I agree. Vanilla Town is my favorite. Apparently I’m not as good at it as I need to be though.
Paul I was convinced you were cult until recent events. Welcome to the thread btw.
Prof I thought you were probably town. More or less based on watching you play before, FWIW.
I was expecting to be Vanilla Town myself, but when I wasn’t, I more or less followed the play style mentioned above: I played VT and hunted scum. This was easier in this particular game since we had two scum teams and the cult and the serial killer, so there was legitimate hunting to do. Later in the game, I expected that to get harder.
I’m looking forward to trying other games and seeing how they compare to this one.
I’m pretty sure that Town won all of Pleonast’s Conspiracy games, but if not Town definitely won the first one.
I’m playing around with a system of vote analysis to detect scum;of course, in a real game, I wouldn’t base votes entirely on this, but I am interested to see if it might be helpful.
According to this system, the eight scummiest people, in approximate order of scummitude, currently are:
Lakai, Cygnus, Koldanar, bufftabby, Hoopy Frood, Total Lost, Zeriel, and Hal Briston.
To be clear, this is NOT spoiled information, but the results of an objective analysis of votes.
I’m surprised there isn’t more of a call to lynch cultists. After sach’s death reveal didn’t include the words “non-hostile” as crys’ claimed PM did, I would be treating them as scum.
I’ll be curious to see your system works out. Mine apparently sucks.
Eh… it makes sense for non-hostile to appear in a role PM, not as much in a death reveal. Leaving the info out doesn’t really hurt any faction, including it would be a big benefit for town.
Sent from my SM-G930T using Tapatalk
I haven’t been spoiled yet, but I would be surprised if Hoopy Frood is scum.
What is your system based on?
Sent from my SM-G930T using Tapatalk
Out of curiosity, how does this analysis work?
Lemme fix that for you.
I still haven’t gotten around to getting my new-and-improved vote analysis program working. It’s the same method as the old one, but modified to handle two Scum factions and the larger game, and I’m pretty sure that it’s the theoretically best analysis possible given the data of who voted and unvoted whom on what Days. Of course, there’s still information in when during a Day a vote was placed, and in what reasons were given for a vote, and so on, but I’m not sure how to analyze that.
My analysis is based on the assumption that Scum, as a whole, shows a nonzero tendency to avoid voting for their own. That is to say, Scum will vote for their own to a lesser degree than Town will vote for Scum. This is counting all votes cast by Town, but only votes in place at the end of the Day for Scum, on the theory that even unvoted Town votes were still meant sincerely, but that unvoted Scum votes might well just be a ruse.
Then, it basically goes through every possible arrangement of Town and Scum (or, for a super-large game like this, a representative random sampling of every arrangement), and for each arrangement, checks to see whether that condition is true (as well as other known conditions, like confirmed players). Out of all of the possible arrangements where the condition holds, it then counts up how often each player is Scum.
This method works beautifully in most games, but a sufficiently-skilled Scum team can totally wreck it. I suspect it would have worked well in this game, just because all 14 Scum can’t possibly be skilled enough.
Except that I still haven’t gotten it working, and am about five Days behind on the data, even if it were.
I’m spoiled now… so I won’t comment anymore on the game. If you are not spoiled and want me to check the output of your system I can tell you if it’s working or not. If you are spoiled I’d still like to see what it says.
Sent from my SM-G930T using Tapatalk
I’m spoiled, but my system doesn’t use any spoiled data, so I was planning on posting the results in here (where anyone spoiled can see for themselves how good they are, and anyone not can sit back and wonder).
If I can get it working.
My system is based largely onthis with a few tweaks.
For Days in which Scum were in danger of lynch (only Day 4 in this game), it straightforwardly gives credit for voting for Scum and casts shade on voters for Town who were in danger of lynch (Diggit, in this case). This is probably entirely too unsophisticated, but the only adjustment I have made for the two scum team condition is to reduce the magnitude of the rewards and punishments for these votes. I haven’t made any adjustment at all for the multi-lynch, which is potentially a serious confounder.
For Days in which Scum weren’t in danger, I assume that Scum will want to avoid voting as a bloc and will be dispersed among the wagons, so that any reasonably sized wagon should consist of Town and Scum in approximately the same ratio as the game overall. Then it waits for flips and adjusts the probability of surviving wagoneers accordingly. ISTM that no adjustments are needed here; although there will probably be more random noise in the data, overall we would expect two small Scum teams without knowledge of each other to act the same as one large Scum team.
For example, currently the system thinks very highly of those who voted for ToeJam on Day 2. There were 14 such players, and on Day Two 28.5% of living players were Scum. We therefore assume that 4 (actually 3.99) is the likeliest number of Scum on that wagon. Five of those players have since died, and three flipped Scum. So we assume that there is most likely only one Scum left among the surviving group of nine, giving each of them only an 11.1% chance of being Scum, much less than the 26.7% chance of the average living player. So players get positive or negative points for their votes on each Day, and these are summed to produce an overall scumminess quotient. The whole thing has to be redone every time someone flips, which is a bit cumbersome for this game but wouldn’t be so bad in a normally sized game.
Of course, like any such system, it would be trivially easy for Scum to defeat if people started to take it too seriously.
Currently, it ranks the LEAST scummy players as Biotop, Capt. Klutz, Cookies, Johnny B, MHaye, nesta, Plumpudding and sinjin.
I’m a-gonna guess LightFoot’s RL problems are keeping her from playing. I can’t think of any reason for her just to not post otherwise.
Ignore the above, obviously.
It’s a lot more interesting watching these games while spoiled. I’m learning a lot about how people are playing their roles and considering how I might be playing differently.
I wish there was more posting going on. It seems the conversation is being continued by only 4 or 5 players. That’s bad for everyone. I’m also noticing that time differences are starting to group players in people’s minds with some thinking players are more or less active than they seem to be to me based on how often the interact with each other. Obviously if players are online at the same time conversation speeds up, reinforcing their concept of those players being active while also making slower conversations with other players making them seem less active.
Sent from my SM-G930T using Tapatalk
Sorry. Real life again has taken up (and i know if I was still alive in the game at this point, I’d probably need to be modkilled for inactivity, I’ve just been busy as heck.
Anyways. You asked me a Q.
A: Because I was a doctor role. I did not want to actually fake being a cop. Because that’s not helpful for the Town. However, what I wanted was for the SCUM to think I was a fake Cop. And because coming right out and saying I’m not a cop- it’s completely truthful (thus wouldn’t trigger a real cop to counterclaim), and it will largely be ignored as a throwaway comment.
However, What it will do is actually is create Wifom. I don’t need scum to believe that I’m a cop 100%, I just need to start planting the seed for it, because my goal was for the scum to NK me (knowing that I could block them). The WIFOM only applies to the scum (which I view as a good thing- anything that makes scum have to wonder/doubt themselves on is a good thing, because it’s a distraction from them trying to find the real cops).
This is sorta that reason. As a Doc–>FakeCop, I wanted to draw the attention of the scum only. I don’t really breadcrumb with the assumption that Townies are looking for it (because most townies are inherently lazier than scum), and Scum are more likely to pick up on breadcrumbs. It was something **Storyteller **DID say a while back, and that was my goal this time around. I haven’t read the scumboards yet, so I dont know if they actually did consider it or not. But that’s why my goal was to try to create any type of association with Cop with My name. even if it’s on an unconscious level - without trying to screw with Townies too much/throwing them onto my trail.
-R
The 2nd half of that strategy was (as a lazy townie) to get the attention of the scum, and see if they would go after me during the daytime to try and lynch me. Which is how I enjoy playing Mafia- I tend to be vocal about being lynched and want to see arguements for why players want to lynch me, just to see if the voting records will turn up scum.
I am curious- i should go back and look at how many of those early players I kept calling out from N0 and D1/2 were actually scum.
I see **SunnyDaze ** was Scum. Her, sach, and ThingFish amuses me greatly. I need to see what **Sinjin ** and **guiri **are.
I have pm’ed all quiet players.