Mafia: NPF edition - Page 69

Sithdarth · 07-17-2006, 01:39 AM

Quote:

Stating facts about what probablitity is does not lessen the probability that you are not mafia.
Just serves to confuse.
Smoke and mirrors=Mafia in my eyes.

1) I never claimed it reduced the probability was mafia. People wanted explinations I gave them explinations.

2) I mean it is simple multiplication of fractions resulting in numbers that are either larger or smaller than each other. I fail to see what is confussing you about that. If you truely are confused explain to me what confused you and I'll try and explain it better.

3) Going with the flow is bad even if you are a known townie.

h4x.m4g3 · 07-17-2006, 01:44 AM

Could someone link to the post in which Sith started mentioning the probabillity of mafia members next to eachother, I do not recall under what circumstances he mentioned it.

Sithdarth · 07-17-2006, 01:47 AM

Here it is. Its post 610 if you want what is around it. I'd link the page but I have 30 post per page and that might not match with yours.

Mondt · 07-17-2006, 01:53 AM

Quote:

Originally Posted by Sithdarth

confusing filler post with numbers and whatnot

One: Why are we trying to base stuff off of if there is a evil role next to another on the list? That makes no sense at all!!!

Two: Probability doesn't help us at all. I could say that there is a 1/5* chance of us hitting a mafiate, but does it matter? Not really. Add a bunch of fractions and decimals and it serves to confuse people.

*Totally random number. It's not a miscount. It's random. I tohught of it on the spot.

Sithdarth · 07-17-2006, 02:02 AM

Fine I give up. If you aren't going to read what I post nothing I could say is going to convince you. But if you do take a look at say the "confusing posts with numbers" you can see that having more than one mafiate in a given sample space, say 2 people, is less likely than having only one mafiate there. Its simple probability, in that probability is a number between 0 and 1; 0 being impossible and 1 being certain. So the lower the number the less likely the outcome. Has no one ever once learned anything about probability?

Mesden · 07-17-2006, 02:05 AM

Quote:

Originally Posted by Sithdarth

Fine I give up. If you aren't going to read what I post nothing I could say is going to convince you. But if you do take a look at say the "confusing posts with numbers" you can see that having more than one mafiate in a given sample space, say 2 people, is less likely than having only one mafiate there. Its simple probability, in that probability is a number between 0 and 1; 0 being impossible and 1 being certain. So the lower the number the less likely the outcome. Has no one ever once learned anything about probability?

I have and this is all giving me a good giggle as I understand everything you've been saying.

It's not confusion, he's just stating facts. It's TECHNICALLY more reliable than our suspicion, especially at this point.

Truce · 07-17-2006, 02:12 AM

Eh, what Sithdarth says makes sense, but at this point (based off what we know), the best way to win right now is to eliminate the unknowns from the system, and you don't need a bunch of calculations to do that. Sithdarth just happens to be one of those unknowns.

1 Lynch, 1 PO, 1 Vig with quite a few known townie numbers>3 Mafiates with 1 kill.

Bailey · 07-17-2006, 06:13 AM

I'm not saying that his number are bad, I'm saying that he's trying to get them to prove something they don't.

If he's saying that it's more likely for mafiates to not be next to each other, then why does that exclude all the people who have a townie on either side of them?

Why does that make the people in unknown patches more suspicious?

By my way of thinking, with it being random, we shouldn't be able to guess anything based on positioning within the list, even if some numbers do prove that it's more likely for mafiates to be seperated.

Guess what? In my game, in which I randomized the list, the mafia was clumped 2, 2, and 1.

You can say all you like that a coin will land on heads exactly 50% of the time, but that doesn't change the fact that when you go and actually flip the coin, you may well find that you get 7 heads and 3 tails.

See, that's the thing about random. It's random.

Sithdarth · 07-17-2006, 08:41 AM

Quote:

Originally Posted by Newb

I'm not saying that his number are bad, I'm saying that he's trying to get them to prove something they don't.

If he's saying that it's more likely for mafiates to not be next to each other, then why does that exclude all the people who have a townie on either side of them?

Why does that make the people in unknown patches more suspicious?

By my way of thinking, with it being random, we shouldn't be able to guess anything based on positioning within the list, even if some numbers do prove that it's more likely for mafiates to be seperated.

Guess what? In my game, in which I randomized the list, the mafia was clumped 2, 2, and 1.

You can say all you like that a coin will land on heads exactly 50% of the time, but that doesn't change the fact that when you go and actually flip the coin, you may well find that you get 7 heads and 3 tails.

See, that's the thing about random. It's random.

Maybe just maybe, and I know this might be difficult, that's why I've said something this like everytime I've drawn up some stats:

Quote:

Originally Posted by Me

Its not definite but its better than just going with my gut, or at the very least as good, and I'd never vote only on the probabilities.

Because you know I'm not actually aiming to prove anything and I've never directly pointed to anyone on stats alone and said lynch 'em. It's not like its anyworse then going on feelings. Also, feel perfectly justified in not using this its more to explain my reasoning then anything. This is my "gut feeling" in a way.

Now if you take a look there are only 3 possible combination of townie and mafiate for two people. (Both mafiate, both townie, or a mix, the mix being twice as likely as it can happen in two ways.) The probability of them both being townies is (4/7) * (3/6) = .286 or 28.6%. The probability of one being and the other not (4/7) * (3/6) + (3/7) * (4/6) = .571 or 57.1%, exactly double the other since there are two distinct ways it can happen. Finally the probability of both being mafiates is (3/7) * (2/6) = .143 or 14.3%. Add them all up and you get 100%. However, the total probability of finding at least one mafiate in a group of two unkown is .143 + .571 = .714 or 71.4%.

The chances of just a single unknown person in the list being mafiate is (3/7) or .429 or 42.9%. Also, the chances of a single unknown person being townie is (4/7) or .571 or 57.1%. So if you really have nothing to go on but randomly picking an unknown person pick one from the groups of two unknown people. You have a higher chance of hitting a mafaite.

Mesden · 07-17-2006, 08:56 AM

I kind of checked over the past bit of the thread, and out of all the people, I get the most off feeling about MoM.

He just seems to agree with people at every bend and just adds little posts here and there, never really voting when it was at any progressive point to vote.

Just...everything he's done doesnt' sit well with me...

So, Vote: MoM. I don't normally use gut feelings, but this one hits hard.

07-17-2006, 08:56 AM	#690
Mesden There is no Toph, only Melon Lord! Join Date: Jan 2006 Location: Inside of a box inside of a smaller box Posts: 4,310	I kind of checked over the past bit of the thread, and out of all the people, I get the most off feeling about MoM. He just seems to agree with people at every bend and just adds little posts here and there, never really voting when it was at any progressive point to vote. Just...everything he's done doesnt' sit well with me... So, Vote: MoM. I don't normally use gut feelings, but this one hits hard. __________________ I can tell you're lying.

07-17-2006, 01:44 AM	#682
h4x.m4g3 Lakitu Join Date: Nov 2003 Posts: 2,152	Could someone link to the post in which Sith started mentioning the probabillity of mafia members next to eachother, I do not recall under what circumstances he mentioned it.

07-17-2006, 01:47 AM	#683
Sithdarth Friendly Neighborhood Quantum Hobo Join Date: Mar 2004 Location: Outside the M-brane look'n in Posts: 5,403	Here it is. Its post 610 if you want what is around it. I'd link the page but I have 30 post per page and that might not match with yours.

07-17-2006, 02:02 AM	#685
Sithdarth Friendly Neighborhood Quantum Hobo Join Date: Mar 2004 Location: Outside the M-brane look'n in Posts: 5,403	Fine I give up. If you aren't going to read what I post nothing I could say is going to convince you. But if you do take a look at say the "confusing posts with numbers" you can see that having more than one mafiate in a given sample space, say 2 people, is less likely than having only one mafiate there. Its simple probability, in that probability is a number between 0 and 1; 0 being impossible and 1 being certain. So the lower the number the less likely the outcome. Has no one ever once learned anything about probability?

07-17-2006, 02:12 AM	#687
Truce Lakitu Join Date: Nov 2003 Posts: 3,856	Eh, what Sithdarth says makes sense, but at this point (based off what we know), the best way to win right now is to eliminate the unknowns from the system, and you don't need a bunch of calculations to do that. Sithdarth just happens to be one of those unknowns. 1 Lynch, 1 PO, 1 Vig with quite a few known townie numbers>3 Mafiates with 1 kill.