Should a vig who shoots N1 be Modkilled?

[LlamaFluff]

Modkilled? No

Viewed as anti-town? Yes.

Vigs are one of those roles that will basically lose or win a game for town every time. I just see it being lose the game far more than win the game, and given how inaccurate vigs are (using an RNG is more accurate than vigs on the site) the case for them not shooting is actually there.

Worst is a player who thinks shooting as a one shot vig in a setup that has an odd amount of players is a remotely good thing, THAT is almost playing against your win condition to shoot. I would think if you had three opens

Setup A: 1x One shot Vig, 8xVT, 2x Goon
Setup B: 9x VT, 2x Goon
Setup C: 1x Named Townie, 8x VT, 2x Goon

Setup C would have the highest town win rate. A and B would be neglegably close in win rates.

[LlamaFluff]

also if you don't shoot as a vig you are a wimp who can't trust your own reads and if you can't bring yourself to believe in your own reads then lmao what are you doing

Because vigs are anti-town... if it brings and leaves the game at odd players then its an okay shot. In any other situation you are probably costing the town a lynch unless you will 100% get two shots off, and even then you are removing a lynch for two vigs which strips town of quite a bit of information.

I think last time it was looked at vigs and no EV for town in balanced 13P setups, so they are not pro-town in any way. It basically is going to increase the chance of scum crushing win by a lot, and town crushing win by a little. A shot with no claim, no further interactions and no discussion is not going to be too accurate, and stats back that one up big time. The only way you can argue that a vig is good is that it decreases the lynch pool, but even then half the time the lynch pool shifts so continuously during the game that its even hard to really say thats a good thing.

Best thing a vig can do is get the game to odd numbers, anything else has huge chances of damaging town for low reward. Even if you want to say scum make up 2/3 of the game at that point, chances are twice as good you hurt town than help town... so how is that a smart move to ever make?

[LlamaFluff]

I guess look at my three setups:

With a vig, if they hit town they get confirmed but are at an even number of players, so if the shot was early chances are endgame looks something like 3:1 with no clears (best case 3:1 with a clear). If they hit scum, same boat.

With just claiming vig though, especially if they keep quiet about it long enough to make it late, 2:1 has a confirmed town in it. So town chances to win increase in the endgame by 17% because the vig didnt shoot.

So if the vig doesnt shoot, +17% for town in endgame. They also go up from five lynches to six.

Its an extreme case, but you are sacraficing 17% endgame odds and one lynch by shooting in that setup.

[LlamaFluff]

LLAMAFLUFF

DEMON CORE IS CALLING YOU

I'm confused....

Im one of the most well known "Vigs should never shoot" players.

N0 in a large game is one of those times that is acceptable to policy vig someone though because its almost impossible to tell what the game will end up looking like and there are players who will decrease town chances of winning even if they are town. Outside of those only spots where I think vigs should shoot would be: confirmed scum, forced win path, game will stay at odd players post shot.

Even the "two vig shots vs lynch" arguement I dont love for reasoning to shoot. Lets say even 70% of the time the vig gets two shots, 30% of the time you are changing a lynch into a vig shot, and thats a really bad 30%. Im going to see if I can find my old 'vig accuracy' numbers though because those were really eye opening for how hazardous to town vigs are in normals. You literally could just kill the first player alphabetically and kill more scum then what the average vig does.

Cop is far far far stronger than a vig. I would put cop on equal ground of vig+1 or 2 VT in a closed. Something like Cop+8VT vs 2 Goons I think town wins more than Vig + 10VT vs 2 Goons

[LlamaFluff]

I found some of the old data, think this was the first 100 or so MN games after we started checking them, it needs an update which I will do eventually. A good vig may help the town, but I think more often then not the vig ends up doing more harm than good.

N1 vigs Shoot Scum 12.5% of time
N1 random vig is 25.9%

N2 vigs shoot scum 0% of time
N2 random vig is 27.9%

N3 vig shoots scum 0% of time
N3 random vig is 16.7% of time

Total vig hits scum 7.1% of time
At random it should be 26.2% of time

[LlamaFluff]

This thread is bad and you should feel bad.

Thanks for posting those stats Llamafluff.

It seems that shooting N1 is the only time a vig should shoot.

Ehhh... stats are limited especially since given small sample its more like 0/8 or so for N2 and 0/4 for N3+ which is why I said they need updating. Its just one of those little stats that shot the only spot you can really try and sell the vig advantage is removing mislynches which shift enough to not be a reliable arguement to me

[LlamaFluff]

Did doomyoshi draw vig and end up misvigging every night? Is that why he has a bad attitude toward the role?

Or he could be like me and just realizes a "small chance high upside - medium chance high downside" role shouldnt be used.

My theory is that's because vigs take too many hero shots and miss

That sounds about right. If vig falls into the hands of a poor player it may actually have a worse than random hit rate because they'll go after their incorrect suspicions that the town doesn't agree with them on.

That wouldnt explain my data though. Its far more common than not for the vig to be worse than random, so claiming this means you claim that the average player is worse than random. Hero vigs are an issue, but the bigger issue is that you are skipping a few major aspects of the lynch that make it better than a vig kill. There is no claim, no wagon, no tells during votes... I think we all have wagoned someone and then suddenly went "oh yeah they are town" due to something that happens. Vig removes any chance of that.

I would bet vig actually hits PRs at same rate as scum. Even if you want to be nice and say "choose one setup"

Setup A: 10x VT, 1x Named Townie, 2x Goon
Setup B: 10x VT, 1x Named Townie, 2x Goon - One player a night randomly dies, 80% chance this player is town, NT cant die

Which would you rather have

[LlamaFluff]

If all the randomly dying players in setup B were bad players/ players agreed to be scummy then setup B.

That doesnt happen though, and for me its usually more frustrating since I tend to be really good with townreads and can get wagons off players. If they just get killed it really damages what players who play heavily on townreads and defense can do. Really though, vigs tend to love to shoot the middling reads group who they think is scummy. Its normally possible to tell the vig from scum kill, but at times its not that easy or at least you are not completely sure.

What are the stats for lynching scum the day after a vig shot instead of after a night without one?

End results are whats important really... if a town kill doubles the chance for a scum lynch, then town should still have a higher EV when vigs are around. The fact that the EV doesnt move really makes it not something of consequence because there is no long run difference.

[LlamaFluff]

if you look at mafia in a really stupid way, all you are doing as town and as scum is reduce the lynchpool for the last day. Come the last day, scum wants as many bad townies (possible mislynches) still in the game, while town wants as little as possible. It's the last lynch that decides the game. Vig kills and lynches aren't too different from that point of view, and having twice as much influence as town on who's there at the last day is pretty huge.

Way I still see it you take away the chance of the lynchpool player being called town for some reason instead. Lets say that through role claims, tells and other stuff 30% of the lynch pool is going to be called town instead at some point in the game. That means if the vig shoots twice there is about a 50% chance that they kill a player that would become town at some point in the game.

The vig kill is massively different from the lynch because there is no claim and no way for any player who has a different read to make the case of why you shouldnt be touching them. Vigs are a high risk high reward type role, and good town doesnt need a high risk to win a game - if they do the balance isnt there. I would take a VT over vig as town any day of the week and a vig over VT as scum any day. Far far more I see games where the town would have done better with no vig then games where the town won due to the vig. Best thing a vig does is basically confirm its existance.

I guess here's the most striking example. Imagine that the game is 3:1, and you're town and have a day ability that kills a random townie. Should you use it?

Thats a horrible example because thats one of the only situations where you should use a vig. What if its 4:1 and you have the ability to kill a random townie? What if its 6:1?

[LlamaFluff]

The only way a cop can fail is if the scum kill him early.

See Llamarble's post from earlier. There are downsides to a cop that a vig doesn't have.

This is just a bad arguement.

Lets say you are a cop in a 13P game where there is one role that will give you a false result (if its more than one mod is doing something very questionable) and a VT was lynched day one

You have a 10/11 chance of getting a true result (91%)
You have a 1/11 chance of getting a false result (9%)

Even if you extend this to another VT lynched and your result doesnt die N2

7/8 chance of true result (87.5%)
1/8 chance of false result (12.5%)

So...

79.6% chance two true
20.4% false result

I like that....

Even if you want to say something like 24P with two false:

21/23 True N1
2/23 False N1

17/19 True N2
2/19 False N2

Thats even HIGHER chances of two accurate results.

Vigs are swingy. If you are at all wrong, which chances ALWAYS side to, you are going to damage the town to some extent. Its why in more games then not with a vig they played poorly then were a key part of the win. I would without question take a VT over a vig without relooking at balance, I would heavily consider taking an empty slot over a vig and probably would take it.

[LlamaFluff]

i feel like that would harm the vig accuracy simply because it would lead to the vig getting manipulated by people's overall reads + the voice of the scum

So double day is anti-town?

[LlamaFluff]

i feel like that would harm the vig accuracy simply because it would lead to the vig getting manipulated by people's overall reads + the voice of the scum

So double day is anti-town?

That's not even remotely implied. You could say he thinks a lynch is less accurate than a vig, but that'd not make lynching anti town. So...uh yeah?

Well if the arguement is that a vig shouldnt let their shot be concensus... concensus is basically a lynch creating a game like

You can lynch A and B, both flips occur at once (which might make for an interesting game really).

Given the INaccuracy of vigs, lynches are more accurate in normals. Maybe somehow this shifted recently, but I dont think it has.