Print

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

[silverspawn]

*that you're wrong

[A Drowned Kernel]

We knew there was no ub cuz our strongman was one shot so MMM was impossible

[Axxle]

We knew there was no ub cuz our strongman was one shot so MMM was impossible

Yeah, it was safe to fakeclaim UB, but I think we're arguing if it was safe for scum to fakeclaim masons if ss and chairs were scum.

[Axxle]

well given that many people here reject maths to different degrees I can't claim that you're right. Still, it is clearly not 'safe'.

I guess we're just arguing the semantics of "safe". I agree with you it's a very risky play.

[silverspawn]

We knew there was no ub cuz our strongman was one shot so MMM was impossible

Wait... wait. waitwaitwaitwait. No. What we're discussing right now doesn't make sense. Whether or not there is an UB depends on the Masons.

So here is the deal. If there is no UB, that confirms the masons, because a single M roll is the only way there is no UB. So, in your scenario where there are no masons and scum wants to fakeclaim them, there is always an UB. In a scenario where there is an M and scum fakeclaims masons, you have two claimed mason teams.

So it always subtracts 3 T's, meaning that scum can only fakeclaim if there are 5 or 6 real Ts around.

[silverspawn]

As seen in this game: there were masons, scum fakeclaimed UB - same probabilities - I immediately called RR scum despite having had a town read on him.

[faust]

At no point in this debate did I get the impression that silver is actually considering being wrong.

[Awaclus]

The other important point is that there is no reason to lynch one of the masons before at LyLo. Literally zero- I said this a million times in the game thread, and it's 100% true. So even if town assumes masons are lying, you still have a shot at lynching one player, and if he flips scum, you're fine.

It is 100% false and it doesn't become true no matter how many times you repeat it without addressing my arguments that actually show why it is false.

[Axxle]

The other important point is that there is no reason to lynch one of the masons before at LyLo. Literally zero- I said this a million times in the game thread, and it's 100% true. So even if town assumes masons are lying, you still have a shot at lynching one player, and if he flips scum, you're fine.

It is 100% false and it doesn't become true no matter how many times you repeat it without addressing my arguments that actually show why it is false.

Can you link those arguments?

[Awaclus]

The other important point is that there is no reason to lynch one of the masons before at LyLo. Literally zero- I said this a million times in the game thread, and it's 100% true. So even if town assumes masons are lying, you still have a shot at lynching one player, and if he flips scum, you're fine.

It is 100% false and it doesn't become true no matter how many times you repeat it without addressing my arguments that actually show why it is false.

Can you link those arguments?

Sure.

http://forum.dominionstrategy.com/index.php?topic=14803.msg577444#msg577444
http://forum.dominionstrategy.com/index.php?topic=14803.msg577463#msg577463
http://forum.dominionstrategy.com/index.php?topic=14803.msg578052#msg578052
http://forum.dominionstrategy.com/index.php?topic=14803.msg578066#msg578066

[Witherweaver]

well given that many people here reject maths to different degrees I can't claim that you're right. Still, it is clearly not 'safe'.

This is another example of your implicitly condescending tone.  No one here is 'rejecting maths".  Someone here (you) are attempting to draw invalid conclusions from some tangentially relevant arithmetic.

[silverspawn]

At no point in this debate did I get the impression that silver is actually considering being wrong.

Well I already admitted to being wrong in the manner debate. The other debate; I laid things out in more detail.


Can scum fakeclaim Masons safely?

To simplify, we can assume scum always has full information over the setup. That only benefits them.

Now, different cases:

Case 1: There has been 1 M rolled.

This means there is no Universal Backup and 2 Masons. In a massclaim, there are 2 claimed mason teams. There is no setup with 2 mason teams and no UB, therefore one of the teams must be lying

Case 2: There have been 3 Ms rolled.

This means there is a Universal Backup and 2 Masons. In a massclaim, there are 2 claimed Mason Teams. The only setup that allows for 2 Mason teams is MMMMMM. If any other PR has claimed, you know that someone must be lying. If not, you either play MMMTTT or MMMMMM. Odds that it is MMMMMM = 8/1000 < 1%

Case 3: There have been more than 3 Ms rolled.

In a massclaim, this creates an impossible setup, because the fakeclaim requires 3 Ms and there are only 6.

Case 4: There's been 2 or 0 M's rolled.

Here we have a UB and no real masons. In a massclaim, the setup that town is presented with has exactly 3 fewer Ts than the real Setup.

So if the real Setup has

0 or 1 or 2 Ts -> impossible
3 Ts -> Two setups presented are {the3PRs}MMM vs {same3PRs}TTT. Odds for the first are again 8/1000
4 Ts -> Two setups presented are {the2PRs}MMMT vs {same2PRs}TTTT. Odds for the first one are 32/1000 = 3,2%
5+ Ts -> a much higher number

Conclusion:

Scum can only fakeclaim masons safely in a setup with 5 or 6 Ts. In other setups, it is either guaranteed or extremely likely that one claimed PR is lying.

... okay?

[silverspawn]

well given that many people here reject maths to different degrees I can't claim that you're right. Still, it is clearly not 'safe'.

This is another example of your implicitly condescending tone.  No one here is 'rejecting maths".  Someone here (you) are attempting to draw invalid conclusions from some tangentially relevant arithmetic.

I'm sorry, I've come across numerous examples of people rejecting maths and its implications in mafia.

[Witherweaver]

We already knew all this long ago when I said that Mafia knows whether or not they can safely claim. 

[Witherweaver]

The calculations involved here, like every other time, are highly irrelevant, because they take into account none of the other information available.  It's, basically, ignoring thousands of posts of information and pretending the given game state only depends on two or three posts.

[Witherweaver]

And really can we stop calling this math every time it comes up?  Because it isn't.

[silverspawn]

We already knew all this long ago when I said that Mafia knows whether or not they can safely claim.

So do you agree with what I said? One thing that means is that scum could not fakeclaim safely in the game that we just played.

[silverspawn]

The calculations involved here, like every other time, are highly irrelevant, because they take into account none of the other information available.  It's, basically, ignoring thousands of posts of information and pretending the given game state only depends on two or three posts.

See: this is what I mean when I say players are rejecting maths. I didn't say they do it for no reason. Obviously they think they have a reason. But they are rejecting it, so why are you calling me out for saying it? I also reject things. I won't call you out if you say that I reject religion. Your quote above is you rejecting my maths.

Anyway. The reason why this all works despite ignoring all the other stuff you talk about is that all the other stuff does is amount to reads. Reads only change accuracy by a few percent from what is statistically random.

So let's say the statistically random is 50% and you have a strong read on someone. Let's say that the chance for your read to be wrong is only 1 in 5. That is a super accurate read. If town had those reads, we would win almost every game. But let's say that you have it.

Now let's say that the read favors the player who is being discriminated by the setup. Now you have a 1/5 = 20% chance for your read to be inaccurate, and on the other side a 3% chance for the setup to have rolled the unlikely way. And you know one of those is true. The odds that the setup ended up rolling the inaccurate way is 3%/20% = 15%.

This is how it works. This is why it works. Denying that is rejecting maths. And you do not even have this level of accuracy in normal reads.

[Axxle]

We already knew all this long ago when I said that Mafia knows whether or not they can safely claim.

So do you agree with what I said? One thing that means is that scum could not fakeclaim safely in the game that we just played.

Look, I don't think any of us actually care if it was "safe" or not, just if it's possible. This is because Mafia is a psychology game not a math game.

Also, not everyone can follow complicated math that well (including me). I keep reading your math posts but it seems like you keep using variables that don't matter. Like why include the cases with 0,1,2 T? Of course scum can't fakeclaim masons there. Why include 5 or 6 Ts? You claimed day 3 when everyone knew of the B and V rolls. Like I read it and am concerned that you're pulling the wool over my eyes somehow.

[Witherweaver]

I am not rejecting math.  Math is about understanding the essence of something.  What you are doing is not math.

[silverspawn]

We already knew all this long ago when I said that Mafia knows whether or not they can safely claim.

So do you agree with what I said? One thing that means is that scum could not fakeclaim safely in the game that we just played.

Look, I don't think any of us actually care if it was "safe" or not, just if it's possible. This is because Mafia is a psychology game not a math game.

Also, not everyone can follow complicated math that well (including me). I keep reading your math posts but it seems like you keep using variables that don't matter. Like why include the cases with 0,1,2 T? Of course scum can't fakeclaim masons there. Why include 5 or 6 Ts? You claimed day 3 when everyone knew of the B and V rolls. Like I read it and am concerned that you're pulling the wool over my eyes somehow.

well I wanted to be thorough, accounting for all cases. And I'm not even that good at maths. I make mistakes all the time; SP and Egork are both much better than I am. But I at least understand the principle.

You or anyone else doesn't have to follow it, but when WW accuses the pan of being bad and then faust accuses me having confirmation bias, well what am I supposed to do? I can either just take it or provide evidence for the contrary. I believe in evidence.

Also, the latest point of debate with WW is a recurring theme that has been relevant a lot of times in the past. Someone makes a math argument, someone else says that doesn't work because you're ignoring this and that, I've heard it before. You can weigh reads and setup probabilities against each other, that's why setup probabilities that are so clear should be treated identically to proofs. 97% should almost always be treated as 100% in this context. I guarantee you that it would lead to better results if everyone did it.

I am not rejecting math.  Math is about understanding the essence of something.  What you are doing is not math.

believe what you want then

[yuma]

The calculations involved here, like every other time, are highly irrelevant, because they take into account none of the other information available.  It's, basically, ignoring thousands of posts of information and pretending the given game state only depends on two or three posts.

You kept accusing me of this.... Except... I was taking into consideration those posts, as were others. And then applying the math to it. I don't know why you felt I wasn't? Maybe because I wasn't agreeing with you so thereby I must not have been taking into consideration the math?

I don't know. That really isn't fair.

[ashersky]

I don't know, or particularly care, if the "math" is math, or even right.

If there is a 3% chance that something occurred, and I decide as scum to try to convince everyone that it happened, that's a valid, if reckless plan.

If I'm town in the same scenario, does it make me more likely scum because it only happens 3% of the time?  Nope.  When I'm town, I'm town.  And yet, regardless of reads and play, ss's argument would be to lynch me as a fake claimant.

And if I fake claim town as scum (higher probability of rolling town than anything else) you don't automatically buy that just because of math.

I think that's the point "against math."

[Awaclus]

If I'm town in the same scenario, does it make me more likely scum because it only happens 3% of the time?  Nope.  When I'm town, I'm town.  And yet, regardless of reads and play, ss's argument would be to lynch me as a fake claimant.

The point is that the expected number of correct lynches when you lynch you in that situation is 0.97. That's like 2-4 times the expected number of correct lynches when lynching someone else in that situation. 3% of the time, it ends up being a bad idea, but simply because it is a good idea 97% of the time, it's worth doing every time unless there is an extremely good reason to believe otherwise.

[ashersky]

If I'm town in the same scenario, does it make me more likely scum because it only happens 3% of the time?  Nope.  When I'm town, I'm town.  And yet, regardless of reads and play, ss's argument would be to lynch me as a fake claimant.

The point is that the expected number of correct lynches when you lynch you in that situation is 0.97. That's like 2-4 times the expected number of correct lynches when lynching someone else in that situation. 3% of the time, it ends up being a bad idea, but simply because it is a good idea 97% of the time, it's worth doing every time unless there is an extremely good reason to believe otherwise.

So could you just figure out the percentages for every player in a given game and then lynch by number?