Print

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

[Witherweaver]

Just the moment I hoped my math masters and participation in school/university olympiads meant something

okay, this is only partially directed at you, but it is a very valuable life lesson, and I feel the need to say it.

experience in a certain area is a really shitty metric for who's likely right in an argument that takes place in this area. this is especially true if both persons discussing are reasonably intelligent and both know a little bit about the subject. In fact it's not only not a strong indication, it's almost entirely worthless.

the probability discussion is a great example. I am snapping here because this comment you made bothers me

Silver, if you is not persuaded by arguments provided by Haddock (for some reason, they are quite clear and convincing), can you just take it from someone who was candidate for Russian IMO team and won both Google CodeJam and TopCoder Open (and probability is quite popular theme in this competitions) that he is correct.

People who have spent a lifetime studying something reach vastly different conclusions all the time, every day. Just take a look at politics.

That's why I studied math and not politics. And I stopped to read that topic when it had gone from 2 pages to 5 in like an hour. I still firmly believe that given the assumptions Haddock made his answer is right. Assumptions may be wrong though

Anyway, what I said was a bit arrogant, sorry. I just wanted people to note that the fact that scum NK and yuma kill can target same person makes yuma kill a bit better. I maybe got too excited about that

Again, if we're operating under the assumption that the setup is balanced, this would have been taken into account.

I'm not sure if Ash would actually run a simulation, though.

[Teproc]

So from a game-theory POV, I probably agree with what Egor said - shooting will probably be beneficial. I mean, I don't like the role, but it also has the plus side of semi-confirming yuma, yes? Because scum can't have two night kills. He could be third-party I guess, but he's be confirmed non-mafia.

Alright, unvote I think... this is not the worst plan ever.

There might be two scum factions (either mafia+SK or mafia+werewolves)

[Witherweaver]

I'm going to

Vote: that Yuma uses his power if we lynch Town today

[Teproc]

Even from a game theory PoV I disagree. I don't believe games are actually balanced for full randomness : doesn't the normal setup favor scum normally ? Because in reality town is supposed to do better than random.

Are we really doing this ? This seriously makes me angry. Why are we even playing this game if you all would rather just press a button on random.org ?

[EgorK]

Even from a game theory PoV I disagree. I don't believe games are actually balanced for full randomness : doesn't the normal setup favor scum normally ? Because in reality town is supposed to do better than random.

Are we really doing this ? This seriously makes me angry. Why are we even playing this game if you all would rather just press a button on random.org ?

Suppose his role has 50% chance to hit town and 50% chance to hit scum. Would you still not go for it?

[Ampharos]

My concern: we are doing nothing but advancing scum wincon in the majority of cases.  Logically to me, we should never use this power.  Townies will be dropping quickly enough as is.

[Teproc]

Even from a game theory PoV I disagree. I don't believe games are actually balanced for full randomness : doesn't the normal setup favor scum normally ? Because in reality town is supposed to do better than random.

Are we really doing this ? This seriously makes me angry. Why are we even playing this game if you all would rather just press a button on random.org ?

Suppose his role has 50% chance to hit town and 50% chance to hit scum. Would you still not go for it?

I'm not sure, probably not. My initial thought would be that the odds would need to be at least 60% in favor for me to be comfortable with it. I understand that might be wrong, and I would likely be fine being overruled in such a situation, though I'd probably still oppose it. As it stands though, the probability is much lower than that.

[faust]

So I need to figure out stuff. This has been a long day and I don't really know who to vote for anymore. From memory:

Witherweaver - he has been around a lot, huh? Also actually generating content I think. Vocal about lynching Ampharos. Seems pretty normal so far.
yuma - feels townier lately. I am profoundly bad at reading him, so. But if we can actually confirm him, that would be good. So no lynching  here I think.
silverspawn - very active. Active is good. I remember feeling bad about him early on, can't remember what that was about. Seems normal lately.
Teproc - I feel like he is defending a lot of people. I think that's kinda his thing, do I get that meta correctly? I feel like his play is very pro-town. Not lynching.
iguanaiguana - huh. There was some fighting. I think he came out of that somewhat townie-looking. Seems much like in RMM28. I think he should stay alive.
Hydrad - I could lynch Hydrad, I don't remember any of his stances. I feels cheap though. Well, what can you do?
Roadrunner7671 - Gut says town. My gut has been horribly wrong before with him. Still, he's not a preferred candidate.
2.7... - seems eager to push the game forward. Also I want him to live to D2 for once. So sympathy pass.
faust
Ampharos - still leaves a scummy feel. I should probably reread to see what's really there. I like this lynch, but it's so popular.
EgorK - unusually involved after some lurky time. That's good, though he does a lot of theory talk and less scumhunting.
Awaclus - his thing with yuma is cheap and allows him to do nothing else. Doing nothing else is anti-town. He also does not make a ton of sense to me, which is always a red flag.
Haddock - I do not know. I don't really know how to tell scum!Haddock from town!Haddock. The two games I've seen him in he felt pretty much the same.

Strikethrough for not lynching, bold for preferred, in case that's not clear.

PPE: 6

[Ampharos]

Do not forget about the potential to lose a town powerrole to the random yuma kill.

[Witherweaver]

scum!Teproc is pretty pro-town, I think.

[Teproc]

scum!Teproc is pretty pro-town, I think.

I'm like a reverse Awaclus.

@faust : We can't confirm yuma just by looking at how many die, because we don't know how many scum factions there are.

[faust]

So from a game-theory POV, I probably agree with what Egor said - shooting will probably be beneficial. I mean, I don't like the role, but it also has the plus side of semi-confirming yuma, yes? Because scum can't have two night kills. He could be third-party I guess, but he's be confirmed non-mafia.

Alright, unvote I think... this is not the worst plan ever.

There might be two scum factions (either mafia+SK or mafia+werewolves)

Well, yes... so what? If yuma is second scum faction, then this debate is moot anyway. And if he's town and there are two scum factions, that makes our plan all the better since it increases the chance to hit scum.

[Roadrunner7671]

And secondly, I am not sure how much use you are to town if you come back here, but refuse to scumhunt (scumhunting kinda necessitates reading everything).

But if we are lynching based on not reading the thread, we need to lynch RR here (and probably every game) because he's constantly posting stuff like this:

Why did you come back?

This. Not sure how I feel about this. It's pretty true, but I feel like I shouldn't like it.

[silverspawn]

Okay, let me say something about the shoot thing.

This is a complex issue and nothing you say is likely to come anywhere close to modelling it accurately. First of all you need a model for how likely scum is lynched over town, which is already super difficult. With random lynches and no night powers, town wins  20.77% of the time. If e.g.  you assume that each scum player is 1.1 / 1.3 / 1.5 / 1.7 / 1.9 / 2.1 times as likely to be lynched as town on days 1 / 2 / 3 / 4 / 5 / 6, town still only wins 41.86% of the time.

The real problem, though, is that the town winrate increases in big part because of investigative results, which are completely impossible to model because the setup is closed.

There are times when trying to talk about maths is productive in mafia. This is not one of them. I suggest that we don't discuss this further and yuma doesn't shoot, or if people think he shoots he can shoot, but don't argue with maths.

That's why I studied math and not politics. And I stopped to read that topic when it had gone from 2 pages to 5 in like an hour. I still firmly believe that given the assumptions Haddock made his answer is right. Assumptions may be wrong though

Haddock himself realized he was wrong, but if you want to believe it despite not having read the thread, sure.

[faust]

scum!Teproc is pretty pro-town, I think.

I'm like a reverse Awaclus.

@faust : We can't confirm yuma just by looking at how many die, because we don't know how many scum factions there are.

Well, it still gives us information that can be used to our advantage. I would argue that scum!yuma can only make this fakeclaim and shoot if he's not in the primary scum faction, so at least we can eliminate partner interactions down the road and yuma's chance of being scum drops.

PPE 2

[faust]

There are times when trying to talk about maths is productive in mafia. This is not one of them. I suggest that we don't discuss this further and yuma doesn't shoot, or if people think he shoots he can shoot, but don't argue with maths.

The pony talks the sense.

I use the Checkov's gun argument: Because the role exists, we should use it.

[Teproc]

scum!Teproc is pretty pro-town, I think.

I'm like a reverse Awaclus.

@faust : We can't confirm yuma just by looking at how many die, because we don't know how many scum factions there are.

Well, it still gives us information that can be used to our advantage. I would argue that scum!yuma can only make this fakeclaim and shoot if he's not in the primary scum faction, so at least we can eliminate partner interactions down the road and yuma's chance of being scum drops.

PPE 2

Except this could be 2v2 multiball. Not to mention that other vigs could exists, bombs, etc.

[Teproc]

There are times when trying to talk about maths is productive in mafia. This is not one of them. I suggest that we don't discuss this further and yuma doesn't shoot, or if people think he shoots he can shoot, but don't argue with maths.

The pony talks the sense.

I use the Checkov's gun argument: Because the role exists, we should use it.

That is perhaps the only argument I could see in favor of using it : if ash put in the setup, it's not because he thought the best play was not to shoot 100% of the time. Still firmly against it though.

[silverspawn]

There are times when trying to talk about maths is productive in mafia. This is not one of them. I suggest that we don't discuss this further and yuma doesn't shoot, or if people think he shoots he can shoot, but don't argue with maths.

The pony talks the sense.

I use the Checkov's gun argument: Because the role exists, we should use it.

I don't know the argument, but doesn't it more say 'if it exists, there are times when using it is good' and not 'if it exists, using it right away is good?'

Using it right away is just super swingy. It gets a bit better if people claimed stuff, and maybe if townies die.

Also note that yuma could be mafia or SK. If we let him use it, he gets a perfectly acceptable excuse for shooting someone.

[EgorK]

That's why I studied math and not politics. And I stopped to read that topic when it had gone from 2 pages to 5 in like an hour. I still firmly believe that given the assumptions Haddock made his answer is right. Assumptions may be wrong though

Haddock himself realized he was wrong, but if you want to believe it despite not having read the thread, sure.

That means he had wrong model (assumptions), not that his math after that was wrong

Anyway, let's move on. I will do promised e reread right now

[EgorK]

By the way, I think RR flavor claim had gone by mostly unnoticed. I'm not even sure if it was claim or joke

[Witherweaver]

There are times when trying to talk about maths is productive in mafia. This is not one of them. I suggest that we don't discuss this further and yuma doesn't shoot, or if people think he shoots he can shoot, but don't argue with maths.

The pony talks the sense.

I use the Checkov's gun argument: Because the role exists, we should use it.

That is perhaps the only argument I could see in favor of using it : if ash put in the setup, it's not because he thought the best play was not to shoot 100% of the time. Still firmly against it though.

That's not really any different than the balancing argument.

At any rate, I still support Yuma using it.

[Witherweaver]

By the way, I think RR flavor claim had gone by mostly unnoticed. I'm not even sure if it was claim or joke

I think it was a claim. 

[faust]

I don't know the argument, but doesn't it more say 'if it exists, there are times when using it is good' and not 'if it exists, using it right away is good?'

Using it right away is just super swingy. It gets a bit better if people claimed stuff, and maybe if townies die.

Also note that yuma could be mafia or SK. If we let him use it, he gets a perfectly acceptable excuse for shooting someone.

Well, if you can lay out to me a situation when using this is better than using it N1, go ahead. I think if we use it N1, we can utilize whatever information we get from it for the longest period of time.

Also if yuma is scum, I highly doubt he will not shoot just because we tell him to. At least he would be forced to shoot someone "random", because it will still be scummy if he shoots top town reads.

Also if there is a mafia team that has reason to believe that there is another scum team out there, they will kill yuma sooner or later, and the problem takes care of itself.

[faust]

That's not really any different than the balancing argument.

It is different in that it sounds way cooler.