My reaction was like: Hammer? [SUBMIT]
2 seconds later
WAIT THAT WAS HAMMER
Posted: Sat Jun 26, 2021 2:32 pm
by Toogeloo
Do you have a game where you did townie things and got town read for them, T3?
Posted: Sat Jun 26, 2021 2:33 pm
by Jake The Wolfie
In post 199, Almost50 wrote:@Jake: No, that's totally fine. 6 out of 9, and 8 scums netted. Now do you think this supports the rule or not?
I think we need to investigate why this rule holds
Posted: Sat Jun 26, 2021 2:36 pm
by Almost50
Simply because it hasn't been used that often, so it remains the case. Once it gains momentun Scum are going to try to be "late to the party" I guess
Posted: Sat Jun 26, 2021 2:44 pm
by Ircher
With 13 alive, it takes 7 votes to fade a player.
Posted: Sat Jun 26, 2021 2:51 pm
by T3
lmao
Posted: Sat Jun 26, 2021 2:52 pm
by T3
In post 202, Toogeloo wrote:Do you have a game where you did townie things and got town read for them, T3?
Mini Normal 2205? I think.
Posted: Sat Jun 26, 2021 2:53 pm
by MURDERCAT
I am only doing this for theme reasons, feel free to ignore
Spoiler: statistics
You have a hypothesis that “[The f]irst three posters of any given game contain 1 scum 80% of the time.” We will compare this to the expected amount given 3 random players instead.
Because we are sampling 3 players from the list without replacement, we need to use the hypergeometric distribution. The hypergeometric distribution tells us the number of successes k (here defined as the observed number of nontown players) given two parameters: N (the population size) and K (the total number of nontown players). Using this distribution, we can determine the expected number of nontown players in a random selection of 3 players and compare to what you have observed with the first 3 players to post.
The problem with this approach is that we can only use games with 3 scum and 13 total players. Alternatively, we could use the binomial distribution (which assumes sampling with replacement) but given the small populations we are dealing with I don’t expect that to be particularly accurate. Besides, this situation is the most comparable to our game.
Let’s calculate the expected number of scum in a random group of 3 players. We can do this is R with the following command
to get a value of 0.4720 but there is a simple intuition behind this. There are 3 ways to get a single scum player in a group of 3 players, we can either pick a scum player with only our first draw, second draw, or third draw. Intuitively, you might sense that these are equally likely scenarios, but we will prove that in a second. First, let’s calculate the odds of getting scum on only the first draw. In our setup, that’s 3/13 * 10/12 * 9/11 which is equal to ~0.1573. The odds of getting scum on only the second draw are 10/13 * 3/12 * 9/11 which is also equal to ~0.1573. If you look closely, you’ll see that it’s the same numbers in both expressions. So, our odds of having exactly 1 scum in a random group of 3 is equal to 3 * ~0.1573 which equals 0.4719 (this is equal to the above, there is just a rounding error).
We now have the expected odds of having scum in the first 3 posts
if and only if
the first 3 players are randomly selected. You claim that it’s not random however and the odds of there being a single scum in that group is higher. Let’s look at the data now to see if that holds. The games you provided all have a differing number of players and scum, so I’m pulling 13p games myself.
Here are the games that I found (for convenience I only looked at opens) and the number of scum who were in the first 3 posters:
This leaves us with:
0: 9 (39%)
1: 10 (43%)
2: 4 (17%)
So compare our 43% versus our expected value of 47%. Pretty close! That took a really long time so I’m not going to bother to do the hypothesis test right now, but I might do that later. Also there might be mistakes related to failing to find replacements or things, so feel free to check my work.
Posted: Sat Jun 26, 2021 2:54 pm
by T3
Wait no definitely not that one.
I can give you my first game when I made 5 posts, went AWOL, and was killed night 1.
Posted: Sat Jun 26, 2021 2:55 pm
by Almost50
@MC: Do you want to claim, perhaps?
Posted: Sat Jun 26, 2021 2:56 pm
by T3
UNVOTE: MURDERCAT
Posted: Sat Jun 26, 2021 2:56 pm
by MURDERCAT
Wow I spend 2 hours trying to teach you all stats and you try to kill me I see how it is
Posted: Sat Jun 26, 2021 2:57 pm
by MURDERCAT
There is no way this wagon on me is real, these votes are awful and trying to run me up to a claim is terrible
Posted: Sat Jun 26, 2021 2:57 pm
by T3
I agree with the reads on him but it's way too early. slimer voting twice is uh.....
Posted: Sat Jun 26, 2021 2:58 pm
by T3
Honestly I really don't like MCS rection and I have no idea why.
Posted: Sat Jun 26, 2021 3:21 pm
by Almost50
I think T3 should put his vote back on MC. I don't believe this "rigid face" reaction from him. I think it's more likely to come from Scum who knew they weren't hammered.
Posted: Sat Jun 26, 2021 3:27 pm
by Jake The Wolfie
A50, do you think MC would freak out if they were town who knew they weren't hammered?
Posted: Sat Jun 26, 2021 3:28 pm
by Titus
MC was hammered. I just pmed the mod. They forgot my vote.
Posted: Sat Jun 26, 2021 3:29 pm
by Titus
Nevermind lol.
Posted: Sat Jun 26, 2021 3:29 pm
by T3
In post 218, Titus wrote:MC was hammered. I just pmed the mod. They forgot my vote.
In post 216, Almost50 wrote:I think T3 should put his vote back on MC. I don't believe this "rigid face" reaction from him. I think it's more likely to come from Scum who knew they weren't hammered.
This is what I meant to quote
Posted: Sat Jun 26, 2021 3:38 pm
by Almost50
In post 217, Jake The Wolfie wrote:A50, do you think MC would freak out if they were town who knew they weren't hammered?
NO, but he'd be pissed enough to throw out some fumes, not just a one-line comment and disappear.
Posted: Sat Jun 26, 2021 3:40 pm
by Almost50
You still need to claim, MC. There's no reason for this wagon to dissolve if you're acting like it isn't there to begin with.