Mafiascum.net

The only person who seriously bought the innocent result was JSU though.

If we hadn't used PRs on n2 completely ideally, we would have lost.

In post 8525, AngryPidgeon wrote:The only person who seriously bought the innocent result was JSU though.

Maybe. There seemed to be a general consensus to leave sorting you for later or something.

I blame myself in large part for ever backing off you after that inno result claim on day 2. Your day 1 claim was skeevy and the day 2 result claim was skeevy squared.

Anyway. gg scum!

And thanks for the game, Cabd!

/pre-in for whatever you're cooking up next!

I'm not saying I think the setup was imbalanced (I haven't actually looked at it in detail), I'm saying godfathers suck.

And ffery, it's fine to not trust cop results, but you can't ignore the result outright. There's probably a way to formally algorithmize it based on decision-making theory, but I'm too tired/lazy to think of it now.

In post 8528, BROseidon wrote:I'm not saying I think the setup was imbalanced (I haven't actually looked at it in detail), I'm saying godfathers suck.

And ffery, it's fine to not trust cop results, but you can't ignore the result outright. There's probably a way to formally algorithmize it based on decision-making theory, but I'm too tired/lazy to think of it now.

There probably is, but reads. :/

No you don't and don't ignore the result outright. It's one data point about the player - potentially a very important data point. Other data points include whether their claims make sense, whether their play fits their wincon and role, whether other claims conflict with theirs, whether it makes sense that they keep showing up alive the next game day, etc.

There's a whole cluster of behaviors that tend to follow on after becoming all-but-cleared town.

Decision-making theory should take into account reads. I've been trying to develop a way to make decision in-game based on associative reads in any given gamestate (i.e., if x is scum, then y has b chance of being scum, but if z is scum, then m has n chance of being scum, etc), and I'm not sure how doable that is without using external software because of how large the decision tree would become. A cop inno affects the probability of someone being town, regardless of your personal read on that player.

In post 8530, BROseidon wrote:Decision-making theory should take into account reads. I've been trying to develop a way to make decision in-game based on associative reads in any given gamestate (i.e., if x is scum, then y has b chance of being scum, but if z is scum, then m has n chance of being scum, etc), and I'm not sure how doable that is without using external software because of how large the decision tree would become. A cop inno affects the probability of someone being town, regardless of your personal read on that player.

The specific effect on probability has a lot to do with other aspects of a game's design, the mod's experience both as a mod and a player, etc. The player playing like a scumfuck is pretty objective.

I followed you on AP because of your experience playing with him. I thought you would be able to read him better than me. I didn't follow you because of your probabilistic models. I have a lot of learnings from this game. One of them is that I need to actually find out what you're basing your assessment on a player.

One of the things I'm mulling over is how my lack of strong scumreads affected how people were reading me. I think I mentioned in the game that if I were scum I would have just spit out a few scumreads. It would have been easy as scum to sheep some of the town players on the wrong track. I kinda think it would have been better to pretend a confidence I didn't actually have about some of my not-so-town reads. But, that goes against some deeply ingrained habits and principles as a town player.

In post 8531, fferyllt wrote:
One of the things I'm mulling over is how my lack of strong scumreads affected how people were reading me. I think I mentioned in the game that if I were scum I would have just spit out a few scumreads. It would have been easy as scum to sheep some of the town players on the wrong track. I kinda think it would have been better to pretend a confidence I didn't actually have about some of my not-so-town reads. But, that goes against some deeply ingrained habits and principles as a town player.

So do what I do and get universally read as town!

In post 8531, fferyllt wrote: One of the things I'm mulling over is how my lack of strong scumreads affected how people were reading me. I think I mentioned in the game that if I were scum I would have just spit out a few scumreads. It would have been easy as scum to sheep some of the town players on the wrong track. I kinda think it would have been better to pretend a confidence I didn't actually have about some of my not-so-town reads. But, that goes against some deeply ingrained habits and principles as a town player.

I thought this was insane and was what actually had me suspecting the people who were pushing you because to me it made perfect sense for you not to have scum strong scum reads, but I felt like of anyone I was seeing the game through a similar lens so I could see it. And, I think it's super easy as scum to give a fake scum read, especially in a game like this, so you not having one is something that I was lost why people had a hard time with it.

It just looked really honest to me, not like scum afraid to give a scum read. If that makes sense, but I guess it's because I was town reading you and they weren't?

~~~~~

Thanks for running it Cabd!

Congrats scum team.

To me it felt weird that you wouldn't be scum reading people when so few people were particularly obvtown.

That's what null piles are for. :/

Over time, I feel like my criteria for genuinely scumreading players has gotten more and more rigorous.

In post 8534, BROseidon wrote:To me it felt weird that you wouldn't be scum reading people when so few people were particularly obvtown.

Not obvtown =/= scummy

I'd like to point out that my town reads this game were 100% accurate (even if I only had 1 )

In post 3405, Lord Business wrote:Player Roster:
1 Kagura (borkjerfkin + nachomamma8) - weak scum
2 The Fox and the Hound (cephrir and DV) - null
3 Yggdra Union (giffy and pieguyn) 3 Yukari Yakumo - null
4 orcinus_theoriginal - town
5 Titan (Tammy + Sir Arthur Dane) - town
6 Rancid Broderick Drake (zmuffinman and natirasha) - town
7 MastinSSK (Mastin2, MafiaSSK) - town
8 Breakfast With Stalin (Hydra of ffery and beli) - town
9 AngryPidgeon 9 ElementalHawk (Prohawk+3dicerolling) - scum
10 Carbon Fiber (FourTrouble and F16) - weak town
12 Just Sheep Us (Broseidon and Desperado) - weak town
13 Red Gyarados (Brian Skies and notscience) - weak town
14 Cupcake Panda - scum
15 PeregrineV - weak scum
16 Clyton - weak scum
17 Mac - town

I have a couple there I need to make a call on so will ISO them when I have time.

Seems this guy knew what he was talking about didn't he...

No he didn't he called me scum!

http://forum.mafiascum.net/viewtopic.php?f=4&t=58149

<.<
>.>

I approve of the above message.

In post 8539, Cephrir wrote:http://forum.mafiascum.net/viewtopic.php?f=4&t=58149

<.<
>.>

In post 8540, Cabd wrote:I approve of the above message.

In post 8542, Cephrir wrote:

you should consider approving of it enough to sign up, PV

I'll probably signup to replace, need to let some of my gameload fall off first.

Well you can't replace into a game if it doesn't run~

In post 8544, Cephrir wrote:Well you can't replace into a game if it doesn't run~

I'll pimp your game over the next few days.

hey, Yulia. Any chance you can update the post with flips so someone trying to follow along with this behemoth can figure out what the heck happened?

Also, as a favor to me, in future games could you post people's FULL names in the player list (with hydras) so that I can easily send out the survey? Not a huge deal to anyone playing, so it's not required, but it'll get you more accurate survey results.

In post 8515, Kagami wrote:You can't even say, "O, he's just trolling now, he must be a godfather!" because peoples' playstyle tends to change radically even as town once they've been inno'd.

Except it's AP. He explicitly was trolling. (This is why lynching the player who can actually READ AP is a Very Bad Idea

(TM)

.)

In post 8523, fferyllt wrote:This is some good reinforcement for why I value reads over nearly anything else in mafia most of the time. I learned to play mafia in an environment where something like 90% of the games had godfathers.

I, too, typically trust reads over roles. (And while I grew up on MS.net, I also play mafia on a site that doesn't go a game withOUT a role called 'Godfather', though because roles aren't standardized there, what said Godfather does changes each game, but having investigation immunity happens an average of 50-75% of the time.)

The correct approach to having a result is of course to bring it forth, but also explain why you think what you do about the result: its validity or falsehood. This is one of the reasons I placed so much hope on F-16; he actually did exactly this--analyzed the result, and explained exactly why he thought AP was scum exploiting it for all its worth. (He was.) The result was not blindly discarded, which would happen if in spite of the result, people pushed AP without thinking of it and maybe without even mentioning it. (Objectively bad play, even if correct in this particular game.) The result should also have continued to not be blindly followed, ESPECIALLY given that non-conftown were dieing in droves over allegedly-conftown. (PV and Beli both had suspicion on them and were not conftown. Why'd they die over AP who allegedly was? Answer: because AP was scum.)

In post 8524, AngryPidgeon wrote:The town power wasn't trivial though lol. If I got lynched d2, scum would have lost. If we hadn't used PRs on n2 completely ideally, we would have lost. Almost 100% guarantee.

This, too.
Blaming the setup is a cop-out. (Ha.) It's easy to say that it's the setup's fault. Just like it's easy to say that it's not the town players' on a mislynch fault. It's easy to place the blame elsewhere. And perhaps some exists. (Even though I think the game was balanced.) But you'll never learn from your mistakes if you always think the fault lies elsewhere.

In post 8546, zoraster wrote:hey, Yulia. Any chance you can update the post with flips so someone trying to follow along with this behemoth can figure out what the heck happened?

Also, as a favor to me, in future games could you post people's FULL names in the player list (with hydras) so that I can easily send out the survey? Not a huge deal to anyone playing, so it's not required, but it'll get you more accurate survey results.

Yeah I can do this hang on.