Newbie 783 - Mongol Mafia - Game Over

[Pablo Molinero]

Yeah, I need to see the current count to decide what to do with my vote.

[Debonair Danny DiPietro]

First a quick look at his predecessor, Shadeux. The biggest thing that strikes me right now is the inconsistency. I still don’t think his vote was that bad when viewed on its own, but in retrospect it was exceedingly aggressive and comes right on the heels of him saying things like…

But honestly though? I think he's just mistaken (this *is* a newbie game after all), but better not to take the chance.

I would wait to hear his reply though, before jumping into voting.

Which seem very conservative in nature promoting a patient outlook to the game. Thus his actions and words don’t appear to mesh together.

On to tracker himself, he comes in and immediately goes on a lurker hunt and I’ve made my opinion clear that it’s not a productive behavior and quite possibly scummy. He then goes on to post an obviously ordered list, but claims it’s unordered which if true would defeat the purpose of posting a list in the first place. This seems like a convenient way to not be held accountable for his opinion because if someone objected to someone’s placement then he could always claim they weren’t really meant to be as low or as high as they were.

His last two posts are the proverbial nails in my opinion. In ISO 18 he reneges completely on the lurker hunt and admits to it being not a solid scum-tell, so his biggest contribution to the game was a pointless endeavor. Then in his last post he votes for SOG putting him at L-1, despite having expressed basically no interest in him previously and with no reasons in his post that I can see. He’s basically pushing the only other serious wagon on then his own for what appears to be that reason, it’s the only serious wagon other than his own.

[Pablo Molinero]

He’s basically pushing the only other serious wagon on then his own for what appears to be that reason, it’s the only serious wagon other than his own.

To be fair, with only two days left before deadline, it is typical for the people on the chopping block to jump on the other for survival reasons only. After all, if you know yourself as town, the other guy is always a better option. However, with a game as active as this, I think we could still swing any lynch of whoever we wanted even in "just" two days, so he loses points for taking the easy way out, but it's not as

completely

damning as you make it out to be.

[Pablo Molinero]

ivanavich, do you realize that your vote is still on SOG? Back it up or give some analysis of his play, plz. You've said only one good statement about him so far. Are PP's transgressions enough to get him lynched?

You’ve said:

I played with SOG last game, and this game I have come to realize I still disagree with most of what he says. On the other hand I have no problem with his assesment of players and encourage it. I do want to know what he doesn't like and shame on those of you saying discouraging it. If he has something not to like about every player here why would that be a problem? Take what he gives and use it if you can. As long as he is not being manipulative and giving bad info it should be usefull info.

I hardly think what SOG has had to say has been blanket statements. A couple points he made had escaped my notice until he mentioned them.

Which is hardly damning the guy.

[Debonair Danny DiPietro]

He’s basically pushing the only other serious wagon on then his own for what appears to be that reason, it’s the only serious wagon other than his own.

To be fair, with only two days left before deadline, it is typical for the people on the chopping block to jump on the other for survival reasons only. After all, if you know yourself as town, the other guy is always a better option. However, with a game as active as this, I think we could still swing any lynch of whoever we wanted even in "just" two days, so he loses points for taking the easy way out, but it's not as

completely

damning as you make it out to be.

Maybe not when just looking at that one aspect of it, but when coupled with the fact that there's absolutely no indication that he thinks it's a good let alone the best lynch or even that he has any specific opinions on SOG suggests to me that it is that damning.

[semioldguy]

Regarding my comments on Pablo Molinero:
I don't and didn't find Pablo Molinero to be scummy.

I didn't like that you asked that question and will stand by my reasoning of why I think it was a bad and anti-town question to be asked.

Everyone who I thought of as scummy at the time of my post I specifically wrote that I felt them to be scummy, Pablo Molinero was not mentioned as such.

And which question would that be?

This one:

What is everyone's favorite role (or what you most want to play as if you've never played before)?

As I mention in Post 200

[semioldguy]

I had a vote on me until just a moment ago (from tracker)

Tracker was not voting for you.

Also, why are you only suspicious of the players who are suspicious of you?

[Iecerint]

@ SOG

I had a vote on me until just a moment ago (from tracker)

Tracker was not voting for you.

Also, why are you only suspicious of the players who are suspicious of you?

You're right about the tracker vote; it was you rather than him. I had a lone vote from Shadeaux most of the game that I got used to seeing there. My mistake.

For the second -- first, the main reason for my suspicion of you and tracker is the business on page 3 from before I was even in the game. That part has less to do with your suspicion of me and more to do with poor/scummy play. So that, at least, is separate from any OMGUSing on my part. Apart from that, though -- I guess I'm reasoning that in general players that seem more predisposed to lynching me are slightly more likely to be mafia. This is because there are exactly two players -- the scum players -- who know for certain that they would be OK to see me go. This isn't the only information I've been using to make judgments, but it is the surest piece of information I have.

That said, there are only 2 scum, so I'm trying to take that into account, too. eKiM was the first player to question my posting style, but I think of him as leaning-town by (not-quite-error-proof) process of elimination at this point. This could change if tracker turns up town, etc. Actually, the only players who haven't posted suspicion of me in some form are (I think; forgive me if I missed something) Katy, ivan, and Pablo. So I'm also just more likely to hit suspecters than anyone else.

@ DDD

I think we've had some miscommunication, I thought you were still referencing the SOG analysis and his comments on the wagons and that's what my comments were based around. If you're going to change tracks you need to make that clearer and then provide quotes or post numbers for me/other people to reference.

I'm sorry to've been vague. I tried to be clear by referring specifically to PP and Shadeaux by name rather than their modern incarnations or a hybrid name; I'll try to be more specific in the future.

That said, now that things have been cleared up, I would appreciate your thoughts on the question -- that is, the remote possibility you appeared to have prior ascribed to (that Shadeaux was not scummy, but PP was) in light of PP's vote for Pablo following his vote for Shadeaux (page 3). It's apparent from your discussion of tracker that you've rethought your position somewhat -- Shadeaux = tracker, etc. -- but I'd still like to hear what had prior motivated it. That eventuality (PP-onry) is so remote to me that I don't think your having ascribed to it is scummy, really; rather, your having ascribed to it leads me to believe that I have missed something.

If this is where you were coming from, I think you were being a little misleading. After the first night we'll know the town/scum identities of 0-2 players (allowing the remote possibilities of doctorwin and nolynchvote), which will presumably alter and ease scumhunting (i.e. reducing the number of possibilities down from 5, assuming your somewhat suspect a priori "one from each group" assumption). I think it would be a stretch to suggest that tracker was really advocating ignoring all non-lurking/otherwise information as we vote for the remainder of the game. (Not that I don't think tracker's the better lynch.)

Way to post some generalized statements and present it as opposition. Yes, scumhunting should be easier tomorrow with either a successful lynch to draw connections from or at very least a narrower pool of targets, but that doesn't mean pushing bad logic on day one is acceptable. The strongest statement that's been made is that as a scumtell there's one scum in a group of four, statistically insignificant from a random lynch of the entire vote pool and that's ignoring the person presenting the logic. So far, that's the only logic that I've seen tracker present and it's not making the grade.

Way to be defensive and snide? I still believe you were using math based on the false assumption that a lurker-lynch policy requires a random lynch from the remaining 5 non-lurker players on Day 2. I don't believe that tracker was implying this at all, but you've not really corrected yourself. (For that matter, tracker didn't really question it, either, so it may largely be a moot point.) You're also not taking into account the possibility that two lurker players (or none) are scum, which would change average probability of a successful lynch, especially if the prior probability of 1 or 2 players is relatively high. (This may just be because tracker listed he suspected that only 1 lurker was scum, but since you were responding to my own math, which included the 50% both-are-scum possibility, it seems to be an inappropriate omission.) Were you trying to see whether tracker would take issue with your straw-manning?

I'm not arguing that tracker had some kind of brilliant strategy that we should all ascribe to. I've already posted why I think he's the best lynch, even though I kinda do enjoy reading his posts. Rather, I'm pointing out that you seem to be (unnecessarily) using some pretty biased, straw-man math to attack him. I hoped that in pointing out why I thought the math was biased you would do one of a) explain why the math was not biased or based on false assumptions b) clarify the rationale for the built-in assumptions (e.g. "it was a trap!") or c) apologize and agree that you were oversimplifying tracker's position. The fact that you've instead accused me of "generalizing" for pointing out that your math was itself generalizing tracker's perspective is more suspect than anything else. (May just be that you're sometimes not very friendly, which I can't say is a total impossibility.)

Also: DDD, why haven't you voted for me through all this (if you never suspected Shadetrack and feel that SOG has exonerated PP)? I had a vote on me until just a moment ago (from tracker), but all you've done is unvote SOG. Are you trying to decide between me and FT (the other player you've questioned a bit)? If so, have you ignored him because of his somewhat-troubling lurking?

Just because I disagree with you doesn't mean I think you're scum. Conversely, trying to bait me into voting for you doesn't inspire me to think you're pro-town either.

I'm just trying to understand your actions. You explained, up to a point, why you didn't vote for Shadetrack or SOGPP upon unvoting the latter (i.e. you never liked the Shadeaux wagon and you liked SOG's analysis), but didn't really act until others already had. Your behavior could be construed as waiting to see how things would pan out before weighing in on it too whole-heartedly. Since the voting abstention seemed to be coupled with slightly ratcheting-up the rhetoric against me (or just voicing stronger-than-prior agreement with others'), I don't think it's unusual for me to question it.

[Iecerint]

@ SOG (correction)

Can you explain the part I bolded then? If your suspicions changed, then what made them change between posts 205 and 213?

Prior to your replacement, I was only really attacked by eKIM. DDD sort of weakly agreed with him, but said he had problems more with my style than anything I said; he added that it was "nothing lynchworthy." When you both replaced in, tracker posted first. I inferred from his poor spelling and punctuation that he, if scum, would probably focus on his scumpartner's posts. The scumpartner that would best lead to his focus on me was eKIM, so he was my best guess.

Since the replacements, on the other hand, DDD has come out more strongly against me, but eKIM has mostly ignored me. This could be eKIM trying to avoid being vilified in the event of your mislynching me, but DDD's relative grouchiness and opportunistic attack-time (waiting until he had some wind at his back before pretty much just referencing others' arguments) led me to suspect him over eKIM. Since there's at most 1 non-tracker/SOG scum (if neither of those are scum, we're in bad shape), DDD's guilt means eKIM's probable innocence. (And if both tracker and SOG are scum, eKIM is similarly innocent.) So eKIM's moved to leaning-town. DDD is only neutral because it would take pretty damning evidence for anyone to move to tracker/SOG level.

I've just noticed that I put eKIM into the leaning-town tier PRIOR to DDD starting things up, so I thought I'd point out that I messed up. Looking back with this in mind, I'd guess that the real reason eKIM and DDD were leaning-town and neutral, respectively, in the second post was that I thought it was a cop-out to list 4 players as neutral, so I had to promote one of them to leaning-town. Knowing what I did at the time, I probably just chose eKIM because he's a little friendlier (yeah, not a good basis; I know). Sorry about that; I was fixated on how I came to suspect DDD a bit more (which made eKIM look even better) a few posts later when he joined in on the back of attacks from tracker and SOG and got my chronology a little mixed-up.

The post still accurately reflects my current perspective, though.

[Kublai Khan]

Early-Day Wednesday May 27th 2009 Votecount

tracker - 3 - Katy, ekiM, Iecerint

semioldguy - 3 - ivanavich, Pablo Molinero, tracker

Iecerint - 1 - semioldguy

Debonair Danny DiPietro
Katy
Pablo Molinero
ekiM
Furpants_Tom
ivanavich

Not Voting: Furpants_Tom, Debonair Danny DiPietro

With 9 players, it takes 5 to Lynch.

Note: Sorry about the double vote confusion on Iecerint.

Italisized

player will be lynched at deadline. Which is in approx. 16 hours.

[ivanavich]

ivanavich wrote:
A couple points he made had escaped my notice until he mentioned them.

Care to share them?

sure.

tracker – How do you not have a scum feel after reading eight pages.

and

Furpants_Tom – I didn’t like his entrance into the game. He put a vote on penguana and his reason for abandoning that vote was bad.

ivanavich, do you realize that your vote is still on SOG? Back it up or give some analysis of his play, plz. You've said only one good statement about him so far. Are PP's transgressions enough to get him lynched?

Yes and Yes. I will also wait for the day to be over before making any changes if any. I may keep it though since I do believe his transgressions are a good reason to lynch. Lets just say my vote is no longer by far the scummiest. I've already given analysis of his play. If he is scum SOG is doing a good job as the replacement.

Either way I want to read the last few pages over again before changing my vote. I'm a lil busy today so home to log on later tonight.

[Debonair Danny DiPietro]

That said, now that things have been cleared up, I would appreciate your thoughts on the question -- that is, the remote possibility you appeared to have prior ascribed to (that Shadeaux was not scummy, but PP was) in light of PP's vote for Pablo following his vote for Shadeaux (page 3). It's apparent from your discussion of tracker that you've rethought your position somewhat -- Shadeaux = tracker, etc. -- but I'd still like to hear what had prior motivated it. That eventuality (PP-onry) is so remote to me that I don't think your having ascribed to it is scummy, really; rather, your having ascribed to it leads me to believe that I have missed something.

I viewed them as unrelated events at the time and still don't see this hard and fast connection you're trying to draw. PP's vote there seems to be completely devoid of recent context (eg the Jarmo wagon and Shadeux's L-1 vote) which suggests to me that she didn't take into account any of the recent activity.

Way to be defensive and snide? I still believe you were using math based on the false assumption that a lurker-lynch policy requires a random lynch from the remaining 5 non-lurker players on Day 2. I don't believe that tracker was implying this at all, but you've not really corrected yourself. (For that matter, tracker didn't really question it, either, so it may largely be a moot point.) You're also not taking into account the possibility that two lurker players (or none) are scum, which would change average probability of a successful lynch, especially if the prior probability of 1 or 2 players is relatively high. (This may just be because tracker listed he suspected that only 1 lurker was scum, but since you were responding to my own math, which included the 50% both-are-scum possibility, it seems to be an inappropriate omission.) Were you trying to see whether tracker would take issue with your straw-manning?

It's not a false assumption, the guy promoting the theory admitted that he only expected one scum to be found in that group. Meaning I was working off of his

best

case scenario. And if you're including the 50% both are scum possibility why are you not including the statistically more likely (because of a larger pool size) 0% none are scum scenario. And I've presented no strawmen, I've simply used the information tracker has given us as his motivations and carried it to the logical end.

My point was and still is, any "scum-tell" which merely splits the town in half and each half being equally likely to have scum isn't a scum-tell at all, it's basically no better than a random lynch. But it looks busy, can easily be directed against a townie, and has plausible deniability: all advantages to scum.

I'm just trying to understand your actions. You explained, up to a point, why you didn't vote for Shadetrack or SOGPP upon unvoting the latter (i.e. you never liked the Shadeaux wagon and you liked SOG's analysis), but didn't really act until others already had. Your behavior could be construed as waiting to see how things would pan out before weighing in on it too whole-heartedly. Since the voting abstention seemed to be coupled with slightly ratcheting-up the rhetoric against me (or just voicing stronger-than-prior agreement with others'), I don't think it's unusual for me to question it.

I have no idea why you continue to twist my actions. I've certainly argued with you more since my unvote because it almost seems like you've adopted tracker's lynch lurkers logic, but I haven't spoken towards your alignment at all except for post 249 after you call me out for not voting for you.

~

Vote: tracker

(L-1)

Case is in post 251.

[Pablo Molinero]

unvote

I'm rereading Shadeaux and liking him/tracker less and less. More later.

[ekiM]

ok

i should clarify the list,

i didn't arrange them in groups for nothing

the top 2 groups were the scummier

middle were the nuetral reads

bottom were the townish reads,

if the village wants to post lists in order, i'll post lists in order, but i think that it's a mistake, and unless someone can convince me otherwise, i don't think i'll change it.

You already posted lists in order but said that you didn't.

lurking isn't a solid scum tell, that's why i'm hesitant to vote on it, or anything like it. the only reason to vote on it is to encourage lurkers to stop,

the game is picking up now so it's not as much of a problem

Backtracking.

ekim you said that he hated my defense of my shadow's action, can you be more specific?

preferabably quotes, but you can paraphrase if you like.

You basically repeated what Shadow had said, which had already been found wanting.

Deadline is in the middle of the night for me. I'll be able to check the thread again before I go to bed in ~5 hours time. After that though I won't be around to move my vote.

[semioldguy]

Italisized

player will be lynched at deadline. Which is in approx. 16 hours.

The first post in the thread lists the deadline as being May 28th at 11:59pm, which is also said to be the deadline in your post on Friday, which would have been about

one day and 16 hours

from your post. Today is only May 27th. I have been under the impression that we still had another day.

Also can Furpants_Tom be prodded?

[Iecerint]

That said, now that things have been cleared up, I would appreciate your thoughts on the question -- that is, the remote possibility you appeared to have prior ascribed to (that Shadeaux was not scummy, but PP was) in light of PP's vote for Pablo following his vote for Shadeaux (page 3). It's apparent from your discussion of tracker that you've rethought your position somewhat -- Shadeaux = tracker, etc. -- but I'd still like to hear what had prior motivated it. That eventuality (PP-onry) is so remote to me that I don't think your having ascribed to it is scummy, really; rather, your having ascribed to it leads me to believe that I have missed something.

I viewed them as unrelated events at the time and still don't see this hard and fast connection you're trying to draw. PP's vote there seems to be completely devoid of recent context (eg the Jarmo wagon and Shadeux's L-1 vote) which suggests to me that she didn't take into account any of the recent activity.

Thanks for this! Looking back, I think this may be correct, after all. For whatever reason, I'd thought that PP's Pablo quote was from the post where he'd just voted for Shadeaux. That's why I thought there was a connection, and that's why I thought PP was so scummy. (Did someone post something to this effect somewhere from page 3 to page 6? Or did I totally make it up?) Were we just wailing on PP because she freaked out when we questioned her?

I wish we hadn't had that miscommunication earlier. I'd found it suspicious that your tracker decision came on the heels of my (now perhaps flawed) can't-be-only-PP analysis without your alluding to my post. (Actually, no one really responds to my posts anymore except DDD and SOG. Are they incomprehensible, or just too long?)

Way to be defensive and snide? I still believe you were using math based on the false assumption that a lurker-lynch policy requires a random lynch from the remaining 5 non-lurker players on Day 2. I don't believe that tracker was implying this at all, but you've not really corrected yourself. (For that matter, tracker didn't really question it, either, so it may largely be a moot point.) You're also not taking into account the possibility that two lurker players (or none) are scum, which would change average probability of a successful lynch, especially if the prior probability of 1 or 2 players is relatively high. (This may just be because tracker listed he suspected that only 1 lurker was scum, but since you were responding to my own math, which included the 50% both-are-scum possibility, it seems to be an inappropriate omission.) Were you trying to see whether tracker would take issue with your straw-manning?

It's not a false assumption, the guy promoting the theory admitted that he only expected one scum to be found in that group. Meaning I was working off of his

best

case scenario. And if you're including the 50% both are scum possibility why are you not including the statistically more likely (because of a larger pool size) 0% none are scum scenario. And I've presented no strawmen, I've simply used the information tracker has given us as his motivations and carried it to the logical end.

The reason I was put off by this was that the only player who used numbers with respect to the lurker business was me (in an "in fairness to tracker" apologist post), so it seemed that your numbers were intended to be a response to my numbers ("nope, that point is wrong in the general case") rather than a specific application to tracker's point ("well, in the context of what tracker said and in the absence of other data/forms of evidence he might present us with, his method statistically performs at chance"). I think you already mentioned that you intended it to reflect only tracker's point, but I was so convinced by my interpretation of your original intent that I had to ask one more time. My interpretation of your "my numbers were about tracker's comments onry" post was that you were trying to discredit me (with math that wouldn't apply to my own point) in preparation for Day 2 stuff. (Yeah, I know, I'm kinda a narcissist.)

With regard to using 50% but not 0%: it's true that the greater number of townies implies that 0% scum probability is greater than that of both scum with respect to a Maximum Likelihood sample, but the use of a high "lurker" prior probability may compensate for that (e.g. via Bayesian inference). Given tracker's model (prior probability is somewhat greater for lurkers, but not much more), though, I agree with you that you'd need to use 0% if you used 50%, and the result would be an even-worse scumfinding rate. (I mean, I haven't done the math by hand, but I strongly suspect the prior probabilities aren't different enough.)

My point was and still is, any "scum-tell" which merely splits the town in half and each half being equally likely to have scum isn't a scum-tell at all, it's basically no better than a random lynch. But it looks busy, can easily be directed against a townie, and has plausible deniability: all advantages to scum.

I agree with this in the general case, but I have some problems with it in this case. I still think it's hard to believe that tracker really meant to imply that the Day 2 lynch would be a random vote among active players (as your "1 out of 5" 20% vote in his "best case" seems to imply). Sure, tracker never really corrected you on this (and, as I now understand it, you consequently never worried about accounting for it in your math), but I'm as prone to interpret tracker's failure to rebuff you as meekness rather than guilt. After all, I posted a response to your math; it's not as if he was irrevocably trapped. Based on this interpretation of his behavior, I think you're giving tracker too much credit.

(I guess you could argue that he didn't really need to respond because I was sort of doing it for him. Sorry if I messed up something you were trying to set-up. It looks like he's being lynched, anyway.)

This is all very beside-the-point, though, because we seem to agree that tracker is the best lynch.

I've certainly argued with you more since my unvote because it almost seems like you've adopted tracker's lynch lurkers logic[...]

My intent was to see where your 20% number was coming from, and to determine what you were trying to do with it (force tracker to explicitly come up with auxiliary scumhunting techniques to boost the probability of a successful Day 2 lynch? maliciously misrepresent tracker's position? make a simple error?). I agree with your discussion of "lurker lynch as fake scumhunting" above, but some details of your argument seemed questionable to me. Any sketchiness perceived there will be nullified for me when/if tracker comes up scum.

[Iecerint]

Also can Furpants_Tom be prodded?

Yes, please.

[Katy]

Today is only May 27th. I have been under the impression that we still had another day.

I was under this impression too.

(Actually, no one really responds to my posts anymore except DDD and SOG. Are they incomprehensible, or just too long?)

On this point, it seems the more recent conversation has been regarding this statistical business and I really have nothing to contribute there, although I am actually reading every post and trying to follow along.

I understand the value of that kind of analysis, but my mind doesn't easily work that way, so for me it's easier to focus on a behavior/psychology look at the subject as far as my contributions go and just read any statistical/numbers analyses that people do and figure that into my thinking in the background.

Also, there hasn't been much for me to add today. I am fine with my vote, as I have been all along, and I haven't seen anything yet I wildly disagree with or have a reaction to.

[Furpants_Tom]

Oh man, I sure have sucked at life for the past couple of weeks.

I still have no pbpa. And I can't claim to have read the thread closely enough to have gotten a better feel for either semioldguy or tracker. I do know, however, that I thought Purple Princess was much dodgier than Shadeaux, and that carries over to SOG.

Vote: Semioldguy

I realise that my vote at this point has no effect barring a switch from one of the tracker voters; but it's important to flag my suspicions in a verifiable way, so that the rest of you can properly analyse me later. I also promise that if I haven't managed to put in some proper analysis by about a week into the next day, I'll request a replacement. Sorry about this.

[semioldguy]

Oh man, I sure have sucked at life for the past couple of weeks.

I still have no pbpa. And I can't claim to have read the thread closely enough to have gotten a better feel for either semioldguy or tracker. I do know, however, that I thought Purple Princess was much dodgier than Shadeaux, and that carries over to SOG.

Vote: Semioldguy

I realise that my vote at this point has no effect barring a switch from one of the tracker voters; but it's important to flag my suspicions in a verifiable way, so that the rest of you can properly analyse me later. I also promise that if I haven't managed to put in some proper analysis by about a week into the next day, I'll request a replacement. Sorry about this.

How do you expect us to hold you accountable and analyze you later for your suspicions when you admit to not being up to date enough to get more recent reads on any of the players?

[Kublai Khan]

End of Day Wednesday May 27th 2009 Votecount

tracker - 4 - Katy, ekiM, Iecerint, Debonair Danny DiPietro

semioldguy - 3 - ivanavich, tracker, Furpants_Tom

Iecerint - 1 - semioldguy

Debonair Danny DiPietro
Katy
Pablo Molinero
ekiM
Furpants_Tom
ivanavich

Not Voting: Pablo Molinero

With 9 players, it takes 5 to Lynch.

Note:

Italisized

player will be lynched at deadline. Which is in approx. 16 hours.

The first post in the thread lists the deadline as being May 28th at 11:59pm, which is also said to be the deadline in your post on Friday, which would have been about

one day and 16 hours

from your post. Today is only May 27th. I have been under the impression that we still had another day.

My bad. My calender was off by a day. Deadline is 24 hours from now.

[Iecerint]

Oh man, I sure have sucked at life for the past couple of weeks.

I still have no pbpa. And I can't claim to have read the thread closely enough to have gotten a better feel for either semioldguy or tracker. I do know, however, that I thought Purple Princess was much dodgier than Shadeaux, and that carries over to SOG.

Vote: Semioldguy

I realise that my vote at this point has no effect barring a switch from one of the tracker voters; but it's important to flag my suspicions in a verifiable way, so that the rest of you can properly analyse me later. I also promise that if I haven't managed to put in some proper analysis by about a week into the next day, I'll request a replacement. Sorry about this.

How do you expect us to hold you accountable and analyze you later for your suspicions when you admit to not being up to date enough to get more recent reads on any of the players?

Indeed.

[Debonair Danny DiPietro]

Bla bla bla, I'm mostly wrong because I misinterpret everything Danny says so I can disagree with him.

Or something like that. At least that's what I got out of the first two overly long paragraphs you wrote.

With regard to using 50% but not 0%: it's true that the greater number of townies implies that 0% scum probability is greater than that of both scum with respect to a Maximum Likelihood sample, but the use of a high "lurker" prior probability may compensate for that (e.g. via Bayesian inference). Given tracker's model (prior probability is somewhat greater for lurkers, but not much more), though, I agree with you that you'd need to use 0% if you used 50%, and the result would be an even-worse scumfinding rate. (I mean, I haven't done the math by hand, but I strongly suspect the prior probabilities aren't different enough.)

Correct, and part of my contention was that my experience shows that tracker's prior distribution is unsound which would turn the statistics even further against him. As it stands a reasonable assumption of randomness still doesn't back his theory.

But the important thing to be noted about this paragraph is that mafia isn't just a game of being right (though that helps alot), it's also about persuasion. And while I as a trained statistician understand what you're talking about with maximum likelihood functions and Bayesian assumptions, someone without specific training like Katy appears to be lost. If you've lost your audience you surely aren't persuading them and thus failing at one of your primary duties. /Fake IC Mode

I agree with this in the general case, but I have some problems with it in this case. I still think it's hard to believe that tracker really meant to imply that the Day 2 lynch would be a random vote among active players (as your "1 out of 5" 20% vote in his "best case" seems to imply).

No, he never said that, but if we make the assumption of one scum in each side of the equation then it doesn't make any sense to ignore the other half of the equation because doing so loses the town the game. This wasn't a point tracker made at all, but something I raised to address the flaw in your analysis.

My intent was to see where your 20% number was coming from, and to determine what you were trying to do with it (force tracker to explicitly come up with auxiliary scumhunting techniques to boost the probability of a successful Day 2 lynch? maliciously misrepresent tracker's position? make a simple error?). I agree with your discussion of "lurker lynch as fake scumhunting" above, but some details of your argument seemed questionable to me. Any sketchiness perceived there will be nullified for me when/if tracker comes up scum.

And again the 20% argument came in rebuttal to your points, stop trying to shift all of this off onto tracker. He's the one who pushed the general idea, but you're the one who has been arguing the soundness of it, except you're not even pushing it as a legitimate course of action. Hence I think you're arguing with me just to argue with me and not out of any real conviction for the idea and while I appreciate the mental workout it distracts us from our purpose of finding scum.

[Furpants_Tom]

unvote
vote: Semioldguy

putting SOG at L-1, nobody shove him into the noose w/o more info however, it's not that time yet. we need more analysis before lynching,

Wow, just read up to this. If you don't want him lynched, don't put him at L-1. I get that you're not keen on a lynching, but this is just raw scumtell. How is tracker not dead already?

His last two posts are the proverbial nails in my opinion. In ISO 18 he reneges completely on the lurker hunt and admits to it being not a solid scum-tell, so his biggest contribution to the game was a pointless endeavor. Then in his last post he votes for SOG putting him at L-1, despite having expressed basically no interest in him previously and with no reasons in his post that I can see.

QFT

It's hammer-time.

Unvote
Vote: tracker

[Iecerint]

Bla bla bla, I'm mostly wrong because I misinterpret everything Danny says so I can disagree with him.

Or something like that. At least that's what I got out of the first two overly long paragraphs you wrote.

Nah. I misinterpreted other things you said, but not here. But it's true that it's me admitting I'd been wrong.

With regard to using 50% but not 0%: it's true that the greater number of townies implies that 0% scum probability is greater than that of both scum with respect to a Maximum Likelihood sample, but the use of a high "lurker" prior probability may compensate for that (e.g. via Bayesian inference). Given tracker's model (prior probability is somewhat greater for lurkers, but not much more), though, I agree with you that you'd need to use 0% if you used 50%, and the result would be an even-worse scumfinding rate. (I mean, I haven't done the math by hand, but I strongly suspect the prior probabilities aren't different enough.)

Correct, and part of my contention was that my experience shows that tracker's prior distribution is unsound which would turn the statistics even further against him. As it stands a reasonable assumption of randomness still doesn't back his theory.

But the important thing to be noted about this paragraph is that mafia isn't just a game of being right (though that helps alot), it's also about persuasion. And while I as a trained statistician understand what you're talking about with maximum likelihood functions and Bayesian assumptions, someone without specific training like Katy appears to be lost. If you've lost your audience you surely aren't persuading them and thus failing at one of your primary duties. /Fake IC Mode

Good point.

I agree with this in the general case, but I have some problems with it in this case. I still think it's hard to believe that tracker really meant to imply that the Day 2 lynch would be a random vote among active players (as your "1 out of 5" 20% vote in his "best case" seems to imply).

No, he never said that, but if we make the assumption of one scum in each side of the equation then it doesn't make any sense to ignore the other half of the equation because doing so loses the town the game. This wasn't a point tracker made at all, but something I raised to address the flaw in your analysis.

What if I argued that Night actions/miscellaneous data might exonerate/kill enough players that the Day 2 lynch was reliably better than 20%? This is my main gripe with that number, but you seem pretty convinced it's a silly thing to argue. Why?

My intent was to see where your 20% number was coming from, and to determine what you were trying to do with it (force tracker to explicitly come up with auxiliary scumhunting techniques to boost the probability of a successful Day 2 lynch? maliciously misrepresent tracker's position? make a simple error?). I agree with your discussion of "lurker lynch as fake scumhunting" above, but some details of your argument seemed questionable to me. Any sketchiness perceived there will be nullified for me when/if tracker comes up scum.

And again the 20% argument came in rebuttal to your points, stop trying to shift all of this off onto tracker. He's the one who pushed the general idea, but you're the one who has been arguing the soundness of it, except you're not even pushing it as a legitimate course of action.

It's true that the argument responded to my request that you help me understand what I had missed, but you've argued that you based it on tracker's one-scum-in-each-group assumption; ergo, I imagined that it was (intended) as much targeted at tracker as at me. Also, it's a stretch that I'm arguing the "soundness" of it; I'm just arguing that it might be at-least-slightly-better-than-random, whereas you have argued that it is no better than random.

Hence I think you're arguing with me just to argue with me and not out of any real conviction for the idea and while I appreciate the mental workout it distracts us from our purpose of finding scum.

This is mostly accurate, but it ignores the possibility that you are among the scum. In that case, the discussion might hypothetically have helped to expose scumminess.

It's also worth pointing out that my two DDD conversation pieces are related to, respectively, the Shadeaux/PP business on page 3 and to tracker's lurker lynch business. I thought I'd caught you defending an untenable position in the first case, but it turned out I'd misremembered the relevant parts of the thread. Similarly, I thought I'd caught you exaggerating the silliness (in this game) of a Day 1 lurker lynch. So I don't think that it was unwise or diversionary for me to try to clarify your positions on those issues.