Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[liopoil]

I'm looking at the scenarios for lynch vs no-lynch.

mail-mi, if Town refuses to lynch and scum refuses to kill, what happens?

Scum forgets about the keys, Andy forgets about you and throws the chest away. Everyone loses.

but when exactly does this happen? If scum no kill then town no lynches, does everyone lose then or is scum forced to kill? what if the kill is just blocked in some way?

[yuma]

So I think that no-lynch is the best option.... if not my favorite

However! I don't think we should vote to no-lynch. I think we should use the full day to scum hunt and try and figure out who the fire mafia is so ice mafia can better shoot them during the night and let the day timeout into a no-lynch. Quickly no-lynching I think wastes this time that we have been given... especially time with our IC.

So I am going to unvote, but not move to a no-lynch vote...

[sudgy]

So I think that no-lynch is the best option.... if not my favorite

However! I don't think we should vote to no-lynch. I think we should use the full day to scum hunt and try and figure out who the fire mafia is so ice mafia can better shoot them during the night and let the day timeout into a no-lynch. Quickly no-lynching I think wastes this time that we have been given... especially time with our IC.

So I am going to unvote, but not move to a no-lynch vote...

Ooh, I like the idea of timing out to give us the most time.  Unvote.

If everybody says they want to no lynch but aren't voting, you'll still wait for the day to end, right?

[yuma]

Ooh, I like the idea of timing out to give us the most time.  Unvote.

Although if we do stagnate and stall I do think moving to a voting no-lynch is better. But hopefully we won't get to that stage for a few days at least... We have some talking to do still, especially as this topic--a necessary one yes--has completely derailed the scum hunting conversation, something I hope we can get back into tomorrow and through the weekend.

[mail-mi]

I'm looking at the scenarios for lynch vs no-lynch.

mail-mi, if Town refuses to lynch and scum refuses to kill, what happens?

Scum forgets about the keys, Andy forgets about you and throws the chest away. Everyone loses.

but when exactly does this happen? If scum no kill then town no lynches, does everyone lose then or is scum forced to kill? what if the kill is just blocked in some way?

That's what happens over a course of days. If town no-lynch twice and scum no kill twice, everyone loses.

[Gveoniz]

The no lynch option seems interesting.

Ok.  I am repeating what Voltaire said in a different format

I spent some time to make graph showing options and out comes: 
D3: green
N3: red
D4: purple

It is still very messy (I only tried to make that the text don't overlap) and need a lot more improvements, I will also have to do some maths. I will come back soon to fix the graph and post my own analysis.
(note that the relative position of the thing does not matter, only the arrows do.)

[Jimmmmm]

The no lynch option seems interesting.

Ok.  I am repeating what Voltaire said in a different format

I spent some time to make graph showing options and out comes: 
D3: green
N3: red
D4: purple

It is still very messy (I only tried to make that the text don't overlap) and need a lot more improvements, I will also have to do some maths. I will come back soon to fix the graph and post my own analysis.
(note that the relative position of the thing does not matter, only the arrows do.)

I've been working on a less graphical version with percentages. Much messier than I anticipated.

[liopoil]

I've been working on a less graphical version with percentages. Much messier than I anticipated.

As have I. It's incredibly messy, especially because of the nightkills w/two scum alive. It's a mini-game, and it's a fair bit of work to find the optimal strategy for each scum, and I have to do that for each nightstate possible w/two scum alive... I likely won't finish. Just with 2 scum and 1 VT, the analysis is pretty in depth.

[yuma]

I've been working on a less graphical version with percentages. Much messier than I anticipated.

As have I. It's incredibly messy, especially because of the nightkills w/two scum alive. It's a mini-game, and it's a fair bit of work to find the optimal strategy for each scum, and I have to do that for each nightstate possible w/two scum alive... I likely won't finish. Just with 2 scum and 1 VT, the analysis is pretty in depth.

While I appreciate the work and enthusiasm people have here... I question whether or not we really need it. Do we really need three separate people working... and likely spending large amounts of time on this? Obviously we do want to make the right decision here so maybe someone should spend some time on it--and then others can double check the maths?

Because ultimately this game is more about scumhunting... something that hasn't happened in the last ~48 hours--myself included I will add--because we are all talking about theory instead.

Really this game reminds me of MV--I know none of you were in that game, so let me say a little about it. It was a game where ultimately a handful of players decided to take a completely numbers approach to the game. They made charts, spreadsheets and calculated percentages. And the end result? It advantaged me as scum and my partner O. Ultimately we lost that game. Why? Because Frisk decided to go against the charts and the spreadsheets and the calculations and went with his reads and shot me to save the game for town.

So ultimately while I think having this data is great I am not going to go solely off it. I am also going to go off my reads, as everyone else should... hence why I think it is important that people focus just not on the data side.

Here I am mostly talking about Jimmmm and Gveo. Lio has already scumhunted a lot more than the two of you combined. But if you other two have enough time to make charts and figure out percentages, you have enough time to scumhunt I think... and ultimately I think your scumhunting is going to be more worthwhile than your charts and graphs--though I appreciate the effort that went into making them--to town in helping us find scum or helping us find out if you are scum.

[liopoil]

I'm at the point where I'm pretty sure that the only way to feasibly do this is to program some sort of mafia simulator... working out the theoretical probabilities/optimal play is just too complex, it would be better to just see what strategy does best over thousands of trials. I know a bit of java, but this project is probably too hard for me to complete within a reasonable amount of time. So Yuma is likely correct. Unless Jimmmmm has a better idea...

[yuma]

Alright... after getting baby down I've got about 15 minutes before I need to head to bed myself... let's see if I can't get something out of it hunting wise...

So right now the idea is that ice mafia will try and shoot fire mafia... while fire mafia will likely take Voltaire out as fire is currently in the lead. That means, the more important interactions today are going to be with eevee, the dead ice goon.

So my main thought is that scum is going to exist in this area (bolded):

but for now, I'm just mystical and give reads without explanations

town: yuma (!), Voltaire (duh)

scum: robz, e, sudgy, archetype

wouldn't have remembered they are playing had e not just referenced them (nice job with the summary posts btw, your effort is appreciated): liopoil (!), jimm, whoever else isn't mentioned except for gweoniz who I need to reread, jury is still out.

I feel that this is where eevee would put a partner. liopoil is arguing that town is where eevee put a partner (that is me...) but as I know that is incorrect I am looking elsewhere. I suppose we could get into a discussion about the theory behind that, but right now I am not up for that sort of a discussion as I am more interested in finding scum then attempting to defend myself...

I will say the reason I think eevee puts his partner here is because it is the safer place... it doesn't require a lot of accountability... he can just say... "they are in the game, who knew?!? and not have to say anything else. While I think 2.7 and sudgy are both dangerous to put in the scummy slot as scumpartners. Too easy for a lynch to take off and he is caught in a situation where he has to buss or has to do some sort of a reversal... again I think bussing in this situation is very, very unlikely as if a buss was successful, in all likelyhood ice mafia would gone right now with a eevee-partner lynch and an eevee nk... not a good way to go when there is an opposing scum team that can easily take you out during the night...

now he does go into more depth about them later: key point bolded by me

Gweoniz, Jimmmm, liopoil: I'm pretty much totally null for the lack of information (either because they haven't posted enough, because nothing in their posts has stood out to me enough or simply because I can't remember their posts). I think having a null category this size at this point in the game is ok. I could develop meaningful and informed reads on them by putting in more work, and that will become necessary later days, but given I have a healthy pool of scum reads already (so I expect these guys to mostly fill the town read category), I don't feel the effort it would take is necessary (=advances the accuracy of my reads enough compared to the unpleasant effort it would take). Basically, without being on the reread all the time, my brain can hold a meaningul, informed read on only so many players at the time.

again, eevee is putting three players in a nice, comfy position where he doesn't have to back up anything said about a potential partner. If eevee did this to an isolated player I think it would be pretty obvious... but he didn't... he grouped together the lurkers--one of which I think was his partner--and just expressed a null read on them and said that he felt what he had was good enough... compare that to his comments on Arch, who was just as much of a lurker, if not more than the others, but eevee has plenty of thoughts on him... so why not separate out these other three completely null reads? Because I think one is his partner and he is hiding him with the cover of two other lurkers....

I guess the objective from this point out is going to be trying to separate those three out... hopefully one is the fire mafia.... wait.... did i just do this completely backwards!!!

Ugh... yes I did. We don't want to find the ice mafia today, we want to find the fire mafia...? so Robz's partner? so ice mafia can shoot them Well that is deflating... sheesh....

But I can't stay up any longer, so that will be a project for another day, but I am going to post this anyways so we have it later if we need it....

[Jimmmmm]

I thought this would be easier if I assume the Fire Mafia will kill Volt, and it turns out it is... slightly.

Assumptions:
Fire Mafia will Kill Volt and then Ice hunt.
Ice Mafia will Fire hunt.
Volt will not use his Doctor power to protect anyone.

Note: Underlined, italicised text means options that are statistically worse for Town and should therefore theoretically not be taken. I have omitted the details by which I determined which was worse, but it's not too hard to work out if necessary.

1 IC, 4 VT, 2 scum

Lynch VT (2/3): 1 IC, 3 VT, 2 scum

• Ice kills VT (3/4): 2 VT, 2 scum
• Lynch VT (1/2): 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): LOSE (total = 2/3*3/4*1/2*1/2 = 1/8)
• 2 scum die (1/4): WIN (total = 2/3*3/4*1/2*1/4 = 1/16)
• Nobody dies (1/4): LOSE* (total = 2/3*3/4*1/2*1/4 = 1/16)
• Lynch scum (1/2): LOSE (total = 2/3*3/4*1/2 = 1/4)
• No-lynch: 2 VT, 2 scum
• 2 VT die (2/9): LOSE (total = 2/3*3/4*2/9 = 1/9)
• 1 VT dies, 1 scum dies (4/9): LOSE (total = 2/3*3/4*4/9 = 2/9)
• 2 scum die (1/9) WIN (total = 2/3*3/4*1/9 = 1/18)
• Nobody dies (2/9): 1 KVT, 1 VT, 2 scum
• Lynch KVT (1/4): 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): LOSE (total = 2/3*3/4*2/9*1/4*1/2 = 1/72)
• 2 scum die (1/4): WIN (total = 2/3*3/4*2/9*1/4*1/4 = 1/144)
• Nobody dies (1/4): LOSE* (total = 2/3*3/4*2/9*1/4*1/4 = 1/144)
• Lynch VT (1/4): WIN** (total = 2/3*3/4*2/9*1/4 = 1/36)
• Lynch scum (1/2): LOSE (total = 2/3*3/4*2/9*1/2 = 1/18)
• No-lynch: 1 KVT, 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): LOSE (total = 2/3*3/4*2/9*1/2 = 1/18)
• 2 scum die (1/4): WIN (total = 2/3*3/4*2/9*1/4 = 1/36)
• Nobody dies (1/4) 2 KVT, 2 scum => no-lynch: WIN** (total = 2/3*3/4*2/9*1/4 = 1/36)
• Ice kills Fire (1/4): 3 VT, 1 scum => no-lynch: 2 VT, 1 scum
• Lynch VT (2/3): LOSE (total = 2/3*1/4*2/3 = 1/9)
• Lynch scum (1/3): WIN (total = 2/3*1/4*1/3 = 1/18)

Lynch scum (1/3): 1 IC, 3 VT, 1 scum => IC dies: 3 VT, 1 scum => no-lynch: 2 VT, 1 scum

• Lynch VT (2/3): LOSE (total = 1/3*2/3 = 2/9)
• Lynch scum (1/3): WIN (total = 1/3*1/3 = 1/9)

No-lynch: 1 IC, 4 VT, 2 scum

• Ice kills VT (4/5): 3 VT, 2 scum
• Lynch VT (3/5): 2 VT, 2 scum
• 2 VT die (2/9): LOSE (total = 4/5*3/5*2/9 = 8/75)
• 1 VT dies, 1 scum dies (4/9): LOSE (total = 4/5*3/5*4/9 = 16/75)
• 2 scum die (1/9) WIN (total = 4/5*3/5*1/9 = 4/75)
• Nobody dies (2/9): 1 KVT, 1 VT, 2 scum
• Lynch KVT (1/4): 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): LOSE (total = 4/5*3/5*2/9*1/4*1/2 = 1/75)
• 2 scum die (1/4): WIN (total = 4/5*3/5*2/9*1/4*1/4 = 1/150)
• Nobody dies (1/4): LOSE* (total = 4/5*3/5*2/9*1/4*1/4 = 1/150)
• Lynch VT (1/4): WIN** (total = 4/5*3/5*2/9*1/4 = 2/75)
• Lynch scum (1/2): LOSE (total = 4/5*3/5*2/9*1/2 = 4/75)
• No-lynch: 1 KVT, 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): LOSE (total = 4/5*3/5*2/9*1/2 = 4/75)
• 2 scum die (1/4): WIN (total = 4/5*3/5*2/9*1/4 = 2/75)
• Nobody dies (1/4) 2 KVT, 2 scum => no-lynch: WIN** (total = 4/5*3/5*2/9*1/4 = 2/75)
• Lynch scum (2/5): 3 VT, 1 scum => VT dies: 2 VT, 1 scum
• Lynch VT (2/3): LOSE (total = 4/5*2/5*2/3 = 16/75)
• Lynch scum (1/3): WIN (total = 4/5*2/5*1/3 = 8/75)
• No-lynch: 3 VT, 2 scum
• 2 VT die (3/8): LOSE* (total = 4/5*3/8 =3/10)
• 1 VT dies, 1 scum dies (3/8): 2 VT, 1 scum
• Lynch VT (2/3): LOSE (total = 4/5*3/8*2/3 = 1/5)
• Lynch scum (1/3): WIN (total = 4/5*3/8*1/3 = 1/10)
• 2 scum die (1/16): WIN (total = 4/5*1/16 = 1/20)
• Nobody dies (3/16): 1 KVT, 2 VT, 2 scum
• Lynch KVT (1/5): 2 VT, 2 scum
• 2 VT die (2/9): LOSE (total = 4/5*3/16*1/5*2/9 = 1/150)
• 1 VT dies, 1 scum dies (4/9): LOSE (total = 4/5*3/16*1/5*4/9 = 1/75)
• 2 scum die (1/9) WIN (total = 4/5*3/16*1/5*1/9 = 1/300)
• Nobody dies (2/9): 1 KVT, 1 VT, 2 scum
• Lynch KVT (1/4): 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): LOSE (total = 4/5*3/16*1/5*2/9*1/4*1/2 = 1/1200)
• 2 scum die (1/4): WIN (total = 4/5*3/16*1/5*2/9*1/4*1/4 = 1/2400)
• Nobody dies (1/4): LOSE* (total = 4/5*3/16*1/5*2/9*1/4*1/4 = 1/2400)
• Lynch VT (1/4): WIN** (total = 4/5*3/16*1/5*2/9*1/4 = 1/600)
• Lynch scum (1/2): LOSE (total = 4/5*3/16*1/5*2/9*1/2 = 1/300)
• No-lynch: 1 KVT, 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): LOSE (total = 4/5*3/16*1/5*2/9*1/2 = 1/300)
• 2 scum die (1/4): WIN (total = 4/5*3/16*1/5*2/9*1/4 = 1/600)
• Nobody dies (1/4) 2 KVT, 2 scum => no-lynch: WIN** (total = 4/5*3/16*1/5*2/9*1/4 = 1/600)
• Lynch VT (2/5): 1 KVT, 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): LOSE (total = 4/5*3/16*2/5*1/2 = 3/100)
• 2 scum die (1/4): WIN (total = 4/5*3/16*2/5*1/4 = 3/200)
• Nobody dies (1/4): WIN** (total = 4/5*3/16*2/5*1/4 = 3/200)
• Lynch scum (2/5): 2 VT, 1 scum
• Lynch VT (2/3): LOSE (total = 4/5*3/16*2/5*2/3 = 1/25)
• Lynch scum (1/3): WIN (total = 4/5*3/16*2/5*1/3 = 1/50)
• No-lynch: 1 KVT, 2 VT, 2 scum
• 2 VT die (2/9): LOSE* (total = 4/5*3/16*2/9 = 1/30)
• 1 VT die, 1 scum die (4/9): 2 VT, 1 scum
• Lynch VT (2/3): LOSE (total = 4/5*3/16*4/9*2/3 = 2/45)
• Lynch scum (1/3): WIN (total = 4/5*3/16*4/9*2/3 = 1/45)
• 2 scum die (1/9): WIN (total = 4/5*3/16*1/9 = 1/60)
• Nobody dies (2/9): 2 KVT, 1 VT, 2 scum
• Lynch KVT (2/5): 1 KVT, 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): 2 VT, 1 scum
• Lynch VT (2/3): LOSE (total = 4/5*3/16*2/9*2/5*1/2*2/3 = 1/225)
• Lynch scum (1/3): WIN (total = 4/5*3/16*2/9*2/5*1/2*1/3 = 1/450)
• 2 scum die (1/4): WIN (total = 4/5*3/16*2/9*2/5*1/4 = 1/300)
• Nobody dies (1/4): WIN** (total = 4/5*3/16*2/9*1/4 = 1/300)
• Lynch VT (1/5): WIN** (total = 4/5*3/16*2/9*1/5 = 1/150)
• Lynch scum (2/5): 2 VT, 1 scum
• Lynch VT (2/3): LOSE (total = 4/5*3/16*2/9*2/5*2/3 = 2/225)
• Lynch scum (1/3): WIN (total = 4/5*3/16*2/9*2/5*1/3 = 1/225)
• No-lynch: 2 KVT, 1 VT, 2 scum
• 1 VT dies, 1 scum dies (1/2): 3 VT, 1 scum => no-lynch: 2 VT, 1 scum
• Lynch VT (2/3): LOSE (total = 4/5*3/16*2/9*1/2*2/3 = 1/90)
• Lynch scum (1/3): WIN (total = 4/5*3/16*2/9*1/2*1/3 = 1/180)
• 2 scum die (1/4): WIN (total = 4/5*3/16*2/9*1/4 = 1/120)
• Nobody dies (1/4): WIN** (total = 4/5*3/16*2/9*1/4 = 1/120)
• Ice kills Fire (1/5): 4 VT, 1 scum
• Lynch VT (4/5): 3 VT, 1 scum => VT dies: 2 VT, 1 scum
• Lynch VT (2/3): LOSE (total = 1/5*4/5*2/3 = 8/75)
• Lynch scum (1/3): WIN (total = 1/5*4/5*1/3 = 4/75)
• Lynch scum (1/5): WIN (total = 1/5*1/5 = 1/25)
  

So if we lynch:
Pr(WIN) = 1/18+1/36+1/36+1/18+1/9 = (2+1+1+2+4)/36 = 10/36 = 5/18 ~ 27.8%
Pr(LOSE) = 1/9+2/9+1/18+1/9+2/9 = (2+4+1+2+4)/18 = 13/18 ~ 72.2%

If we no-lynch:
Pr(WIN) = 1/10+1/20+1/45+1/60+1/180+1/120+1/120+4/75+1/25 = (180+90+40+30+10+15+15+96+72)/1800 = 137/450 ~ 30.4%
Pr(LOSE) = 3/10+1/5+1/30+2/45+1/90+8/75 = (135+90+15+20+5+48)/450 = 313/450 ~ 69.6%

*I am considering cases with 2 scum and 1 Town a loss for Town. Perhaps the Townie could convince one of the scum to no-lynch, but I wouldn't bet on it.
**If we assume that both scum are scumhunting, then in situations where it is Night and they both know exactly who the other is, then in theory that results in a win for Town. Obviously one may decide to kingmake and give the other the win, but if they play optimally they kill each other and hope the other decided to kingmake.


Now obviously we can't be sure that the Fire Mafia will kill Volt, but if he doesn't that just makes the no-lynch even better.

So statistically no-lynching (and continuing to no-lynch when appropriate) is better than lynching (and then no-lynching when appropriate) by around 2.6%. Of course if we compared no-lynching when appropriate to always lynching, it would be a significantly bigger difference. Of course if we trust our collective reads more than the individual reads and desire to scumhunt of the two remaining scum by at least 2.6%, then maybe we should lynch Today anyway. But I'm not sure that we do.

So as boring as it is, I think our best option is to Vote: no lynch

[2.71828.....]

I am out now.  Will be following the game but won't be able to post long thoughts until Monday, only one liners at best.  I am happy if we decide to no lynch before then, but FOS to anyone in my list here who hasn't completed my "assigned homework" before a lynch/no lynch.

FoS to everyone but Yuma so far. Discussing odds won't help. Reads will help. Lio needs to make his case. Sudgy needs to post his notes. Jimmmmm and gveo need some reads to start with.

[liopoil]

Assumptions:
Fire Mafia will Kill Volt and then Ice hunt.
Ice Mafia will Fire hunt.
Volt will not use his Doctor power to protect anyone.

--snip--

*I am considering cases with 2 scum and 1 Town a loss for Town. Perhaps the Townie could convince one of the scum to no-lynch, but I wouldn't bet on it.
**If we assume that both scum are scumhunting, then in situations where it is Night and they both know exactly who the other is, then in theory that results in a win for Town. Obviously one may decide to kingmake and give the other the win, but if they play optimally they kill each other and hope the other decided to kingmake.

Very nice. I was trying to work out something that finds the best strategies for scum as well, which is much harder. I'm very unsure that it is best for the two scum to act like you think they will... and it may be good for voltaire to save people. Cases with 2 scum and 1 town are not a loss for town at all. The Ice mafia will agree to a no lynch for sure, because the Ice mafia loses if they lynch anyone but the fire mafia, and since neither the fire mafia nor the town is going to be willing to lynch the fire mafia... ice mafia will go for no lynch. Fire mafia has to go for no lynch too because they need to pretend to be town, since no lynch is happening anyway. If we reach such a scenario, I suspect that the day will be over after 2 posts! then, that night, it's really complicated. After quite a  bit of work I am still not sure what happens. Most likely the best option is for scum to choose randomly, a la rock-paper-scissors, but not purely random. But town has a non-zero chance that night for sure. How would the scum ever know who each other are? You use "KVT" a few times, what does that mean? known VT? I don't think we can ever create a Known VT. It seems you think after a night w/out kills one is created? I'm pretty sure this is not the case.

[Jimmmmm]

You use "KVT" a few times, what does that mean? known VT? I don't think we can ever create a Known VT. It seems you think after a night w/out kills one is created? I'm pretty sure this is not the case.

Oh sorry! Forgot to explain that one. It means "Known (to scum) VT". Basically it's someone who has survived a NK because both scum targeted them and therefore know they're Town and therefore will not be targeted again since we're assuming both scum are scumhunting.

[liopoil]

You use "KVT" a few times, what does that mean? known VT? I don't think we can ever create a Known VT. It seems you think after a night w/out kills one is created? I'm pretty sure this is not the case.

Oh sorry! Forgot to explain that one. It means "Known (to scum) VT". Basically it's someone who has survived a NK because both scum targeted them and therefore know they're Town and therefore will not be targeted again since we're assuming both scum are scumhunting.

ahhh, neat. However, this is not quite the case, because there is a possibility of voltaire saving someone/scum killing themselves.
[/quote]
however, your analysis is a pretty good estimation, so we should probably go with it. Later in the game we will be able to make a more exact analysis.

[sudgy]

So, I was going to get around to my reread today, but I forgot that I'm doing something early for Thanksgiving...

[Gveoniz]

I am not able to get anything done today, sorry.

I do agree that we are done with the theory now, I think I would go no lynch unless we suddenly get some very certain scum.

I will try to get reads out instead of the percentages for my graph in later, hopefully tomorrow.

[liopoil]

I just played 100 million games of mafia.

I wrote a simulator. The code is pretty messy though and likely has some logic errors. Does someone here know java? I can post the code for people to look at.

It has town winning 45% of the time, and each scum virtually splitting it right down the middle at about 27.5% each. this is pretty different from Jimmmm's calculations.

It does not account for KVT at the moment...

I programmed a few different strategies. Ice mafia always shoot someone who isn't themselves or voltaire. I programmed two different strategies for fire mafia - one which shoots voltaire, one which doesn't. And get this - the one that doesn't shoot voltaire preformed MUCH better.

For voltaire, there were two strategies - always save, and never save. I couldn't find a difference for which one he did. For town, there were 3 strategies - always no lynch, never no lynch, and no lynch today only. The one that always no lynched did poorly, and I couldn't find a difference between the no lynch today and never no lynch strategies.

I must have done something wrong here...

[yuma]

But if you have enough time to make charts and figure out percentages, you have enough time to scumhunt I think... and ultimately I think your scumhunting is going to be more worthwhile than your charts and graphs--though I appreciate the effort that went into making them--to town in helping us find scum or helping us find out if you are scum.

[liopoil]

But if you have enough time to make charts and figure out percentages, you have enough time to scumhunt I think... and ultimately I think your scumhunting is going to be more worthwhile than your charts and graphs--though I appreciate the effort that went into making them--to town in helping us find scum or helping us find out if you are scum.

Yeah I know... I was just going to start it to see if it was plausible and then kept going... I'm more just enjoying it than actually doing it for whatever negligible benefit it may give us. But yeah, I'll stop now.

[yuma]

But if you have enough time to make charts and figure out percentages, you have enough time to scumhunt I think... and ultimately I think your scumhunting is going to be more worthwhile than your charts and graphs--though I appreciate the effort that went into making them--to town in helping us find scum or helping us find out if you are scum.

Yeah I know... I was just going to start it to see if it was plausible and then kept going... I'm more just enjoying it than actually doing it for whatever negligible benefit it may give us. But yeah, I'll stop now.

Well if you enjoy it, then by all means... don't let me be someone to tell you to stop!

[liopoil]

well, the thing is, I do have things I should be doing...

[yuma]

Ok:

So I am going to retry my attempt to figure this out if I can. So... what we actually want is ice mafia to shoot fire mafia... so we need to find Robz (the fire goon)'s partner.

That means taking a closer look at Robz, not Eevee (although it is worthwhile to take a look at eevee, but not in lieu of analyzing robz for tonight's actions...

So robz:

I don't want to lynch myself, Voltaire, yuma, or 2.7, basically.

Eevee and sudgy are really the only two players I've looked at very in depth. And I don't come away with much more than, well they could be scum.

Scum have to be lurking, right? Maybe a lurker lynch? Liopoil? Archetype? Gveoniz?

I mentioned this post before. And I think it has the same significance that the post from eevee I mentioned does, but that kinda makes me value it less.

Like I said before I think sudgy isn't robz's partner... nor do I think 2.7. But I feel about this less strongly than I do in regard to eevee. Robz I think could very easily put a scum partner as a strong town read. It is very much within the realm of possibility to consider for him (see MXIX where he lied to protect his partner mcmc!!)

I do find it strange that he omitted Jimmmm from this list here. That used to be seen as a towntell (or rather a tell that they aren't partners, but that has kinda morphed over time) and like I mentioned Robz does later end up voting for Jimmmm a few posts later. So I guess I don't think Jimmm is the most likely... because I do think Robz would be very hesitant to buss a teammate in this situation... again see MXIX where Robz had a chance to buss mcmc, but didn't (or did he toward the end?, that is worth checking.... he did. He ended up being the hammer... so Robz was willing to hammer a partner when it was obvious that he was going down... but I don't think he would want to start up a wagon on a partner in a multiball game. So I would theorize that Jimmm isn't his partner.

So that leaves liopoil, gveo, 2.7 to an extent as I have eliminated Jimmmmm and sudgy and myself...

two non-comital reads here about lipoil and gveo:

Of our lurkers... NHS isn't actually a lurker, just someone who probably needs to be replaced.

Archetype is very often mislynched, and pretty lurky is town.

Sudgy is lurking and fits some scum criteria for me, so cool.

As for liopoil, I have no idea what liopoil's meta is, except I think I've seen him lurk as town but try to have good content, sort of like theorel but less extreme. Do the vets agree/disagree with me here?

Gveoniz is just getting into the game I think and I try not to lynch newbies on Day 1 without exceptionally strong reasons.

Others have mentioned this quote as well:

I'm pretty sure we'll find more scum OFF the faust wagon, but I need to know exactly who was on the damn wagon in the first place.

which I think from a point of view points more toward players on the wagon (so that would be liopoil of the people still alive that I haven't already eliminated)

and toward the end of day he says this:

Vote: Archetype

Will also vote for Gveoniz. Liopoil and Jimm have done enough to avoid the lurker lynch, at which point I would need like other reasons to lynch them, and I don't have any.

If Archetype is scum, I think Eevee is suspicious. I know Eevee is voting for Arch, but I feel like he keeps trying to subtly push things in a different direction. Like just now when he said, okay doesn't matter which lurker, let's do the one Voltaire thinks (even though we all mostly prefer Arch), and then tried to find Jimm scummy for something that even he doesn't really think is scummy.

which I think puts gveo more on the less likely to be partners side... Maybe. Maybe robz is setting himself to potentially hammer gveo is the lynch gets to that point... But he does give liopoil a safety net and I already talked about how I dont think robz would have started a wagon on partner jimmm....



I have to end this here... baby needs some help... so I think fire mafia if they want to hit ice mafia should target in this order:

liopoil, gveo, jimmmm, 2.7, sudgy, volt/yuma.

[liopoil]

I think fire mafia if they want to hit ice mafia should target in this order:

you got it backwards again!