain random elements to mafia, the Night 0 targets being the best example of this, that are simply unavoidable if you don't want to run a game where 9/10 of the town are boring VTs, and the mafia just all claim VT, and everyone winds up lurking because there's nothing to talk about.

Again, this is a good reason not to have Night 0 and to start with Day 1. It's perfectly possible to design a game that complies with my rules and has zero vanilla townies. You just have to avoid reaching for the Doc/Cop/Vigilante box straightaway.

Claim analysis is an integral part of mafia, because it creates something to discuss in a game where there is otherwise nothing to discuss. Without initial discussion, you can't analyze behavior, because there's been no behavior to analyze.

Of course there's something to discuss: choices. Who sides with whom. Who lynched the townie. Who defended scum.

Even when you have claims to analyse, you can analyse the choice and not the flavour. Isn't that Miller Vigilante claim convenient?

Safe claims are important, but give too many or too strong, and scum will breeze through the game.

If the town are morons, or if the game is not properly balanced for scum who aren't going to be tripped up by flavour.

because town are either going to be lynching randomly because scum aren't under any real pressure when asked to claim, or playing follow-the-cop, and shutting down follow-the-cop isn't hard in a well balanced game, because follow-the-cop otherwise is broken.

Yes. If the town are morons.

Mafia isn't a puzzle. It's a social game. It should be about what the players say and do, rather than about how likely it is that Mr Tickle is a Role Cop.

Quoted from Julius999:

Yes. Give someone or some faction a reason to want them dead. Give them a win condition that isn't simply to survive. Basically, don't include in your mafia game a player who is neither town nor scum unless you can think of something interesting.

I don't feel like the utter abandonment of the third party as a concept is a feasible solution to third parties sometimes being at risk of losing early.

Quoted from Julius999:

Yes. So don't include a 3-man mafia in the same game as a SK, Cop and Vigilante. Honestly, I think you're just illustrating ways in which my suggested rule helps avoid defaulting into setups that are mediocre or worse.

And I feel the opposite. In a game with two killing factions, the town needs more than a single investigator to balance out that they're potentially being killed twice as quickly. To arbitrarily decide that they can't have more than one investigator/killing role total because then there happens to be an astronomically low chance that a scum faction might die if the heavens roll back and the stars align is ludicrous. In all the games I've played, which have pretty much to a man broken this rule, I think I've seen an entire scum faction wiped out on Night 0/Day 1 a single time, and that was in a game that broke the rule to an extreme (there were like five kills on the first night; the game likely wasn't balanced).

Quoted from Julius999:

Again, this is a good reason not to have Night 0 and to start with Day 1. It's perfectly possible to design a game that complies with my rules and has zero vanilla townies. You just have to avoid reaching for the Doc/Cop/Vigilante box straightaway.

This is the same mentality that drove MafiaScum to develop "Cop Head Start", a system wherein mafia doesn't get to act on Night 0. Subsequently, and surely in unrelated news, almost all existing prebuilt setups created by MafiaScum that this was applied to suddenly saw a much larger percentage of town wins to mafia wins, to the point that many of the setups were abandoned as being unbalanced.

Day start is a balance decision, because it gives additional power to the town. Not as much as Cop Head Start, but still some power. It isn't just something you haphazardly throw into a game because one player is probably not going to have maximum fun during this game otherwise. Spoilers: getting lynched on Day 1 sucks too, even with a night start. And if town no lynches, getting killed the subsequent night is really no different than getting killed during a normal Night 0. People die in mafia, and people will generally be salty that they died. It's part of the game. You can't have a competitive game if you hold everyone's hands.

And really? I guess you could use passives, but then you need to consider the repercussions of those passives. I mean, Iron is already a difficult role to balance for, Noble is practically a confirmed towny if you do vote counts, and Mason is one step away from creating a powerful town bloc, but beyond that, a bunch of random passive towny claims is just about as meaningful for town as a bunch of VT claims. I mean, with 1 investigator, 1 protector, and 1 blocker, you still have 6 town roles to fill. A Mason group, a Noble, and an Iron towny would probably be a bit much for a 3-man mafia to handle. You might be able to get away. And random passives aren't that hard for mafia to fake claim as either. With no actions to fake and no fake actions to try and keep track of, it's really a lot easier. "Yes, I'm Iron. Good luck testing it, since town doesn't have a Vig or equivalent role. As long as I'm not counterclaimed by an actual Iron towny, my claim is untouchable."

Quoted from Julius999:

Of course there's something to discuss: choices. Who sides with whom. Who lynched the townie. Who defended scum.

And how do you make choices? I mean, without claiming, what do you use to determine a lynch on the first day? Dice? Can you really then look back at what was ultimately random and fault anyone that was on that bandwagon? "You sided against the towny that we all agreed to lynch because the dice gods demanded it! You must be scum!" "Actually, all of us sided against that towny, because we had nothing better to go on, because we've removed one of the most integral parts of the game."

Trust me, I've been to several communities that don't do claiming at all. It was boring as shit. The day phase was basically a bunch of idle, irrelevant gossiping, and then the town got together and lynched the person they liked the least. Not liked the least because they seemed scummy, but liked the least on the forums as a whole. It was just a big popularity contest. Just a series of shitposting and circle jerking followed by a random bandwagon.

Quoted from Julius999:

Even when you have claims to analyse, you can analyse the choice and not the flavour. Isn't that Miller Vigilante claim convenient?

There is definitely some degree of analysis that can be applied to role claims, but only to some degree. Most claims aren't going to be as painfully obvious as the Miller Vigilante of old. Especially in games where you've stripped the town of power roles because you're afraid that the .003% chance that the 3-man mafia be Night 0 eliminated in a 13-man game (probably actually lower, since most Vigs won't fire on Night 0), the mafia can just claim as VTs. How do you analyze a VT claim, when there's eight of them? "Ya, scum is definitely hiding among the VTs...but since all the claims are completely identical, and we've basically had nothing to base a lynch on and therefore no proper behavior to analyze...I guess we're buggered."

Quoted from Julius999:

If the town are morons, or if the game is not properly balanced for scum who aren't going to be tripped up by flavour.

See the above argument. Flavor claiming is an important part of the game, whether the game be vanilla or role madness. Giving the scum a completely free pass on flavor claiming is just as bad as giving them too powerful of abilities. Fake claiming really isn't that hard if you haven't been caught in a lie by an investigator, and since Follow the Cop is both a terrible practice for town, incredibly boring when it works, and really easy to defeat with one scum blocker or redirector, the alternative is hardly viable.

Quoted from Julius999:

Yes. If the town are morons.

Then what is your alternative? Your one example of how town can catch mafia in your ideal world where flavor might as well not exist and the town has been heavily disempowered to avoid fringe scenarios involves mafia being morons. Who the hell claims as a Miller Vigilante anymore? I don't think that fake claim saw use after 2008, because it was basically always lynched on the spot. What is the town's strategy on Day 1? They've got basically nothing to go on, there's basically nothing to discuss, and all the claims are going to be about the same. I mean, I guess you could just give the town only major roles and force the scum to claim weird or minor powers to avoid counterclaims, but that's just as bad as giving the town all the major characters and forcing scum to only claim minor characters to avoid counterclaims.

Quoted from Julius999:

Mafia isn't a puzzle. It's a social game. It should be about what the players say and do, rather than about how likely it is that Mr Tickle is a Role Cop.

And without some minor puzzle aspects to generate behavior, there's nothing to say or do. It's Day 1. You've gotten three claims so far, all VT. Flavor is ignored. Who do you lynch? Well, you pick one at random, because there's been nothing yet to base behavior on. I mean, it's Day 1. So you lynch one at random. But that isn't a decision with a right or wrong answer, so you can't go back and look at it later and say "these people made the wrong choice, and may be scum" because there was no wrong choice. Same goes for selecting the people to claim. On Day 1, that's usually done at random, or on the very slightest of pretenses. You can't go back and say "you forced x to claim, when y was clearly the better choice", because there is no objectively better choice. Day 2 begins. Since you had no informed decisions yesterday to analyze, Day 2 plays out the exact same way. Unless an investigator has damning evidence, claims are chosen at random, and since the town consists primarily of VTs, most of those claims are VTs. No sane scum is going to actually suggest you lynch the claimed Cop or claimed Doctor or claimed RB or whatever, so you lynch one of the claimed VTs. Which one? Well, the claims are all perfectly identical, so you basically choose one at random. But wait, that choice has no objectively better or worse options either, so now it's Day 3, and you're in the exact same buggery place as before. There's no real behavior to analyze, because there's been basically nothing to discuss. All scum has to do is go with the flow, and occasionally be the ones to select at random which VT to lynch, and they're golden. Town might accidentally happen upon scum, but at this point, it's no longer a game of skill, but a game of complete luck, and 9 times out of 10, the informed minority wins that game. That's not a social game. That's barely even a game.

I don't feel like the utter abandonment of the third party as a concept is a feasible solution to third parties sometimes being at risk of losing early.

That is not what I suggested. You may recall that in my own games I have included a number of third parties, and yet I didn't go down the route of vanilla Survivor.

In a game with two killing factions, the town needs more than a single investigator to balance out that they're potentially being killed twice as quickly. To arbitrarily decide that they can't have more than one investigator/killing role total because then there happens to be an astronomically low chance that a scum faction might die if the heavens roll back and the stars align is ludicrous.

I don't agree with spamming investigative roles as a solution. If you have two killing factions then what the town needs is more people.

This is the same mentality that drove MafiaScum to develop "Cop Head Start", a system wherein mafia doesn't get to act on Night 0.

Although you'll notice it drove me in a completely different direction. Cop Head Start is a bad idea because it makes investigative roles even more important.

It isn't just something you haphazardly throw into a game because one player is probably not going to have maximum fun during this game otherwise.

It is, for me, one of the assumed features around which the game is balanced. I think that is generally better than seeing it as a feature which can be tweaked to assist with the balance.

Spoilers: getting lynched on Day 1 sucks too

True, but at least you got some real involvement in the game. Most of the time your lynch will be attributable to something you did that wasn't advisable.

You can't have a competitive game if you hold everyone's hands.

Clearly not what I am advocating. I am for giving everyone a fair shot so that they can participate in the competitive game.

And really? I guess you could use passives...

The only limit is your imagination. Make up new abilities. Thymole's "propagandist" role who could adjust the number of votes needed to lynch the target was a good example.

And random passives aren't that hard for mafia to fake claim as either. With no actions to fake and no fake actions to try and keep track of, it's really a lot easier. "Yes, I'm Iron. Good luck testing it, since town doesn't have a Vig or equivalent role. As long as I'm not counterclaimed by an actual Iron towny, my claim is untouchable."

I suppose the town will have to do some actual work then.

And how do you make choices? I mean, without claiming, what do you use to determine a lynch on the first day? Dice?

Pick on the good players. Pick on the weak players. Pick on the inactive players. Speculate about people's motivations for choosing one or the other of these options. Pick on anyone who suggests random lynching. You need the mindset of suggesting things and then holding people responsible for their suggestions.

How do you analyze a VT claim, when there's eight of them? "Ya, scum is definitely hiding among the VTs...but since all the claims are completely identical, and we've basically had nothing to base a lynch on and therefore no proper behavior to analyze...I guess we're buggered."

If eight people claim VT in a row, do you think the first of them is more or less likely to be telling the truth than the last one? That's your starting point.

Seriously, you don't have to resort to pure randomness immediately after finding that there's nothing concrete to go on.

Can I persuade anybody to join this mafia game at AOKH?

It develops a room mechanic used by Draconian Devil a few years ago and is hopefully something a little different to the normal games.

^What he said.

The mafia community at AoKH is still active, and I've got a game going. Anyone is welcome to join.

[ Message deleted by Pecunia at 11-22-2022 01:04 PM ]