Season 2: Werewolf - Shadow Pine - Day 3

Hey Guys... i won't be available until at least Monday as i am out of town.

Happy Thanksgiving

Jad

Warning: this post is long. But I promise it's hot shit, not cold diarrhea. Skip to the bottom for the chart if you don't want to read anything.

Brothers, I apologize for losing my cool. There are only so many times I can explain myself before letting the interrogation get to me. Termy is a villager, and I have to appreciate his attempts to get to the truth.

My accusation on Pascal was based on a hunch, based on his general behavior of playing it safe. Villagers make mistakes, but he does not. What struck me most was his constant posts, but early, safe votes on Manygrams and Nocturne. That way no one could accuse him of questionable voting, while avoiding having to make comments on his partner as long as both of them kept talking. I am not the seer, but I threw his name out honestly believing I was going to die. Termy was attacked instead. Part of me takes that as Pascal being scared off and attacking Termy to frame me. That could be nonsense, I freely admit.

But let's get to the meat of what I have to say. Using the village records, I have analysed all of our conversations over the past 2 days and today. I have charted out who has accused or defended or said nothing of substance to everyone else. The goal was to determine guilt on the basis of association. A wolf could be falsely be cleared due to not being in cahoots with a villager, but the system would determine two wolves being easy on each other. I have determined 6 possible wolf pairings. Everyone is involved in at least 1 pairing, and obviously 5 of them are false positives. It is possible none of these pairings are correct due to the wolves accusing each other, but I consider my threshold of accusation required relatively high, and the wolves did not have much incentive to put on much a show this early while silent villagers were being eliminated.

A mea culpa: eliminating the silent villagers has narrowed down the pairings considerably. So take that as an apology. I still have suspicions of Pascal, given he has only voted for Nocturne, myself, and Romulus, the latter two only after we accused him. 

However, to reiterate, anyone could be a wolf. This chart does not help us much this on this vote, but next round it could work wonders to narrow down the suspects. Giving this information out publicly means wolves can use it as a tool, but I doubt by how. Putting on a show now does not change the fact that they didn't before.

Right of the grey middle line: The chart is organized by pairings. The left box in the pairing is what the person on the veritical left axis said about the person on the horizontal top axis. The right box is vice versa. Together we see the relationship between the two villagers.

Ultimately, a green box is best because it means of that pairing, one person voted to lynch the other OR at least one of the two has since been proven a villager. A red box means this villager never commented on the other villager, while a red box with a black X villager 1 never said anything about villager 2 before villager 2 was confirmed not to be a wolf. A white X means, nothing was said by a confirmed villager in that pairing who is now dead.

Right of the grey middle line: duplicates of the pairings that do not have a a green or blue square and are therefore suspicious. I submit these as your possible wolf pairings.




Full size:
http://fc09.deviantart.net/fs71/f/2014/284/1/4/the_werewolf_accusation_matrix_by_wiseguy_174-d82gq27.jpg

I encourage people to dispute (and correct!) the chart.

Let's kill some wolves.

Well, there was bound to be a typo in that magnum opus:

LEFT of the grey middle line: duplicates of the pairings that do not have a a green or blue square and are therefore suspicious. I submit these as your possible wolf pairings.

I did include a pretty iron clad caveat in the fine print of my list of suspected werewolves. Not my fault you missed it.

glendon wrote:

everything glendon said

I'm only 1/18th of the way through Glendon's mega-post but I just want to say that, as flawed as my logic may be, I believe anyone who doe that much work/writing to figure out who the wolf is, is not the wolf.

And G, I'm pretty sure you were kept alive because the tide of public opinion was already turning against you last night. ie. the wolves didn't need to kill you as you were already half-way through tying your own rope.

PS - If Glendon IS a wolf, I will never be able to trust him again. ever.

Pascal_B wrote:

I did include a pretty iron clad caveat in the fine print of my list of suspected werewolves.  Not my fault you missed it.  

Pfft, Fine print?? Nobody ever reads that... *whistles*

Alright I have to admit that I can't make heads or tails of the chart. Sorry. However, what I do gather of it, I do disagree with a decent amount. Sorry again. Namely, the subjective breakpoints in the spectrum of actions (naturally difficult due to the nuances and undertones of language), but also the hierarchy and order of significance of said actions (ex. 'no comment' is ranked as very suspicious in fire-hydrant red).
To me, what I _do_ get from this though is that I feel the following:
[1] Pascal I still don't like
[2] Mongoose seems to be trying to fly under the radar and only surface to defend Pascal
[3] Glendon I trust (sue me)
[4] TBird and Noc -- garbage data in, garbage data out so not much is really gained from analysis

Glendon, I have a qualm with your methodology and your conclusions.

To start, assuming that your methodology is sound, which it isn't, Tbird is party to the most pairings. Why am I still your main suspect if he is party to the most suspected pairings.

Now to the methodology. Your suspected pairings are all based on a lack of evidence to the contrary. All of your most likely cabdidates are where one villager has not accused another. What if they just ignore one another?

But whatever, I'll play along. I will vote based on your results here. When you are proven wrong we will know how you have been working to deliberately deceive us. I'm on a phone, so apologies for the lack of formatting Slippy.



VOTE: TBIRD

@Termy @Glendon @TBird, read the following statements carefully.

[1] Statistically speaking, if the village doesn't lynch a wolf this round it will be impossible to win. Agree? Alright.
[2] Because of that, at this point the wolves don't actually need to 'silence' anybody and don't care who dies as long as it isn't them. Agree as well? Good, good!
[3] Pascal recently wrote "whatever, I'll play along. I will vote based on your results here" and then redirected his attack, seemingly out of good will. However, his olive branch of "alright, let's agree not to kill each other, and kill TBird instead" is 100% in line with the thinking of a wolf--kill an innocent, it doesn't matter which one.

To me I see that the choice is quite clear: Kill Pascal.

Note that this message is directed at everyone BUT Pascal and I request that the following people do NOT respond:
1. Termy
2. Glendon
3. TBird (there's still a chance you're a wolf and Pascal is so sure that he's going to die that he wants to make sure you aren't a suspect.. but I don't buy it)
The reason I ask you not to respond is that between you you have four votes and you can only successfully lynch if the fifth innocent joins in (either Noc or Mongoose). The fifth vote is likely to be the innocent but if the second wolf gets a sense that there kin will die anyway (eg there are already three votes cast) they will likely throw in their vote as well to avoid suspicion. By remaining silent we will see who is actually willing to vote against Pascal: Mongoose or Nocturne.

PS - Term/Glendon/Tbird, even if you disagree with the following logic, I urge you to be silent at least until one of the two has voted. Srs, please dawgz
PPS - Sorry for the alternating "bold, not bold" text
PPPS - Silencio!

Pascal wrote:

I did include a pretty iron clad caveat in the fine print of my list of suspected werewolves. Not my fault you missed it.

Ok, so an orange square becomes a yellow square. Doesn't cross the pairing off the list.

Romulus wrote:

Namely, the subjective breakpoints in the spectrum of actions (naturally difficult due to the nuances and undertones of language), but also the hierarchy and order of significance of said actions (ex. 'no comment' is ranked as very suspicious in fire-hydrant red).

I agree with the point about the breakpoints being subjective. The difference between an orange and a yellow square is irrelevant as the result is still the same. But between and blue and a yellow it's huge. Look at the glendon/Nocturne pairing: change his comment to yellow and then there's a seventh pairing in contention. As for hierachry, I stand by my chart. The fact that the no comment type is red should stand out because it means these two have never talked about one other significantly the entire time. 

Romulus wrote:

[1] Pascal I still don't like
[2] Mongoose seems to be trying to fly under the radar and only surface to defend Pascal
[3] Glendon I trust (sue me)
[4] TBird and Noc -- garbage data in, garbage data out so not much is really gained from analysis

1. Agree. 2. I can see that. If you change those blues to yellows then there's 2 more pairings in contention. 3. Doesn't mean I trust you! 4. Disagree. The chart is doing its job. Each pairing needs to be judged on an individual basis. Tbird flares up a lot of red because he doesn't say much. Nocturne, less so, but one is on there 3 times and the other twice. 

Pascal wrote:

Tbird is party to the most pairings. Why am I still your main suspect if he is party to the most suspected pairings. 

Because having the most pairings doesn't automatically make you a wolf. The chart is a guide, not a manual.

Pascal wrote:

Your suspected pairings are all based on a lack of evidence to the contrary. All of your most likely cabdidates are where one villager has not accused another. What if they just ignore one another?

That's why there's 6 pairings, not 1. I explained earlier, there are 5 false positives. And even more if a less restrictive standard is used.

EDIT: I see as I was gathering my thoughts Romulus wrote a post. Having read the post, I don't think what I've written here ruins what he's trying to accomplish. I think one thing must be followed though: under no circumstances can there be a lynch during Thanksgiving weekend out of fairness to those who have temporarily left the village.

I also advocated for Glendon to bring himself off from the brink. Does that not count as a data point, or does it get thrown out because it doesn't support your erroneous hypothesis?

Nevermind, I change my vote. Glendon's methodology is stupid. And as Romulus said, if we lynch wrong we lose, and it's not worth throwing away a vote on Tbird to prove a point. Romulus is a wolf. His incessant and at times rabid pursuit of me being lynched is proof enough for me. He is a dog that needs to be put down. Again apologies for the lack of formatting.


UNVOTE: TBIRD

VOTE: ROMULUS

glendon wrote:

So many things

[slightly offtopic] 

I feel like I just got in over my head. 

But really, there's some serious sample size issues here.  And chart design.  No offense, but that thing belongs on http://www.reddit.com/r/dataisugly.

If you can find a better way to track and codify that information, you could write an awesome, if not necessarily useful, paper on predictive analysis of Werewolf/mafia games [/slightly offtopic]

Pascal_B wrote:

Nevermind, I change my vote. Glendon's methodology is stupid. And as Romulus said, if we lynch wrong we lose, and it's not worth throwing away a vote on Tbird to prove a point. Romulus is a wolf. His incessant and at times rabid pursuit of me being lynched is proof enough for me. He is a dog that needs to be put down. Again apologies for the lack of formatting. 


UNVOTE: TBIRD

VOTE: ROMULUS

You were already voting Romulus, weren't you?

Please let me if I'm getting the math wrong but I don't see how if we lynch wrong this round we lose. There's five villagers and two wolves. A lynched villager and a night killed villager leaves us with three villagers, two wolves, and another vote.

Killing the wolf cub however gives the wolves two night kills and then it's over, so be wary.

Killing the cub would make it 3v to 1w, which still seems to be tenable to me.

Wow, I really should get my own math right before correcting others' mistakes. Regardless, that is good news.

For formatting purposes

VOTE: ROMULUS

Wow! My head is spinning with all theses charts, graphs,hypothesis, statistical analysis, and whatever else has happened in the last three days. My opinion is unchanged for all the reasons I've previously stated.

It's amusing to me that the two people who are voting to lynch me are the only two in "fire hydrant red" on my chart. Personally, I'm not sure if tbird is a wolf but that last post is just off to me. 

Tbird wrote:

My opinion is unchanged for all the reasons I've previously stated.

Ha! The only opinions you've stated are that Manygrams is too quiet, Wltgreenwood should be killed because he hampers our ability to get a majority vote (I'll give you that one. An early, astute post that I don't think a wolf would just throw out as a first post), Termy might be tired from night actions, Nocturne has "an interesting choice of words", and I'm suspicious because I didn't want to murder a villager. That's it. You've said nothing of Pascal, Romulus, Witty (before he was confirmed), or Mongoose.

For someone whose lynch vote against me is based on me protecting a quiet villager, you're not much of a chatterbox yourself.

Pascal, Romulus: I believe one of you is a wolf, but I can't decide which is which. My pick for most likely wolf pairings are Pascal & Mongoose or Romulus & Nocturne.

What's almost guaranteed to happen is Termy will be nightkilled tonight. It makes no sense for the wolves to leave some above suspicion in play. The worst thing the village could do is lynch me and then we are left with those four plus Tbird. If that is the case though, my guess is it will come down to the wolf pair trying to turn the village pair against each other, or a wolf pair sacrifices their own so the other can appear innocent and win the game. Each team will jockey for Tbird's support. Or Tbird is a wolf, and then it's even harder to guess who his partner is. Or am wrong, and will feel really, really, bad.

And now to where shit gets real. Killing Tbird. We need to narrow down the suspect list. As Pascal said, he's said the least to anyone. And bonus, he may actually be a wolf this time.

Vote: Tbird.

Tbird, can I count on your support since we can't allow the quiet villagers to live?

Everyone, after reading this post wrote:

But glendon, you didn't want to kill Greenwood because he's silent! You're a hypocrite!

And you lack reading comprehension
Greenwood wasn't around for a week before the night phase so I thought it was pretty obvious that he was a villager. Tbird is clearly participating but not saying much of anything, and therefore suspicious.

Or I'll vote Pascal or Romulus, if that's the way the wind is blowing. Though it sure gives me pause to vote against the only guy who's defending me. It just makes more sense to me to determine which of those two is guilty by picking off their potential quieter partners, starting with Tbird.

ok, i'm home and i see the village was active the whole weekend...

I would read up everyone's theories, diagrams, charts soon and will reply back.

Have a good night

Termy, welcome back. As a reminder and a plea, that chart includes a legend of what the colours mean, as well as a direct quotation explaining every colour chosen (except for no comment at all) . My post also goes into greater explanations about the uses and caveats of the chart. The full size is about 9 megapixels when fully zoomed, so it's clearly readable. I beg of you not to discount it as cavalierly as the others have.

To the brothers who assume my guilt: if I'm lynched you will realize immediately after I am no wolf. At that point, please review what I've said as you continue your search for the guilty ones. Start with the quiet ones and work your way from there. First lynch Tbird. Then Mongoose or Nocturne. If one of them the latter is guilty, your first suspicions should be on Pascal or Romulus, respectively. Seer, check out Pascal or Romulus. If you find a wolf in either, announce it immediately!