C9++ - Game End - Mafia Win

Unusual_Dood

Member
Joined
Sep 13, 2015
Messages
132
Reaction score
913
We only have to determine what the chances are of there being at least 1 B generated in the rolls that we don't already know. And the probability of a roleblocker doesn't seem lower because of the roles we already know it is lower.
Your statement would be true if people claimed/died in the same order as the letters that got rolled, however they're not.
What do you mean the probably presented isn't accurate? It absolutely is. (...) You're trying to defend weak's claim with faulty statistics.
You are claiming I'm using faulty statistics to show how the estimated probability isn't accurate, yet fail to even challenge my two other points.
Just quickly am going to point out that the x * probability to find to probability of getting something in x rolls, isn't exactly correct. For instance that would mean the chance for getting a role blocker in 20 rolls would be 100% (0.05 * 20), and we all know that there for instance is possible to get 20T's, which means it isn't 100%.
The probability is low, yes, but the probability estimated right now isn't exactly accurate, since the roles in the game is chosen in a more complex way. For instance we haven't in Choco's calculation considered the cannot be only one more V or M (since we don't have an innocent child and can't have two 1-shot vigs).
 

catgirl12466

Member
Joined
Aug 6, 2011
Messages
66
Reaction score
270
Oh my god, ok, wow. What the heck happened. I was waiting for Molten to respond and then this got going and life hit me and oh jeez.

Ok, first off, full disclaimer: I suck at statistics and I've just spent the better part of an hour trying to refresh myself on this stuff because something is irking me about how we're going about this.

Each roll of the 7 is independent, correct? The probability of one roll does not affect the others. One result does not mean the rest of the rolls are suddenly weighed.
For instance that would mean the chance for getting a role blocker in 20 rolls would be 100% (0.05 * 20)
Incorrect. 5% chance over 20 tries would be calculated using the formula 1 - (1 - 0.05) ^ 20. In which case, the chance of one single B in all those rolls becomes 64%.

So the thing that bugs me is I'm not sure we can calculate the probability of a B given everything we know, because the events are independent. What we CAN say is that yes, given what we know, the probability that there is a mafia roleblocker is intuitively 100%. Given the 7 rolls, the probability of a single B is 1 - (1 - 0.05) ^ 7 = 30%. But that goes for literally any of the PRs letters.

Of course, I might still be incredibly unbelievably wrong. I don't trust myself with statistics, and I just took an hour trying to work out why my numbers weren't lining up and trying to figure out if there was a reason everyone's calculations were done using conditional probability.

In this case, there are no conditions. None of the rolls inherently affect the others.

I suppose, what we can say is, knowing that 3 of the rolls were not Bs, for the remaining 4 rolls... the chance of rolling one B out of 4 is 1 - (1 - 0.05) ^ 4 = 18.5%.

I don't really know what else to say. At the least, I have reasonable suspicion not based on statistics on TWG this time, just because this all seems strange. If nobody else has anything to add or wants to dispute my math, I'll hammer it.
 

Timdood3

Member
Mafia Host
Joined
Apr 1, 2014
Messages
2,339
Reaction score
2,824
Your statement would be true if people claimed/died in the same order as the letters that got rolled, however they're not.
The order roles are rolled in is completely and utterly irrelevant.
You are claiming I'm using faulty statistics to show how the estimated probability isn't accurate
And you're not disputing that claim.
yet fail to even challenge my two other points.
I've addressed both of those "points."
The first here: (quote broke)
It's an average. Meaning that on 20 rolls, you'll usually have approximately 1 town roleblocker.
And the second in the same post you just quoted yourself....
We only have to determine what the chances are of there being at least 1 B generated in the rolls that we don't already know.
But that goes for literally any of the PRs letters.
Incorrect, blockers have a 5% chance to get rolled, every other PR is 15%.
 

Alisha

Member
Mafia Host
Mafia Mod
Joined
Aug 6, 2011
Messages
805
Reaction score
2,829
I've got nothing to dispute, math reasonably checks out at a glance (Is off with roleblocker but seems accurate-ish) and isn't misleading to manipulate anyone like what UNU is doing with his multiplication nonsense

Catgirl I wish you posted more because you're a ray of sanity
 

Unusual_Dood

Member
Joined
Sep 13, 2015
Messages
132
Reaction score
913
The order roles are rolled in is completely and utterly irrelevant.
They are, but the order people claim in and die isn't.

I've addressed both of those "points."
The first here: (quote broke)
And your "point" is that I'm using faulty statistics since the estimated probability is only "approximately" when I point at scenarios showing that these calculations don't work?
 

Timdood3

Member
Mafia Host
Joined
Apr 1, 2014
Messages
2,339
Reaction score
2,824
They are, but the order people claim in and die isn't.
So it's relevant how exactly?
And your "point" is that I'm using faulty statistics since the estimated probability is only "approximately" when I point at scenarios showing that these calculations don't work?
And how exactly don't they work?
 

HKCaper

Almighty Goat
Joined
Oct 21, 2014
Messages
110
Reaction score
1,197
The group decides TheWeakGuy48_ is next to go. In no time the goat dashes into the room and rams into TheWeakGuy48_, which sends him flying out of sight. The goat calmy leaves the room. The goat appears to be laughing softly.
Possibly crying...
or both.


TheWeakGuy48_ was: Town Roleblocker
Night 2 start:
- Night ends on Saturday Jan. 26, at 3 PM GMT. (I wanted to start the night before I go to sleep, but I don't want to end it at this time in 24 hours from now).
- Send in your night action (if you have any) via your forum pm.
- If you haven't send in an action before the deadline, that's too bad. (of course if you know in advance you arent going to make a night deadline msg me)​
 

HKCaper

Almighty Goat
Joined
Oct 21, 2014
Messages
110
Reaction score
1,197
The light that I never turned of turns on once again. One player appear to be missing. With MORE blood ALL OVER,
a few HUNDRED hoof prints leading to and from the room, and an AWFUL stench of goat, a new day begins.


Catgirl12466 was: Vaniller Townie
Day 2 start:
- Deadline is hereby set as Wednesday Jan. 30, at 10 PM GMT. (early night end wtf that's kinda neat i guess)
- With 8 players alive it takes 5 votes to lynch and 4 votes to no lynch.
- If a majority hasn't been reached by the time the deadline hits it will default to a no lynch.​
 
Top