Open 680: C9++ (Abandoned due to mod error. See new thread)

[davesaz]

Thanks for that (she says, pretending that she knows anything about coding, or that she remembers anything from logic class).

I have one little question, based on the wiki. If the mod picks seven numbers and repeats are allowed, why do all of the role assignations other than T stop before 7? (E.g. V up to 5, but no further explanation of what would happen if you got 6 or 7 Vs.). Is it just presumed that the pattern keeps repeating, or is there some other rule that prevents all the letters from being non-T?

https://wiki.mafiascum.net/index.php?title=C9%2B%2B

That hopefully concludes the newbie setup questions.

If there were more of the same letter, the number of that PR in the game would max out according to the table, and the mafia/sk roles would still be chosen according to the number of T's rolled. So 7 D's would be 3 docs, goon, roleblocker, godfather.

[davesaz]

And yet another accidental pagetop. Why is it that I can never do that on purpose.

[tojam2]

VOTE: Shannon OMG what is that disgusting thing on your profile picture!

VOTE: Excel Not original by any means but the best I have to go off right now.

I cannot count this vote because I am not sure who you intend to vote for. You are welcome to revote.

[davesaz]

Compensating for something?

[CogMachine]

Gonna be on a plane today so prob slow posting from me.

Noted

[CogMachine]

Hey Umlaut, you mentioned a tool you wrote to work out the setup based on known info. I know we don't really have much info yet, but I'd love to know more about the tool and how it works.

Spoiler: Nerd talk

So, I have

• a bunch of specific predicate functions for taking a C9++ draw and asking things like "Is there any cop?", "Is there a one-shot cop?", "Is there a serial killer?", etc., e.g.

  Code: Select all

  (sk? '(T T T C V D M)) -> #t
  (sk? '(T T C V D D M)) -> #f

  
  
• an initial probability distribution over all possible C9++ draws. I generated this algorithmically but now it's stored as just a big association list.
  
• a function "condition" that accepts a predicate and a probability distribution, and returns the new probability distribution conditional on that predicate's truth.
  
• a function "prob" that takes a predicate and a probability distribution, and returns the probability that that predicate holds.

When I get new information I generate and save a new probability distribution by conditioning on that information. Or I can just condition on something to see what would happen. E.g.

Code: Select all

(prob maf-godfather? (condition (disjoin full-cop? doctor?) *initial-distribution*))

will tell me the probability of a godfather conditional on there being either a cop or a doctor.

This is of course limited to conditioning on information that eliminates possibilities altogether. For example, in principle every lynch of a townie ought to slightly increase our credence that there are two scum as opposed to three (because we're more likely to hit town when there are more town to hit), but there is no way to condition on that with this tool.

Any chance you could throw this up on GitHub? I'd be interested to see how it works exactly. If not no biggie

[Umlaut]

I'm actually cleaning it up now, planning to put it up somewhere when it's clean enough that I'm not embarrassed to show it. (Also probably won't include the precomputed dist but rather compute it at runtime so that it will be verifiably correct.)

Do you speak Scheme?

[TwoFace]

I'm happy with keeping my vote on Excel. He doesn't give reasoning for why Beeboy's posts are scummy, so I guess they just are because he says they are.

does he really need to though? beeboy's posts are sort of scummy, especially that vote on Cog

[shannon]

I took Beeboy's posts as being a bit forced, but not outright scummy. The case against Beeboy seems kinda the same. At this point, it's a wash for me.

[shannon]

VOTE: Shannon OMG what is that disgusting thing on your profile picture!

VOTE: Excel Not original by any means but the best I have to go off right now.

I cannot count this vote because I am not sure who you intend to vote for. You are welcome to revote.

I just had to check to make sure I hadn't uploaded a selfie or something

But not is my adorable Oshawott, who I choose <3

[Alchemist21]

I don't find Beeboy's posts scummy. And the only "forced" post I see is one that should have never been taken seriously by anyone.

[CogMachine]

I'm actually cleaning it up now, planning to put it up somewhere when it's clean enough that I'm not embarrassed to show it. (Also probably won't include the precomputed dist but rather compute it at runtime so that it will be verifiably correct.)

Do you speak Scheme?

All I know about scheme is that it's like Lisp, and that it's a functional language.

How hard could it be?

Also I don't get the Beeboy 'case'. Shannon on the other hand...

UNVOTE:
VOTE: shannon

[defnotmafia]

I don't find it scummy either

[TwoFace]

You don't find it scummy that beeboy voted cog for not voting their top scum read when cog hadn't even given a scumnread yet?

[Alchemist21]

Stop doing lame RVS votes I caught scum :<

Show of hands, how many people still actually thought Beeboy was serious after he made this post?

[TwoFace]

:raises hand:

[Umlaut]

:pointedly does not raise hand:

[FireScreamer]

:Does not raise his pointy hand:

[FireScreamer]

Wow rude you emote jerk

[Ircher]

20 by itself means nothing and can be brushed off easily; but coupled with 24 it suggests a degree of seriousness. That means that 1)beeboy is using himself as a guinea pig for kickstarting the game by doing something scummy, or 2)he failed to see the IC meaning he didn't even pay attention when he put the vote down.
Fake content is usually forced content imo

99% sure it's 1)

[Ircher]

Constantly reminding us how slow a day old rvs game is strikes me as insincere. Trying to get us to talk about ourselves is only something I've ever done if I want to talk about myself in an effort to set something up later.

Ummm..... No......

[Ircher]

Be sure to give lots of strong reads. Even if they're wrong they might end up being indicative.

I can't develop reads until people falsely accuse me of things lol.
I'll try though.

Let me help.
SCUM! SCUM! Already trying to avoid responsibility! Get him guys!

Look at this scum not voting there top scum read.
Case close team, 1 scum down.
VOTE: CogMachine

This looks kinda forced to me

40 pages is a pretty healthy D1 imo. You have to have enough to make decent reads, ya know.

Wouldn't know since most of my games were either newbie or micro bastard.

It was this post Charloux. I didn't find beeboy's response forced, just facetious in a way.

[shannon]

Stop doing lame RVS votes I caught scum :<

Show of hands, how many people still actually thought Beeboy was serious after he made this post?

Not serious, but it was a shit joke.

[Umlaut]

@Mod I'll be V/LA Friday thru Tuesday.

Out of town and probably busy.

Noted. Thank you.

[Umlaut]

Alchemist, I feel like we're on the same page.

How do you feel about Shannon right now?