TWG XL: Mall Madness

[iggymatrixcounter]

Re: TWG XL: Mall Madness

In my example, I assume the SK is public knowledge too since he cannot die from the wolves and the humans would be throwing away the game by lynching him. So the lynch pool is 11 (13 -SK -kagome) for today. 10 for the SK (the same people minus the lynch).

To find the overall percent I did: 1 - (1 - [3/11]) X (1 - [3/10]) = 1 - (.72 X .7) = .496 which I'm not exactly sure is correct, but I'm not sure on how to find the probability of an event occurring when given 2 chances.

[DiscoBobbyPARANOiA]

Re: TWG XL: Mall Madness

That sounds right to me, but I'm not the best in math.

Also, I'm pretty sure the sk will come out tomorrow after he kills who we choose tonight and it's agreed he's not going to get lynched until endgame.

[DiscoBobbyPARANOiA]

Re: TWG XL: Mall Madness

Also, iggy, you seem A LOT more optimistic about this than you did last night, ha.

[iggymatrixcounter]

Re: TWG XL: Mall Madness

Also, I'm pretty sure the sk will come out tomorrow after he kills who we choose tonight and it's agreed he's not going to get lynched until endgame.

He should come out NOW so we don't lynch him and drop our pool of who to kill. The SK can prove himself it's not like a fake wolf could do it. All he has to do is say "I'm the SK and I'm killing X tonight" and if it happens then that's good enough for me.

[iggymatrixcounter]

Re: TWG XL: Mall Madness

Also, iggy, you seem A LOT more optimistic about this than you did last night, ha.

It's all pure speculation. I figure that IF the SK came to you wanting to give up his chance at his own destiny then these numbers would be the case.

But if you're lying then I did a whole lot of work for nothing XD.

[Afrobean]

Re: TWG XL: Mall Madness

Essentially, if you don't understand, is that by killing the SK, we gain 1 day because there aren't 2 kills a night. But we use that extra day to kill the SK so our net gain is nothing. If we keep the SK alive (and he's really going to work with us) then we still get our 3 lynches, PLUS 3 kills from the SK. So 6 chances at getting a wolf is way better than 3 right? [/COLOR]

[COLOR=black]Now depending on when we get a wolf, the math becomes trickier because we DO have to kill the SK eventually. So in practical purposes, we may lose one of the SK's kills. But we are garanteed 4 kills.

Ok, so if we leave the SK alive and he continues making kills, whether they be voted upon (which pretty much can't be done today unless they simply off the person with the second most votes, which, in my opinion would be really stupid) or not, we get around 1 extra kill (saying 4 guaranteed compared to 3 guaranteed)?

If we only get one kill extra as you say, couldn't we allow the SK to live for tonight, let him kill anyone he thinks is a wolf, then lynch him tomorrow?

Even if the math was wrong and we'd actually get a few extra kills, would you seriously want to use them? Seriously, we have trouble killing 1 wolf per day. Do you really think we can win this game when we're killing 2 humans per day? Personally, I'd rather make this game last as long as possible, so as to give ourselves the best chances of deducing things. When we have two lynches per day, we make the game end two times as fast.

[nforcer06164]

Re: TWG XL: Mall Madness

This doesn't work for me; that is, DBP knowing who the SK is. Why? Simply this: the SK, if anybody knows who he is, has given up his chance of winning the game in full. Once we decide we don't need him (and I think we don't), he's dead. He's already lost. Now, what person in their right mind would just decide to lose the game for no good reason? Either DBP is a lying wolf, or we have some players making really dumb moves this game. Don't write that off; looking at the history, the latter is actually very probable.

[rioting]

Re: TWG XL: Mall Madness

There seems to be a lot of numbers flying around again and I want in. So here's my second shot at trying to provide some useful data:

Being the CS major that I am, I wrote a simple recursive program to find every possible game and its outcome. Here are the results:

H = Humans win
W = Wolves win
SK = Serial Killer wins
J = Joint Wolves/Serial Killer win

Keep the SK alive
Wins: H:2005481106 W:4770057600 SK:3490179840 J:1991485440
Total Games:12257203986 H:16.36% W:38.92% SK:28.47% J:16.25%

Kill the SK now
Wins: H:6807024 W:26490240 SK:0 J:0
Total Games:33297264 H:20.44% W:79.56% SK:0.0% J:0.0%

Killing a wolf now
Wins: H:214970274 W:426384000 SK:342558720 J:162570240
Total Games: 1146483234 H:18.75% W:37.19% SK:29.88% J:14.18%

Accidentally killing a human right now:
Wins: H:150418140 W:384935040 SK:273611520 J:167086080
Total Games: 976050780 H:15.41% W:39.44% SK:28.03% J:17.12%


Finding and killing the SK is, mathematically speaking, the best out course of action. So I agree with nforcer that it is/was a very dumb idea for the SK to reveal themselves to DBP or anyone else because effectively they eliminate their chances for winning. Also note from looking at the results that killing a human now is, as you might expect, the worst possible action so avoiding further reducing the human population in a lynch is of the utmost importance. And, even though they get yelled at for saying it, in looking at the numbers, I have to agree with everyone who has said our chances for pulling off a human win are grim.

[nijatwo]

Re: TWG XL: Mall Madness

Kill the SK now
Wins: H:6807024 W:26490240 SK:0 J:0
Total Games:33297264 H:20.44% W:79.56% SK:0.0% J:0.0%

Notice that though this is the highest outcome for the humans, it's also the highest probability for the wolves. Nice job ignoring that. Remember, when p(sk)=0, p(w) and p(h) both increase.

Another issue is calculating probability across 4 options (what you did) and giving your results as a binary problem (human win vs !human win). It's not wrong, there's just more to it than looking at human win percentages. Data is sound (thanks for writing the program), but I think the conclusion needs to be rethought.

Once I get off work I'll look at the P(w)=1-P(l). Again, thanks for the numbers.

[iggymatrixcounter]

Re: TWG XL: Mall Madness

If we only get one kill extra as you say, couldn't we allow the SK to live for tonight, let him kill anyone he thinks is a wolf, then lynch him tomorrow?

Even if the math was wrong and we'd actually get a few extra kills, would you seriously want to use them? Seriously, we have trouble killing 1 wolf per day. Do you really think we can win this game when we're killing 2 humans per day? Personally, I'd rather make this game last as long as possible, so as to give ourselves the best chances of deducing things. When we have two lynches per day, we make the game end two times as fast.

1) Lynching him tomorrow would still waste the extra day we earned.

2) I look at twg as concrete and objective as I can make it. If given the chance to go with a strategy that yields 4+ kills in 3 days as opposed to one that yields 3 kills over 4 days, I'm picking the one that gets us the most chances to make a kill.

[iggymatrixcounter]

Re: TWG XL: Mall Madness

@rioting: I have no clue how to come up with those numbers. But there is no way that the wolves and SK can win together. There are conditions that make sure one or the other will win, so that fudges your numbers a bit already. Plus realize that I assume random lynchings and things but those assumptions only raise when adding logic and reasoning to the kills. Your model shows things that we wouldn't consider (I.E. lynching kagome, the wolves wolfing themselves... if it maps out EVERY possible outcome for everything.) Those things just aren't practical and is a bad representation of what the current situation is.

I still say that if this whole situation with the SK is real then he should come out (because if the SK really wanted to help out, he would make sure we don't lynch him and not consider him in the voting pool)

I really think that disco (or the SK) should give us this name.

[Kilgamayan]

Re: TWG XL: Mall Madness

Notice that though this is the highest outcome for the humans, it's also the highest probability for the wolves. Nice job ignoring that.

Why does it matter? A wolf win is the same as a serial killer win in the sense that they're both a human loss. What do we care who wins if it's not us?

EDIT: However, I'd like to see a model, if possible, that completely eliminates the possibilities of Kagome getting lynched or shot by the SK. Like iggy said, the numbers are a bit skewed because those events were included, and the percentages are close enough together that their order might change.

[nforcer06164]

Re: TWG XL: Mall Madness

The SK is dead if he revealed himself, as I said. We have to kill him eventually. Therefore, him coming out changes nothing. Unless, of course, DBP is lying. Which I unfortunately consider more and more with each post that goes without a name.

[iggymatrixcounter]

Re: TWG XL: Mall Madness

The SK is dead if he revealed himself

Disco claims he already has. The SK is dead anyways, why not be a public figurehead and actually have fun for losing the game. And at least that way we know another person to NOT lynch.

Lynching the SK in the next two days will yield no extra kills NONE!!!!! That huge post I made confirmed it. It had numbers, logic, facts, everything. I find that anyone that says the SK shouldn't come out (Given the current situation) to be trying to cast doubt so the SK doesn't do it. I hope the SK rubs you out for it.

Kagome also can lower the pool by telling his result to someone or to the thread. Get that info out there. (Normally I'd say keep it all to yourself, but it's a little different circumstances)

[iggymatrixcounter]

Re: TWG XL: Mall Madness

BTW I'm adding more evidence to my vote on nforcer: an inability to understand what possible outcomes there are for the humans. Sure the SK being public knowledge or not changes nothing..... for the SK. But the humans can further benefit from the SK's sacriface than you are looking at.