Mini Normal 2119 [game over]


User avatar
GuiltyLion
GuiltyLion
he/him
Survivor
User avatar
User avatar
GuiltyLion
he/him
Survivor
Survivor
Posts: 13829
Joined: August 19, 2015
Pronoun: he/him
Location: Seattle, WA

Post Post #4125 (ISO) » Tue Mar 31, 2020 4:59 pm

Post by GuiltyLion »

late to the thread but gg everyone! Very well played by the scum team after a difficult D1. Micc I was calling for your head constantly in the dead thread up until LYLO with bji going down but I thought you played incredibly well especially at f3, I feel a little responsible for shading emps so much haha.

thanks for modding the worst! Was a lot of fun. Looking forward to reading the mafia thread
"I think I no longer believe in monsters as faces in the floor or feral infants or vampires or whatever. I think at seventeen now I believe the only real monsters might be the type of liar where there's simply no way to tell. The ones who give nothing away"
User avatar
GuiltyLion
GuiltyLion
he/him
Survivor
User avatar
User avatar
GuiltyLion
he/him
Survivor
Survivor
Posts: 13829
Joined: August 19, 2015
Pronoun: he/him
Location: Seattle, WA

Post Post #4126 (ISO) » Tue Mar 31, 2020 5:00 pm

Post by GuiltyLion »

In post 4067, Micc wrote:viewtopic.php?f=90&t=82253

my notes can be released. my favorite post upon reread:
In post 36, Micc wrote:Hmm if I was scum here I’d be playing such a good game
also I feel this so hard! When I'm town and the game is not going well I'm like "why can't my scum games ever feel like this, nobody is suspecting me despite me voting mislynch after mislynch".
"I think I no longer believe in monsters as faces in the floor or feral infants or vampires or whatever. I think at seventeen now I believe the only real monsters might be the type of liar where there's simply no way to tell. The ones who give nothing away"
User avatar
mastina
mastina
She/Her
False Prophet
User avatar
User avatar
mastina
She/Her
False Prophet
False Prophet
Posts: 16670
Joined: October 7, 2016
Pronoun: She/Her
Location: Between Snohomish and Monroe, WA
Contact:

Post Post #4127 (ISO) » Tue Mar 31, 2020 5:51 pm

Post by mastina »

In post 4123, Menalque wrote:However, the question is how long those things should last before good
play
comes back into things. Why was that slot caught in the first place? GL and icon doing an excellent job of scumhunting on D1, purely based on play and before PRs came into things.
I can answer most concerns with one simple point.

Games cannot be balanced off of what players actually end up playing the game--games can only be balanced in a void of theory and the theoretical outcomes of a game, wherein you have to account for how likely scenarios are purely by statistical numbers, pretty much. You can 'adjust' the numbers to account for things that by play are more likely to happen in reality than in theory/are less likely to happen in reality than theory. (Quintessential example; millers are in theory negative utility but in reality due to site policy of instantly claiming on D1 are net-neutral to possible positive utility in reality. This is what I mean by the sort of theory/pure statistical randomness vs. reality difference.)

But at the end of the day, regardless of whether you make the adjustments or not, you're still doing much the same. You're guessing at what things could happen in theory, how likely they are to happen in reality, but these things can and will differ from what actually ends up happening because when designing the game, you cannot know what players will be in the game and even if you did, you wouldn't know which players would hold which roles (which affects the balance of the setup significantly; different players holding different roles makes huge differences in how the game plays out--easily demonstrable by a 'what if' of you being the strongest town role and bj being the strongman and the Chemist slot being the relatively disposable encryptor).

And by every metric, you cannot balance for the extremes of setups. In the extreme end of a cop + tracker + vtx8, vs. scum rolecop scum roleblocker, it is possible that the town lynch the scum roleblocker N1, the cop guilties the scum rolecop N1, and the tracker guilties the scum goon N1--in that scenario, the town has a perfect win because they killed the strongest scum PR on D1, and both of their investigatives got guilties unimpeded.

But would you call, "Cop, Tracker, VT x 8, versus Mafia Roleblocker, Mafia Rolecop, Mafia Goon", a townsided setup?
Fuck no you wouldn't, that setup is scumsided as fuck because all of the town's strength is on two roles who have a high chance of not getting guilties, and the scum have two role-investigatives per night (kill+rolecop), and the scum have two role-shutdowns per night (kill+roleblocker), which gives them the ability to fish for three town players per night, more town players than there are actual town PRs in the game.

If you math it out, with a VT lynch D1, and three groupscum, scum between their kill, roleblock, and rolecop can target 3/9 town players...on N1. The scum can literally identify/shut down/control/etc. one third of the total town's members, on the first night, and each subsequent night only increases those odds. If a VT mislynch D2 and no PRs hit, then on N2 the scum have a 3/7 chance of hitting a PR, on top of the 3/9 chance from the night before giving them a much much much better idea of where to focus.

So I think you get where that's going.

You cannot account for the best possible play in setup design, for the radical extremes that are incredibly unlikely to happen, because those extremes are...well. Extreme. Statistically unlikely, nigh-on statistically impossible. The odds of the scenario I mentioned before, where the town lynches the roleblocker D1 and the cop guilties the rolecop and the tracker guilties the goon?
1/13 (roleblocker lynch) * 1/11 (cop guiltying the rolecop) * 1/11 (tracker tracking the goon) * 1/2 (goon doing the nightkill) * 4/5 (chance scum don't kill one of the PRs). You know what that comes out to be? 0.0769230769230769 * 0.0909090909090909 * 0.0909090909090909 * .5 * .8 = 0.0254291163382072% chance probability.

The chance of the town pulling that perfect scenario off is less than one tenth of one percent. But by play, by the players coming into the game--it could happen. Something with a percentage so small as to statistically be almost 0%, could happen and make that scumsided-as-fuck setup be an absolute town sweep, purely because of the playerlist, which players got what roles, and how the game plays out.

Do you get what I mean, here?

You cannot anticipate what the storm of players in the game, with the roles they have, will end up creating for balance. And extremes that are statistically close to 0% could end up happening, purely because of that.

But we can't balance for those near-zero-percent outcomes. It's literally impossible to do. You have to balance for the likely outcomes. You have to balance for what is probable to happen, or ideally, extending it to the
plausible
outcomes--yet it is humanly impossible to account for every
possible
outcome because there are literally thousands, maybe even millions, of possible outcomes for any given game. Unique combinations of lynches, role interactions, etc. span into amounts that are borderline-incomprehensible.

In the case of this game to bring it back to your complaint--yes town did well to identify the Chemist slot as scum even before PRs are taken into account, but I ask again; how would it be possible to anticipate that the town would scumhunt well before PRs are taken into account, without knowing which players are which alignment? How can you know that the game will have people like Slaxx and Iconeum as part of the town popping off on D1 and people like Chemist as scum that are easy to catch?

You literally can't.

If instead, the game had some rather mediocre town players in the game and, sayyyy, a don corelone candidate/winner as the slot Chemist was--then guess what? Then by play, the scum stomp on D1 and by play, scum gain a massive edge.

I suppose I can give you a final tl;dr on this.

The NRG fundamentally
cannot
balance around play, because play is determined by players and the alignments of said players...
...And that is a quality determined after the setup has long-since been finalized.
Because we CAN'T balance around play, we balance around 'best guesses' as to the probable scenarios likely to happen in any given game. Those best guesses tend to not factor in the absolute extremes that are statistical abnormalities, so scenarios such as the strongest scum PR being lynched on D1 (something statistically very unlikely to happen), aren't something we can balance around.

This is particularly true in games that are swingy as fuck. It is possible to design games that have minimal swing; these games have less extremes and are more moderate. But therein enters moderator discretion--moderators are allowed to run games that are incredibly swingy, so long as they are balanced, should they choose to. Alisae/the worst were informed this game would be swingy-as-fuck, but they were okay with that; they had the option to design a setup with less swing, but chose not to because this was the setup they wanted to run.

The NRG do their best to give the moderators the setup that the moderators want to run, while maintaining that it is balanced. We can't control every variable, especially not when the moderator wants to run a swingy setup, but moderators have that right.
(Granted, I wouldn't be opposed to an optional opt-in description for game setups to be announced in advance, where mods announce expected swing value of their setup, on a scale of "moderately low, moderately high, high, extremely high", but if you think that's a good idea, you should probably pitch it to implosion and such.)



It was very well played by town and deserved to confer advantages. But after that lynch we were on the back foot for most of the game, and if slaxx hadn’t self voted then winning would have been significantly harder regardless of how well we were playing and how well what we had was used. If the game can go, with one role out of 13 being lynched on D1, to the point of being unwinnable without significant luck and regardless of play on behalf of one side then I’m sorry, but it’s unbalanced in favour of whichever side has that advantage. Like imagine a world in which icon doesn’t misunderstand his result, Aaron doesn’t out, and/or slaxx doesn’t self hammer. Bji and I played well here, I used my role basically perfectly, and we would still have been absolutely losing an overwhelming amount of the time. I don’t see how that’s balanced. If the argument is that we should just have played better or harder — okay, maybe, but at what point do you get to an unrealistic expectation of your scum!players? Again, bji and I played pretty damn well here but those flukes of chance were still NECESSARY for us to turn this from a loss to a win.

Putting us at a major disadvantage for a couple of days? Sure, that seems fair. But by D4 the game still felt stacked in favour of town and that’s after we’d had great strokes of luck regarding early mass claim and lynching a PR. We also managed to get TR throughout most of the game, so if the answer is “scum should just have played better” — okay, how? What could we have done to be playing better here? I’m sure there are things, but fundamentally our play was sound.
User avatar
mastina
mastina
She/Her
False Prophet
User avatar
User avatar
mastina
She/Her
False Prophet
False Prophet
Posts: 16670
Joined: October 7, 2016
Pronoun: She/Her
Location: Between Snohomish and Monroe, WA
Contact:

Post Post #4128 (ISO) » Tue Mar 31, 2020 6:00 pm

Post by mastina »

Whoops, forgot to delete some of Menalque's post. Here's my post, fixed:
In post 4127, mastina wrote:
In post 4123, Menalque wrote:However, the question is how long those things should last before good
play
comes back into things. Why was that slot caught in the first place? GL and icon doing an excellent job of scumhunting on D1, purely based on play and before PRs came into things.
I can answer most concerns with one simple point.

Games cannot be balanced off of what players actually end up playing the game--games can only be balanced in a void of theory and the theoretical outcomes of a game, wherein you have to account for how likely scenarios are purely by statistical numbers, pretty much. You can 'adjust' the numbers to account for things that by play are more likely to happen in reality than in theory/are less likely to happen in reality than theory. (Quintessential example; millers are in theory negative utility but in reality due to site policy of instantly claiming on D1 are net-neutral to possible positive utility in reality. This is what I mean by the sort of theory/pure statistical randomness vs. reality difference.)

But at the end of the day, regardless of whether you make the adjustments or not, you're still doing much the same. You're guessing at what things could happen in theory, how likely they are to happen in reality, but these things can and will differ from what actually ends up happening because when designing the game, you cannot know what players will be in the game and even if you did, you wouldn't know which players would hold which roles (which affects the balance of the setup significantly; different players holding different roles makes huge differences in how the game plays out--easily demonstrable by a 'what if' of you being the strongest town role and bj being the strongman and the Chemist slot being the relatively disposable encryptor).

And by every metric, you cannot balance for the extremes of setups. In the extreme end of a cop + tracker + vtx8, vs. scum rolecop scum roleblocker, it is possible that the town lynch the scum roleblocker N1, the cop guilties the scum rolecop N1, and the tracker guilties the scum goon N1--in that scenario, the town has a perfect win because they killed the strongest scum PR on D1, and both of their investigatives got guilties unimpeded.

But would you call, "Cop, Tracker, VT x 8, versus Mafia Roleblocker, Mafia Rolecop, Mafia Goon", a townsided setup?
Fuck no you wouldn't, that setup is scumsided as fuck because all of the town's strength is on two roles who have a high chance of not getting guilties, and the scum have two role-investigatives per night (kill+rolecop), and the scum have two role-shutdowns per night (kill+roleblocker), which gives them the ability to fish for three town players per night, more town players than there are actual town PRs in the game.

If you math it out, with a VT lynch D1, and three groupscum, scum between their kill, roleblock, and rolecop can target 3/9 town players...on N1. The scum can literally identify/shut down/control/etc. one third of the total town's members, on the first night, and each subsequent night only increases those odds. If a VT mislynch D2 and no PRs hit, then on N2 the scum have a 3/7 chance of hitting a PR, on top of the 3/9 chance from the night before giving them a much much much better idea of where to focus.

So I think you get where that's going.

You cannot account for the best possible play in setup design, for the radical extremes that are incredibly unlikely to happen, because those extremes are...well. Extreme. Statistically unlikely, nigh-on statistically impossible. The odds of the scenario I mentioned before, where the town lynches the roleblocker D1 and the cop guilties the rolecop and the tracker guilties the goon?
1/13 (roleblocker lynch) * 1/11 (cop guiltying the rolecop) * 1/11 (tracker tracking the goon) * 1/2 (goon doing the nightkill) * 4/5 (chance scum don't kill one of the PRs). You know what that comes out to be? 0.0769230769230769 * 0.0909090909090909 * 0.0909090909090909 * .5 * .8 = 0.0254291163382072% chance probability.

The chance of the town pulling that perfect scenario off is less than one tenth of one percent. But by play, by the players coming into the game--it could happen. Something with a percentage so small as to statistically be almost 0%, could happen and make that scumsided-as-fuck setup be an absolute town sweep, purely because of the playerlist, which players got what roles, and how the game plays out.

Do you get what I mean, here?

You cannot anticipate what the storm of players in the game, with the roles they have, will end up creating for balance. And extremes that are statistically close to 0% could end up happening, purely because of that.

But we can't balance for those near-zero-percent outcomes. It's literally impossible to do. You have to balance for the likely outcomes. You have to balance for what is probable to happen, or ideally, extending it to the
plausible
outcomes--yet it is humanly impossible to account for every
possible
outcome because there are literally thousands, maybe even millions, of possible outcomes for any given game. Unique combinations of lynches, role interactions, etc. span into amounts that are borderline-incomprehensible.

In the case of this game to bring it back to your complaint--yes town did well to identify the Chemist slot as scum even before PRs are taken into account, but I ask again; how would it be possible to anticipate that the town would scumhunt well before PRs are taken into account, without knowing which players are which alignment? How can you know that the game will have people like Slaxx and Iconeum as part of the town popping off on D1 and people like Chemist as scum that are easy to catch?

You literally can't.

If instead, the game had some rather mediocre town players in the game and, sayyyy, a don corelone candidate/winner as the slot Chemist was--then guess what? Then by play, the scum stomp on D1 and by play, scum gain a massive edge.

I suppose I can give you a final tl;dr on this.

The NRG fundamentally
cannot
balance around play, because play is determined by players and the alignments of said players...
...And that is a quality determined after the setup has long-since been finalized.
Because we CAN'T balance around play, we balance around 'best guesses' as to the probable scenarios likely to happen in any given game. Those best guesses tend to not factor in the absolute extremes that are statistical abnormalities, so scenarios such as the strongest scum PR being lynched on D1 (something statistically very unlikely to happen), aren't something we can balance around.

This is particularly true in games that are swingy as fuck. It is possible to design games that have minimal swing; these games have less extremes and are more moderate. But therein enters moderator discretion--moderators are allowed to run games that are incredibly swingy, so long as they are balanced, should they choose to. Alisae/the worst were informed this game would be swingy-as-fuck, but they were okay with that; they had the option to design a setup with less swing, but chose not to because this was the setup they wanted to run.

The NRG do their best to give the moderators the setup that the moderators want to run, while maintaining that it is balanced. We can't control every variable, especially not when the moderator wants to run a swingy setup, but moderators have that right.
(Granted, I wouldn't be opposed to an optional opt-in description for game setups to be announced in advance, where mods announce expected swing value of their setup, on a scale of "moderately low, moderately high, high, extremely high", but if you think that's a good idea, you should probably pitch it to implosion and such.)
Would like to mention, among the adjustments you can make is more or less mentally calculating out the
expected
competency level of town versus scum, but that is a dangerous metric to try and delve into because if you guess at the expected competency level of the town/scum and get it wrong, the game is shifted much much much further into a direction it shouldn't have gone.

By which, I mean.
If you design a game expecting there to be an incredibly competent town and not very competent scum, then if you get in your game an incredibly incompetent town and competent scum, the scum roflstomp the town; if you design a game expecting there to be an incredibly incompetent town and competent scum and end up getting a game with an incredibly competent town and not so competent scum, then the town will roflstomp the scum.

So that's why it's dangerous territory to delve into that design idea. You're more or less guessing, and if you guess wrong, the game becomes a farce. It's usually better to stay objective, to not take huge guesses, to stick to the numbers, by and large.
User avatar
the worst
the worst
Snuggly Duckling
User avatar
User avatar
the worst
Snuggly Duckling
Snuggly Duckling
Posts: 36314
Joined: November 7, 2015
Location: pond

Post Post #4129 (ISO) » Tue Mar 31, 2020 8:36 pm

Post by the worst »

Basically second everything mastina said. that's a much better variation of what I wanted to say. :p
who's scum? i haven't read up yet but like, it's me
--
intermittent v/la until late march
User avatar
bji
bji
Mafia Scum
User avatar
User avatar
bji
Mafia Scum
Mafia Scum
Posts: 1597
Joined: December 23, 2014

Post Post #4130 (ISO) » Wed Apr 01, 2020 8:34 am

Post by bji »

In post 4111, mastina wrote:if two goons decide to let their only scum PR die D1
You speak as if we had the entire setup spec to refer to like you do when making your statements.

How could we have known that "doctor activated ascetic" was so important? We were supposed to infer the entirety of our strategy from the existence of this role? And then play it to perfection?

Really I thought that the doctor activated ascetic was there to balance a town vig. I thought it was going to be a game of deciding when to protect a mafia from being shot by the vig versus when to protect itself from being investigated.
User avatar
bji
bji
Mafia Scum
User avatar
User avatar
bji
Mafia Scum
Mafia Scum
Posts: 1597
Joined: December 23, 2014

Post Post #4131 (ISO) » Wed Apr 01, 2020 8:39 am

Post by bji »

In post 4117, Alisae wrote:People should really consider player's roles when bussing
Why? We won. We won using the strategy that always wins in every scum game I have ever played. We acted well. That's all we needed.
User avatar
bji
bji
Mafia Scum
User avatar
User avatar
bji
Mafia Scum
Mafia Scum
Posts: 1597
Joined: December 23, 2014

Post Post #4132 (ISO) » Wed Apr 01, 2020 8:45 am

Post by bji »

Also I still think the game was townsided because I think all of the arguments made about what "should have happened" completely ignore the element of randomness that human interactions bring to this game. The "should have" arguments do seem founded on this notion that players should be able to deduce specifics about setup from limited information where other deductions are equally (or nearly equally) as valid and lead ot vastly different outcomes.

Also I still enjoyed the game and am not complaining.

Also I still think town probably should have won the game.
User avatar
bji
bji
Mafia Scum
User avatar
User avatar
bji
Mafia Scum
Mafia Scum
Posts: 1597
Joined: December 23, 2014

Post Post #4133 (ISO) » Wed Apr 01, 2020 8:49 am

Post by bji »

One thing I definitely learned in this game is to pay very close attention to scum roles and try to deduce what the game designer is trying to get you to do. This is the first game where I've played where I felt like it was set up for the players to play a certain pretty specific way and I will be on the lookout for this in the future.
User avatar
Datisi
Datisi
he/him, it/its
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
User avatar
User avatar
Datisi
he/him, it/its
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
Posts: 26007
Joined: March 28, 2019
Pronoun: he/him, it/its
Location: Croatia
Happy Scumday!

Post Post #4134 (ISO) » Wed Apr 01, 2020 8:51 am

Post by Datisi »

gib scum PT
I will straight up disregard all reason if you have a PR dream again. You can come back and be like, “I dreamt that Locke is a N2 Bulletproof Multitasking Cop and Self-Targeting Doctor,” and I will go, “Okay, Locke kill it is then.”
~M
User avatar
Alisae
Alisae
lolbalance
User avatar
User avatar
Alisae
lolbalance
lolbalance
Posts: 47090
Joined: October 31, 2016
Location: Cali~ (PST)

Post Post #4135 (ISO) » Wed Apr 01, 2020 9:00 am

Post by Alisae »

In post 4131, bji wrote:
In post 4117, Alisae wrote:People should really consider player's roles when bussing
Why? We won. We won using the strategy that always wins in every scum game I have ever played. We acted well. That's all we needed.
im sayin in general
GTKAS
| here.
User avatar
mastina
mastina
She/Her
False Prophet
User avatar
User avatar
mastina
She/Her
False Prophet
False Prophet
Posts: 16670
Joined: October 7, 2016
Pronoun: She/Her
Location: Between Snohomish and Monroe, WA
Contact:

Post Post #4136 (ISO) » Wed Apr 01, 2020 10:15 am

Post by mastina »

In post 4130, bji wrote:
In post 4111, mastina wrote:if two goons decide to let their only scum PR die D1
You speak as if we had the entire setup spec to refer to like you do when making your statements.
Well to be blunt--yes, you did.

The scum are the informed minority--furthermore, in this game, in addition to knowing what roles you had, you had information to the nature of TWO of the town roles. With access to the knowledge of your OWN roles in COMBINATION with the information about half of the town's power roles (knowing there is a cop, knowing that there is a rolestopper), then...

...Absolutely, yes, you had the information needed to deduce that the ascetic doctor was your strongest power role. Knowing that there's a doctor in the game (via it being a mafia role) and knowing that there is a cop in the game, it follows that the game is meant to have a follow-the-cop aesthetic to it, subverted by the expected protective (doctor) being scum instead of town. If you deduce that the game has a subverted follow-the-cop mechanic, then you can also infer that there's probably a town roleblocker in the game, too, a role which would normally be scum but due to your roles you know isn't scum.

By knowing that the town has a cop and a rolestopper, and being able to infer a roleblocker, you can infer that the ascetic doctor is meant to be used to either deny the town information or to look town as a safeclaim of doctor or maybe do both at the same time if claiming doctor while using the ascetic.

Basically, by knowing there's a rolestopper in the game, and with a massive hint that there's a roleblocker in the game via knowing there's both a cop and doc in the game, you can infer that the setup has multiple methods of denying successful night actions, which the activated ascetic fits into.

So, yes. You were in fact meant to infer the importance of the doctor activated ascetic, from the existence of your role.
That's the entire reason scum's weapons are being the INFORMED minority and the nightkill. If scum don't use
the information available to them from the beginning of the game
, if scum don't think about what information is at their disposal, if scum don't think about what roles they have and their likely importance to the game...then they are literally throwing away one of their two strongest tools to win the game.

You can of course, attempt to use your information and be incorrect in the conclusion--but that's, again, something which comes down to PLAY, something we can't know when designing the setup. I've laid out the process above for how players would be able to, with purely the information available to the scum, be able to deduce a significant portion of the setup. Roleblocker hinted at, town ascetic maybe hinted at, rolestopper and cop explicitly known, literally the only role scum had no way of speculating about from on D1 is the follower/tracker.
In post 4130, bji wrote:Really I thought that the doctor activated ascetic was there to balance a town vig. I thought it was going to be a game of deciding when to protect a mafia from being shot by the vig versus when to protect itself from being investigated.
Well in that case, you still should've thought it was your most important power role. If this game did have a vig in it, then letting the mafia doctor die D1 means that the town vig would have free reign to shoot whoever they pleased with no way of the scum stopping them--that'd have still been a very very very bad thing for the scum to let happen. Plus! Games with mafia doctors frequently feature gunsmiths. (And, yes, there can be games where there's both a town cop and a town gunsmith.) If this game had a town gunsmith in it, then the mafia doctor would've been a godfather to it.

To put this in a different way:
If you had followed the process I laid out above to get the information, sure, you would've known that the doc ascetic was your strongest role, but
even with incorrect faulty setup speculation
, you STILL should have known.

If you incorrectly setup specced the existence of a vig, then your doctor's your most important PR.
If you incorrectly setup specced the existence of a gunsmith, then your doctor's your most important PR.
If you correctly setup specced the nature of the 'subvert follow the cop', then your doctor's your most important PR.

I literally don't see how there's any setup spec possible where you'd conclude the doctor
isn't
the most important role.
Ascetic is one of the strongest scum roles in existence, even as an activated action. Doctor is one of the best safeclaims for a scum faction to have. Having both on the same slot gives it versatility and vast utility to be able to be flexible and give the scumteam options.
User avatar
mastina
mastina
She/Her
False Prophet
User avatar
User avatar
mastina
She/Her
False Prophet
False Prophet
Posts: 16670
Joined: October 7, 2016
Pronoun: She/Her
Location: Between Snohomish and Monroe, WA
Contact:

Post Post #4137 (ISO) » Wed Apr 01, 2020 11:02 am

Post by mastina »

In post 4132, bji wrote:Also I still think the game was townsided because I think all of the arguments made about what "should have happened" completely ignore the element of randomness that human interactions bring to this game.
Again--it is impossible. It is literally impossible. To balance around the randomness of human interactions, because those interactions are
random
. You fundamentally cannot know what will happen. Because you fundamentally cannot know what will happen, you can only balance around what is
likely
to happen.

And what is likely to happen? What 'should' happen, as you so put it.

Basically, the only way to balance while taking into account randomness, is to make educated guess via statistical likelihood, statistical probability, combined with some level of human element of knowing the difference between what is likely to happen in theory versus reality (e.g. millers are always going to claim D1).

And by the numbers, scum being the informed minority with a nightkill means that by the numbers, scum are more likely to have a disproportionately accurate read on the gamestate compared to the town. By which, I mean--even ignoring any human element in statistical calculations. If you don't assume scum will, with their information and nightkill, be able to nail the town down. If you just go by pure raw numbers calculations.

Scum's information and them being the minority still leads to them making better actions in a game.

By which, I mean. On D1 there is a 10/13 chance town mislynches, by the numbers.
In this game, there's a 5/10 chance that the town mislynches a power role, by the numbers. (In reality, that's probably only a 5/10 chance of a PR claim, but this is the by the numbers approach.) Assume it's a VT as it likely is, for the true demonstration.
Going into the night, the scum make up 3/12 slots--so there are only 9 unknown slots for them to sort through.
They sort one of these slots with the nightkill--leaving 8 slots left to sort through.
With only 8 slots to sort through, and knowing there's a cop and a rolestopper in the game, the scum have a much better idea of what action to use for the doctor/ascetic.

In contrast, the town power roles have no information at their disposal. The cop knows there is a doctor in the game and doesn't want to 'waste' an investigation on the doctor, but has 11 slots they can target. 3/11 give the cop a guilty, and if the cop has a good read on who the doctor is, that's only 2/10 slots the cop would get a guilty on.

The rolestopper doesn't know who to protect, and has 11 possible targets--with only a 1/11 chance of stopping the nightkill. And there is a 3/11 chance that the rolestopper's action interferes with the action of a town player.
The roleblocker doesn't know who to roleblock, with only a 1/11 chance of stopping the nightkill, possible 1/11 chance of preventing a false innocent, and a 3/11 chance of blocking town.
The tracker/follower doesn't know who to track, with only a 1/11 chance of getting a guilty, and via the doc, a 1/11 chance of a false innocent, or alternatively, a 3/11 chance of their action failing (due to town ascetic, activated ascetic, rolestopper target), not to mention being roleblocked.

By the numbers, the town's roles have a much much much higher chance of doing nothing good, compared to the scum. Scum, just by knowing who is scum, have better target selection, and by also having information from their roles not to mention any claimed town roles (because the more town roles there are in a game, the higher the likelihood of at least one being outed on D1, D2, every day really), have even BETTER target selection options.

In post 4132, bji wrote: The "should have" arguments do seem founded on this notion that players should be able to deduce specifics about setup from limited information where other deductions are equally (or nearly equally) as valid and lead ot vastly different outcomes.
Like I said--I don't see any world where you can think the ascetic doctor isn't the strongest scum PR. What possible setup, when you know there's a cop and rolestopper, leads to it not being important?

Players aren't expected to perfectly deduce the setup as scum. Quite the opposite, scum being able to perfectly deduce the setup is poor setup design. But scum
should
, and are expected to, be able to
partially
deduce the nature of the setup, based on the knowledge provided to them. Scum have the partial solve for the setup from the get-go, just by having knowledge of their own roles.

There are equally valid alternative deductions to be made, sure--but I don't see how any of those alternative deductions lead to the conclusion of "the ascetic doctor can be thrown away"? Deducing wrongly about there being a vig means that the scum are meant to use the doctor to stop the vig. Deducing wrongly about there being a gunsmith means that the scum doc is meant to be a godfather.

Again, ascetic is, generically speaking, one of the strongest scum roles in existence, up there with roleblocker and ungated strongman, comparable in strength to both of those. (And if it were a passive rather than active, to a ninja. It wasn't passive so isn't comparable to a ninja this game, butstill.)

Scum are expected to know the strength of their roles on a generic level, even without setup spec. If you think activated ascetic isn't a strong scum role, then...well. Your opinion of the role is severely underestimating its strength.

This might be more easily explained by using the example of a Mafia Godfather.
When a reviewer is reviewing a game with a Mafia Godfather, the reviewer is going to naturally assume that the mafia players understand that Mafia Godfather is an insanely strong scum role. They are going to assume that the mafia understand what the role does, how it functions, and its interactions with other roles.

To put it another way--a reviewer is going to assume that the scum have, more or less, read the wiki about their roles, familiarized themselves with the role, and from this familiarity, have a good idea of how strong that role in general is, even without knowing what the setup is. This doesn't delve into setup spec at all. This doesn't delve into scum's abilities to guess the nature of the town's roles to more accurately reflect the strength of the role. But in general, reviewers are going to assume scum know how strong, or weak, their role is.

If scum have an ungated strongman, it is expected by reviewers that scum will consider that a strong role, regardless of what the setup is. It is expected by reviewers that the scum will, without delving into setup speculation, know that that role is incredibly strong.
Similarly so for the inverse. If scum have a 1-shot rolecop, it is expected by reviewers that scum will consider that a weak role, regardless of what the setup is. It is expected by reviewers that the scum will, without delving into setup speculation, know that that role is incredibly weak.

And then it is generally assumed that,
armed with the above information
, knowing that an ungated strongman is incredibly strong, knowing that a 1x rolecop is incredibly weak, that the scum would delve into setup spec as to what that means for the nature of the roles in the game.

But if things go wrong in the first step--if the scum assume that an ungated strongman is an incredibly weak role, then...things kinda fall apart.
If the scum assume that a 1x rolecop is a god-tiered scum role, then...things kinda fall apart.
If the scum start from this wrong base when delving into setup spec, then their conclusions probably stray much much further from reality, too.

Basically, let me put it this way.
A reviewer does anticipate it is possible scum will see what role(s) they have and, from the knowledge of these roles, setup spec incorrectly as to what their strength
in that particular game
is. (If scum have a full strongman and incorrectly setup spec that it's there to stop a doc protect when it's really there to be a guilty to a town rolecop, then that's an error that is anticipated.)

A reviewer does not anticipate it is possible scum will see what role(s) they have and from the getgo, without delving into setup spec, incorrectly evaluate their generic strength level
ignoring that particular game
is. (If scum have a full strongman and incorrectly think it's a trash role, then that's an error that isn't anticipated.)

So in this game--with an ascetic doctor, reviewers anticipate that scum will see what role it is, and from that knowledge, setup spec.
If they setup spec incorrectly, then they can misevaluate what the strength of their roles in that particular game are.

But I get the feeling that you don't actually understand how strong that role in general, outside of this setup specifically, is. I get the impression that you don't know how strong a dual scum role with activatable asceticism is, and a reviewer won't anticipate that scum will see what the role is and offhandedly dismiss it as trash without thinking about what the role would be in the setup. To write the role off before doing the setup spec, more or less, and it really feels like that's what you're doing here, thinking that the role wasn't strong in general, that the role might've been strong in this setup specifically but in general wouldn't be, but in general it IS a strong role and in this game specifically was EVEN STRONGER than in general.
User avatar
mastina
mastina
She/Her
False Prophet
User avatar
User avatar
mastina
She/Her
False Prophet
False Prophet
Posts: 16670
Joined: October 7, 2016
Pronoun: She/Her
Location: Between Snohomish and Monroe, WA
Contact:

Post Post #4138 (ISO) » Wed Apr 01, 2020 11:15 am

Post by mastina »

I feel like I can do a better job of explaining that divide.

In a game with an ungated strongman, there is the
generic
expectation, mafia strongman is a strong scum role.
If the town has no protectives or blocking actions, then in that game with an ungated strongman, specific to that game, it is not a strong role.
If the town has nothing but protectives and blocking actions, then in that game with an ungated strongman, specific to that game, it is the
strongest
role.

There are two separate metrics in question--how strong the role generically is, and how strong the role is specifically in the given setup, which can be stronger than normal, weaker than normal, or about the expected level.

Reviewers assume that scum understand the generic strength of the role in question--that if they have an ungated strongman, they know it is a strong scum role.
Reviewers know that scum will not know how strong the role is specifically in the given setup--and expect them to speculate about this, knowing its strength and then to speculate if it is stronger or weaker than it appears on the surface.

In this game, there was an activated ascetic doctor--there was the generic expectation that that'd be seen as a strong scum role.
It was expected that scum would speculate as to how strong the role is specifically in the given setup--
It was not expected that scum would think that it wasn't a strong scum role in general and that it would be seen as equally weak in speculations about the given setup. (Because, in Normals, there's basically a holy quartet of the four strongest scum roles: Strongman, Roleblocker, Ninja, Ascetic. Those four roles are considered to be the best scum roles in general to have in any given Normal game.)

Reviewers don't expect scum to go "oh this is a trash role we need to bus it" to the scum's strongest role--not because it's a strong role in the given setup, but because it's a strong role in general.
User avatar
mastina
mastina
She/Her
False Prophet
User avatar
User avatar
mastina
She/Her
False Prophet
False Prophet
Posts: 16670
Joined: October 7, 2016
Pronoun: She/Her
Location: Between Snohomish and Monroe, WA
Contact:

Post Post #4139 (ISO) » Wed Apr 01, 2020 11:25 am

Post by mastina »

(To wit, the wiki page for ascetic literally tells you "this is a really really strong scum role", emphasizing how it is much much much stronger in the hands of scum than in the hands of town. It's also noted on the page for doctors that the strength of a mafia doctor is comparable to the strength of a town doctor--that is to say, not a weak role; a fairly strong role.)
User avatar
bji
bji
Mafia Scum
User avatar
User avatar
bji
Mafia Scum
Mafia Scum
Posts: 1597
Joined: December 23, 2014

Post Post #4140 (ISO) » Wed Apr 01, 2020 12:54 pm

Post by bji »

In post 4137, mastina wrote:Like I said--I don't see any world where you can think the ascetic doctor isn't the strongest scum PR. What possible setup, when you know there's a cop and rolestopper, leads to it not being important?
Well like I said, I thought it was probably just there to protect against a vig. Objectively wrong conclusion? Perhaps.

But with that conclusion, it didn't seem on balance worth dirtying the entire team to save the slot whose (in my conclusion) purpose was just to protect that same scum team who would be so badly dirtied in the process.

Learning experience though. Weird roles require a lot of thought and planning because they're there for a reason and it may not be the most obvious reason.
User avatar
bji
bji
Mafia Scum
User avatar
User avatar
bji
Mafia Scum
Mafia Scum
Posts: 1597
Joined: December 23, 2014

Post Post #4141 (ISO) » Wed Apr 01, 2020 12:58 pm

Post by bji »

In post 4134, Datisi wrote:gib scum PT
I wrote a PM to the worst reminding him to open the scum PT. I am sure he'll get around to it ...

The scum PT is long though. Hope you enjoy reading.

Also where is this private Datisi notes thread you were mentioning earlier?
User avatar
Datisi
Datisi
he/him, it/its
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
User avatar
User avatar
Datisi
he/him, it/its
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
Posts: 26007
Joined: March 28, 2019
Pronoun: he/him, it/its
Location: Croatia
Happy Scumday!

Post Post #4142 (ISO) » Wed Apr 01, 2020 1:01 pm

Post by Datisi »

I am also waiting on the worst to release it. I do not have the power to do that.

it's gonna be a great read though. page 16 is my favorite. :]
I will straight up disregard all reason if you have a PR dream again. You can come back and be like, “I dreamt that Locke is a N2 Bulletproof Multitasking Cop and Self-Targeting Doctor,” and I will go, “Okay, Locke kill it is then.”
~M
User avatar
bji
bji
Mafia Scum
User avatar
User avatar
bji
Mafia Scum
Mafia Scum
Posts: 1597
Joined: December 23, 2014

Post Post #4143 (ISO) » Fri Apr 03, 2020 5:00 am

Post by bji »

Still waiting for these mod actions ...
User avatar
Alisae
Alisae
lolbalance
User avatar
User avatar
Alisae
lolbalance
lolbalance
Posts: 47090
Joined: October 31, 2016
Location: Cali~ (PST)

Post Post #4144 (ISO) » Sun Apr 05, 2020 12:29 pm

Post by Alisae »

GTKAS
| here.
User avatar
Looker
Looker
the
Stenographer
User avatar
User avatar
Looker
the
Stenographer
Stenographer
Posts: 5304
Joined: February 20, 2009
Pronoun: the

Post Post #4145 (ISO) » Sun Apr 05, 2020 12:48 pm

Post by Looker »

So flattering :lol:
User avatar
Datisi
Datisi
he/him, it/its
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
User avatar
User avatar
Datisi
he/him, it/its
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
Posts: 26007
Joined: March 28, 2019
Pronoun: he/him, it/its
Location: Croatia
Happy Scumday!

Post Post #4146 (ISO) » Sun Apr 05, 2020 1:14 pm

Post by Datisi »

i found yall's crumb hunting funny, but then i read the description of my death and got chills. :lol: will read in detail once it's not 2am.
I will straight up disregard all reason if you have a PR dream again. You can come back and be like, “I dreamt that Locke is a N2 Bulletproof Multitasking Cop and Self-Targeting Doctor,” and I will go, “Okay, Locke kill it is then.”
~M
User avatar
Datisi
Datisi
he/him, it/its
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
User avatar
User avatar
Datisi
he/him, it/its
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
Drawn from Memory

Drawn from Memory<br><img src=https://i.imgur.com/rUdtw5m.gif>
Posts: 26007
Joined: March 28, 2019
Pronoun: he/him, it/its
Location: Croatia
Happy Scumday!

Post Post #4147 (ISO) » Sun Apr 05, 2020 1:18 pm

Post by Datisi »

oh, i just realized my notes haven't been released yet. huh.
I will straight up disregard all reason if you have a PR dream again. You can come back and be like, “I dreamt that Locke is a N2 Bulletproof Multitasking Cop and Self-Targeting Doctor,” and I will go, “Okay, Locke kill it is then.”
~M
User avatar
implosion
implosion
he/him
Polymath
User avatar
User avatar
implosion
he/him
Polymath
Polymath
Posts: 14329
Joined: September 9, 2010
Pronoun: he/him
Location: zoraster's wine cellar

Post Post #4148 (ISO) » Wed Apr 08, 2020 11:44 am

Post by implosion »

If you want to nominate this game or any people in it (moderator or players) for a scummy, please do!
Locked

Return to “Completed Mini Normal Games”