In Soviet Russia, Mafia plays you! (Forbidden Thread for Glasnost Mafia -- No Players!)

[MHaye]

@USCDiver

I don’t believe I would have said that…regardless of my alignment.

Neither would I, but that’s primarily because I saw what happened last time.

In Blaster master’s Cult of Sekham, Hal Briston posted in a Night thread that he couldn’t wait for the Night to end. This was because he had proof positive that Hockey Monkey had told a lie and wanted to expose her lie for all to see and prove she was a Cultist.

The actual Cultists correctly deduced that this was because he was the Cop, and killed him. Unfortunately the protectors didn’t also follow that logic.

The kicker? HM was not a Cultist. She had lied to hide her power. If Hal had lived, HM might have realised her ambition to be recruited (assuming she survived the lynch wagon).

[Tom_Scud]

Now that Idle has come forward as receiving a “chicken” message, the day is basically over; might as well end it.

[Pleonast]

@Tom_Scud

Huh. Just realized that Town could totally break this game:

One mason identifies him/herself (with the implicit threat that one of the two hidden masons could vouch for him if there was a counterclaim). Everyone EXCEPT that mason then is required to vote for him/herself. The jailer (still hidden) jails the mason.

This leaves the scum and the bomber with the option of (1) playing along while hunting for the jailer or (2) trying some sort of last-minute unvote, which has the advantage of giving the possibility of an instant win but the disadvantage of leaving one of their number exposed to retaliation if everyone doesn’t die; depending on how quick people’s reaction times were, they might even get lynched the same day (if say they posted 1 minute ahead and 3 Town posters saw the unvote and voted the unvoter).

Good catch, Tom. It shows how powerful a multiple lynch is.

But I’m not sure it actually breaks the game, because the multiple lynch only has a 1 in 8 chance of occurring. So for the first several Days, scum and bomber will be able to act without much risk. Basically, each Day, town gets a chance to win the lottery, otherwise scum get a free action. Extremely swingy, but not obviously unbalanced in favor of either side.

[Tom_Scud]

Well, if the Scum (a) don’t try to disrupt things with unvotes and (b) otherwise do everything perfectly, Town still has a 41% chance of winning, since they get 4 Days to try

(Night 1: Scum kills jailer. Night 2: Scum kills Mason 1. Day 3: Mason 2 and 3 claim. Nights 3 and 4: Scum kills Masons 2 and 3. Bomber could speed this up if he happened to pick Masons 2 or 3 at random before they claimed).

OTOH, Scum messing around with last-minute unvotes would complicate matters - if nothing else, it would give them extra Nights to go fishing for the jailer and the extra Masons. Also, the game would end up deadly boring, would take forever to finish, and would be nothing at all like Mafia.

[Pleonast]

Tom, I ran the numbers and it looks like you’re right. The “half-life” of the all-kill lynch is about 5 Days (it has about even odds of occurring in any 5-Day interval). Scum need 17 players dead. If only scum and bomber are killing, that’ll take 9 Days, which is almost 2 lotto half-lives. That’s just over 1/4 chance to win for scum. Not fair.

@sachertorte

Frankly, the fact the Pleonast turned up scum is largely coincidence. I suppose I could retroactively look back and see clearer signs (e.g., Pleonast never addressed the statistic that showed he was a good lynch candidate based on his Vanilla Claim) but even that is iffy. Being wrong and annoying about being wrong isn’t a scum tell in my opinion.

Good catch sach, me not addressing that statistic probably was a scum tell. Unconfirmed vanilla town are always preferable lynch candidates unless there is strong indications that another player is scum. It’s especially true in this game. I probably would have talked about the point if I weren’t scum.

Actually, that gives me an idea for a general town strategy. Each Day, one vanilla town truthfully claims such, and is lynched. Town avoids exposing power roles. It’s then a race of building the confirmed pool versus scum finding the power roles. Who wins will depend on the powers in the game, and a little bit of luck. The town should not be able to guarantee a win with this tactic if the game is balanced.

[Chronos]

Actually, that gives me an idea for a general town strategy. Each Day, one vanilla town truthfully claims such, and is lynched. Town avoids exposing power roles. It’s then a race of building the confirmed pool versus scum finding the power roles. Who wins will depend on the powers in the game, and a little bit of luck. The town should not be able to guarantee a win with this tactic if the game is balanced.

I really, really can’t see Town winning with this strategy, since it gives Scum the means to avoid getting lynched for a very long time. And everyone in Town would end up giving Scum information about who the power roles are: Anyone who posts during the Day (any Day) before the first vanilla claim would thereby be considered less likely to be a power role. Plus, of course, Scum can also hunt power roles using their usual method of Scum investigator (which seems pretty common nowadays), looking for folks who seem to be playing a little differently, or just shooting into the crowd. And your first N days would be essentially devoid of information in the voting record, so if the Scum did manage to kill off the Detective before he claimed, Town would have nothing at all to go on.

Tom Scud, sachertorte has a pretty good point about the hammer determining when the next Day will end (if it’s not also hammered). That’d be a valid reason to keep the Day going a bit longer.

[Doctor_Who]

@Pleonast

Actually, that gives me an idea for a general town strategy. Each Day, one vanilla town truthfully claims such, and is lynched. Town avoids exposing power roles. It’s then a race of building the confirmed pool versus scum finding the power roles. Who wins will depend on the powers in the game, and a little bit of luck. The town should not be able to guarantee a win with this tactic if the game is balanced.

If you proposed this in a game I was playing with you, I would honestly probably vote for you. Seems like a scum-friendly strategy that asks Vanilla Town to abdicate their ability to lynch scum and hope and pray the power roles save the day. Blech.

ETA: on preview, looks like Chronos has some similar issues.

[Pleonast]

@Chronos

Anyone who posts during the Day (any Day) before the first vanilla claim would thereby be considered less likely to be a power role.

Can you explain why this would be?

[Chronos]

Well, how would you decide which vanilla ends up being the one to claim each Day? Even if it’s left tacit, some method is bound to emerge. And the simplest method is that the first vanilla to post each day is the one who claims.

[Pleonast]

@Doctor_Who

If you proposed this in a game I was playing with you, I would honestly probably vote for you. Seems like a scum-friendly strategy that asks Vanilla Town to abdicate their ability to lynch scum and hope and pray the power roles save the day. Blech.

Whether it is scum friendly or not depends on the powers in the game. Voting for players for simply stating their ideas is not helpful to the town, especially if it’s a closed game, there’s not enough information to evaluate tactics.

It is deceptively easy to design a game where the town can accumulate confirmed townies rather quickly. All it takes a several masons, an alignment investigator and one or two self-confirming roles (someone who survives a mislynch and is confirmed, or someone claiming a role that is confirmed if there is no counterclaim) and a reduction in size of the unconfirmed pool and the town wins.

For example, let’s take a game of 20 players. Town has an alignment investigator, 3 masons and scotsman who survives a lynch and is mod-confirmed. Remaining 10 town are vanilla. Scum have a blocker, a godfather, and three vanilla scum. That doesn’t seem too unbalanced at first glance.

On Day One, we have 5 “confirmed” townies, 10 “unconfirmed” townies, 5 scum and 5 mislynches available. After one cycle, if we assume a mislynch and kill of vanilla town, and investigated player added to the confirmed pool, we have 6 confirmed, 7 unconfirmed, 5 scum, 4 mislynches. Day Three is 7, 4, 5, 3. Day Four is 8, 1, 5, 2. There are fewer unconfirmed townies than mislynches remaining, which is an automatic town win. And there’s one spare mislynch to account for one of the confirmed killed by scum or the godfather being “confirmed”. If there’s any additional confirmable roles, and all it takes is someone with a role that’s unlikely to be scum, town’s chance improves. I would not call this setup balanced.

Designers should always check that this is not easy to do. My tactic of one vanilla town claiming per Day and then lynching them is way for town to improve their chances of getting a confirmation win. Whether or not it works depends on the details, but this tactic should be taken into account when designing the game.

[Pleonast]

@Chronos

Well, how would you decide which vanilla ends up being the one to claim each Day? Even if it’s left tacit, some method is bound to emerge. And the simplest method is that the first vanilla to post each day is the one who claims.

Well, that would be a bad method, then.

I was thinking along the lines that each vanilla townie consider to claim or not with each post, if no one else has claimed. And leave it vague like that. Eventually, someone will claim.

It could be formalized by having each vanilla townie secretly pick a random time (using random.org) between the start of the Day and maybe 12 hours before the end. If their time is reached and no one has claimed, then they do so at that point.

[Doctor_Who]

@Pleonast

Voting for players for simply stating their ideas is not helpful to the town, especially if it’s a closed game, there’s not enough information to evaluate tactics.

In principle I agree, because I like discussion amongst the town. But it is important to realize that during that discussion Scum members are going to state ideas that are NOT helpful to town. Your idea, to me, seems so anti-Town, so pro-Scum, that it would invoke a powerful feeling of distrust if you were to suggest it in a live game. It’s similar to this: Player X says on Day One, “Let’s have all the power roles claim right now! Then we can be sure we don’t mislynch them.” Player X is just stating an idea. Player X has also locked up my vote (and probably most everybody else’s).

On Day One, we have 5 “confirmed” townies, 10 “unconfirmed” townies, 5 scum and 5 mislynches available.

See, you lose me here. On Day One, everyone is unconfirmed. Confirmation takes time and information. And under your strategy, while VT is sitting around offering themselves up like Christmas Geese, the scum are actively hunting those power roles. What happens when they kill the town investigator on Day Three and his results die with him? What happens when scum kills another of your “confirmed” players? I’m guessing VT goes, “shit… we should have been trying to lynch scum instead of lying back with our throats exposed.”

I guess with that, I will agree to disagree here. I think maybe we just have a profoundly different approach to the game.

[Tom_Scud]

You’re also assuming that the investigator will never investigate one of the other “confirmed” or someone who subsequently offers him/herself up as the vanilla sacrifice du jour.

[Pleonast]

@Doctor_Who

In principle I agree, because I like discussion amongst the town. But it is important to realize that during that discussion Scum members are going to state ideas that are NOT helpful to town. Your idea, to me, seems so anti-Town, so pro-Scum, that it would invoke a powerful feeling of distrust if you were to suggest it in a live game. It’s similar to this: Player X says on Day One, “Let’s have all the power roles claim right now! Then we can be sure we don’t mislynch them.” Player X is just stating an idea. Player X has also locked up my vote (and probably most everybody else’s).

And that is why voting based on ideas (not actions) is bad. Especially in a closed game where it’s difficult to determine if an idea really is pro- or anti-town. There is no reason to think Player X is more likely to be scum than any other player. In fact, if one takes into account the your reaction, one could plausibly say scum would be less likely to state anything that could construed as controversial.

See, you lose me here. On Day One, everyone is unconfirmed. Confirmation takes time and information.

On Day One everyone is unconfirmed only because they have not claimed. For example, a mason can confirm themself at will. My “confirmed” count is the number of players who could be reasonably considered town at that point, if everyone claimed.

And under your strategy, while VT is sitting around offering themselves up like Christmas Geese, the scum are actively hunting those power roles. What happens when they kill the town investigator on Day Three and his results die with him? What happens when scum kills another of your “confirmed” players? I’m guessing VT goes, “shit… we should have been trying to lynch scum instead of lying back with our throats exposed.”

Yes, it’s possible that scum could get lucky. They only get 3 kills though, before the town can do a mass claim and win by simply lynching everyone who is not confirmed. And the numbers show that in most cases, the town will win. That extra mislynch covers for a lot of different potential things that could go wrong for the town.

I didn’t even include a town protector or a town killer; either of those will drastically improve the town’s chances. Besides providing extra confirmable roles, by reducing the chance of a Night kill, or by more quickly eliminating unconfirmed players and conflicting claims.

I guess with that, I will agree to disagree here. I think maybe we just have a profoundly different approach to the game.

Don’t conflate the approach to playing the game with designing a game to be balanced. While you might not like to play the vanilla-sacrifice tactic, it’s existence should be accounted for by the game designer. And even if the town doesn’t take this tactic, they could still use it circumstantially (simply by lynching only vanilla town without forcing any power roles to claim).

The bottom line is, if you were scum in the setup I outlined above, would you feel like you had a fair shot at winning the game? Even if the town started sacrificing vanilla townies, and your only chance of winning was in your three kills, either killing the town investigator or hoping to reduce the number of “confirmed” by two before the mass claim on Day Four?

[Chronos]

Yes, it’s possible that scum could get lucky. They only get 3 kills though, before the town can do a mass claim and win by simply lynching everyone who is not confirmed.

They only get three kills if absolutely nothing goes wrong, but it’s almost guaranteed that something will go wrong. If the Detective investigates any of the other Town powers, it buys Scum more time. If the Detective investigates a vanilla and then that vanilla goes up for the sacrifice, it buys Scum more time. If the Scum hit any of the confirmable roles (not just the Detective), it buys Scum more time. If the Scum hit anyone the Detective has investigated, it buys Scum more time. If the Detective claims early to prevent an investigated Townie from dying, the Scum kill him immediately. And if Town can’t hit the confirmation win, they can’t win at all, because they’ve actively denied themselves any other avenues.

The bottom line is, if you were scum in the setup I outlined above, would you feel like you had a fair shot at winning the game? Even if the town started sacrificing vanilla townies, and your only chance of winning was in your three kills, either killing the town investigator or hoping to reduce the number of “confirmed” by two before the mass claim on Day Four?

No, I wouldn’t feel like I had a fair chance, but I’d be laughing all the way to the bank at the Town for making it so unfair in my favor.

[Pleonast]

@Chronos

They only get three kills if absolutely nothing goes wrong, but it’s almost guaranteed that something will go wrong.

No, they only get three kills if no more than one thing goes wrong (except for the case that the investigator gets killed). Town has a spare mislynch that will accommodate almost any single thing going wrong. Scum need two things to go wrong. Run the numbers, it’s not likely in only three cycles.

And two things going wrong isn’t game over. It puts the town at lynch or lose, having to chose scum from an unconfirmed pool with only one townie in it. Something like 5 scum in a pool of 6 players. Even with blind lynching, town has a much better chance of winning than scum.

Only if three things go wrong in those three cycles will have town lost on Day Four.

No, I wouldn’t feel like I had a fair chance, but I’d be laughing all the way to the bank at the Town for making it so unfair in my favor.

Town doesn’t have to explicitly follow the sacrificial vanilla tactic to win on Day Four. The tactic is simply one way to reach the confirmation win. Surely you can see that a mass claim on Day Four will put the town in a highly advantageous position.

[Chronos]

Don’t forget that one of the “confirmed” players in your setup needs a half a mislynch to confirm him, too.

[Pleonast]

@Chronos

Don’t forget that one of the “confirmed” players in your setup needs a half a mislynch to confirm him, too.

Not really. There’s “confirmed” meaning “absolutely certain they’re town” and there’s “confirmed” meaning “we won’t lynch this person until the unknown pool is eliminated”. I’m using the latter, because that’s all that’s needed trap the scum in the unconfirmed pool.

A claimed scotsman is not a good lynch candidate until the unconfirmed pool is gone. If scum want to play risky, they can falsely claim scotsman, but that’s simply forces the town to make a half-mislynch vs freebie scum lynch coin flip. The expectation value of that trade is not in the favor of scum.

Falsely claim mason is only viable in a lynch-or-lose situation, I think. Falsely claiming investigator is more practical, but it has a cost. Not only is there the one-for-one trade that’s not helpful for scum on average, it also prevents the town from making mistakes with respect to the investigator’s results, when it’s counterclaimed.

[Pleonast]

@Pleonast

Something like 5 scum in a pool of 6 players. Even with blind lynching, town has a much better chance of winning than scum.

Whoops, that’s backwards. With blind lynching, town only has a 1 in 6 chance of winning. But the lynching won’t actually be blind; town will probably have the results of several lynches to help decide who is scum or the last unconfirmed town.

[Chronos]

But the lynching won’t actually be blind; town will probably have the results of several lynches to help decide who is scum or the last unconfirmed town.

They would if they’d been playing conventionally. But if they’ve been following the strategy you proposed, they wouldn’t have any more information than they did on Day 1. The results of the lynches wouldn’t tell anyone anything, since they’ll all be vanilla Town, just like everyone expected, and the voting wouldn’t tell anyone anything, since they’d all be for the designated vanilla, just like everyone expected.