Mini Normal 2119 [game over]

[mastina]

Hi I reviewed this game.

[mastina]

I forgot to mention the f3 lylo rule
It’s on mastina’s wiki iirc
The townier of the two players is probably scum if scum played well this game

I mean, I do preach that but I wouldn't be able to tell you where I gave that advice.

I would be able to tell you different advice that'd have condemned Menalque though.
Occam's razor; which is simpler? That there is an unknown ungated mafia ninja which there is no evidence of who happens to be the last scum alive, or that the innocent result is an actual innocent?

The former requires you to invent a role that there's no evidence of existing and which turns the game hilariously scumsided (because, Iconeum investigated Micc multiple times while there was only one scum alive and got the same result, meaning that it would NEED to be a FULL ninja), whereas the latter is by far the simpler, more likely explanation which is stupidly simple.

[mastina]

Also, keep in mind. Ninja is a role that tends to not exist to counter followers/trackers; it is usually in there to fight a watcher. You can have a ninja to counter a follower/tracker, but it is usually incredibly gated and against a town that has significant strength behind it.

A game like this having a mafia joat with a ninja shot, a game like this having a 1x ninja, maybe maybe a 2x ninja, would be plausible, but a game like this having even a nonconsecutive ninja would be hilariously scumsided. And Micc was via the tracker/follower, ONLY possible to be an ungated Ninja. (Not x-shot, because of the late nights Micc was investigated and the number of times. Not even-night, odd-night, or nonconsecutive, because Micc was tracked two days in a row. Micc would have HAD to be a FULLY FULL, ungated in any way, Ninja, for him to be scum, and a fully full Ninja against a tracker/follower is a role that is incredibly unlikely to make it through a Normal review unless the town has some SIGNIFICANT additional power.)

[mastina]

(Basically. In Normal Games. 97.5% of innocent results, are actually going to be innocent results. Unless there is VERY strong reason to reasonably suspect otherwise, e.g. a game with a town gunsmith and town vig you can reasonably anticipate a mafia doctor and even then MOST innocents are going to be actual innocents.)

[mastina]

Btw, re;being NRG member reviewing this game--I admit that there is fault from me, mostly due to my approach.
As a secondary reviewer, I am usually a lot less thorough than when I am the primary reviewer. This comes from the obvious; as the secondary reviewer, it's more or less my job to take a back seat to the primary reviewer, in that I am more or less, not supposed to hold back the review by being heavily contradictory towards the primary reviewer. (Small differences are fine, obviously, but at the end of the day the primary reviewer's stance should take priority over the secondary reviewer's stance.)

I did chip in and give my thoughts, but I left the majority of the work to PenguinPower, which was probably a contributing factor to the lack of thoroughness and the oversight. I might've done most of the initial work, but if you get to read the review PT, the first page was mostly me while PP was busy, and then the second page was pretty much almost all PP giving his insight and giving it the pass which I was wanting due to him being the primary and trusting his judgement, more or less.

For what it's worth--I do think the setup is ~50/50 on balance, albeit swingy as fuck. It's just, a little unfortunate that I missed a couple of details which I normally wouldn't miss, and that we didn't get to fully discuss the roleblocker/ascetic/action investigation interaction. (I used to discuss that every review, but nowadays I normally don't unless I'm worried the mod will make mistakes on it; I didn't think that between Alisae and the worst that it'd be likely for a mistake to happen. Which I admit, is sloppiness on my part. Basically, perfect storm of unfortunate mindsets; being secondary means I'm more laxed, and reviewing for tw/Alisae meant I was more laxed. Purely a mistake on my part, so I again apologize for it.)

[mastina]

I agree it’s swingy and I think part of the reason I thought it was so townsided was because of the wrong power dying on D1 but I don’t think that should be enough to doom your game in any circumstance and it very nearly did.

Well to be blunt--if the strongest role for a faction dies in the first phase, the other faction should be rewarded for eliminating the strongest threat to their faction. (Or maybe, if the strongest role for a faction dies in the first phase, that faction should be punished for allowing that loss to happen, but I realize that this definition is less optimal because sometimes it's out of your hands. Scum three-man scumblocking, pushing hard, trying to throw everything and the kitchen sink, and yet somehow the town gets a D1 scumlynch on the strongest PR anyway? Not much they can do about that. Scum nightkill the strongest town PR N1? Not much the town can do about that. Which is why I favor the 'reward opposite faction' definition over the 'punish the faction' definition.)

If the town eliminates the strongest scum PR on D1, the town should be rewarded for having done so. Most mini normals only have an average of one scum power role; if two goons decide to let their only scum PR die D1, then no duh they are going to have a much harder, rougher time? Because the game wasn't balanced around that abnormality happening. If something has a 1/13 chance at happening, then 12/13 times, it doesn't happen, which means in 12/13 situations, the town isn't rewarded and the scum aren't punished.

So if a statistical fluke, a statistical abnormality, happens, it is going to skew the balance of the rest of the game.

[mastina]

Activated ascetic doctor — doctor bit does nothing, activated ascetic makes more susp than not

Don't forget that the doctor had information about the cop and the cop had information about the doctor--check out Mystery Box of Silver 4 (I think it was 4 anyway) for how that dynamic plays out. The mafia role instantly pegged the town investigative and N1 nightkilled them, with the town being completely and entirely blindsided 100% unaware of this dynamic with no possible way of having caught it. The town had no way of having figured it out, but the scum did.

The weapons of the scum faction are:
Information and their Nightkill.
The Information of the doctor, and with the ability to easily peg the cop with that information, with the cop thinking that the doctor is town, are strongly advantageous to the doctor.

Additionally, so long as there isn't only one scum alive (and to reiterate--if there's only one scum left alive, the town SHOULD be rewarded for this!), the mafia doctor should never be making the nightkill. Thus, the mafia doctor should be using either their ascetic or their doctor action every single night.

If the doctor uses their ascetic, they are investigation-immune to both the cop and the tracker, and the town doesn't have a way to figure out the source of the failure. (Keep in mind, with both a roleblocker and a rolestopper, there's any number of possible failure causes.)
If the doctor uses their doctor action, they are creating a false innocent for the tracker/follower. This means that if the tracker/follower track/follows the mafia doctor, they will falsely clear the mafia doctor--this is, effectively, a GODFATHER CLEAR.

The Mafia Doctor had a way to make themselves town to BOTH town investigatives--you're dismissing the power of what was essentially a full godfather in a Normal game. It could become town to the cop simply by the cop knowing there's a doctor in the game and the doctor knowing there's a cop in the game so communicating this to each other results in them mutually self-clearing each other; it could become town to the tracker/follower simply by using the protection action on any player the mafia weren't nightkilling that was a plausible protection coming from them.

The strongest role wasn't your strongman; it was your doctor.

[mastina]

Scum ascetic is one of the strongest roles scum can have.
Scum are already the informed minority--giving them additional information increases the odds they will win.
The play on information is likely to falsely conftown a scumslot or falsely damn a town slot, either of which vastly increases the odds the scum will win.
Strongman is one of the strongest scum roles possible, especially in a game where the majority of the town's strength comes from a role the strongman punctures (the rolestopper).
Town roles have incredible anti-synergy. Roleblocker and rolestopper very easily can interfere with each other, roleblocker can block the cop, rolestopper can rolestop the cop's target, and these outcomes make the cop less able to generate reliable results--in a setup where one scum already has immunity to the cop (ascetic).

Literally everything in this setup stacks things in the scum's favor and the town has literally nothing going for it.
Plus, MBoS 4 should be a fair indication of what happens when scum are informed of a town role that is informed of the existence of a role in the game but not what alignment said role was.

You know what happened that game?
The town cop role successfully 'crumbed to the scum role, and vice-versa...allowing the scum to make a N1 nightkill on the cop role without any consequence because nobody in the town had the information to pick up on the 'crumb and the information per Normal standards doesn't flip with them so they had no way of gaining said info.

^This is what I said in the review thread. This was about the originally-proposed setup, sure, but even in the changed-into-final version, my comments here are still applicable for where the scum's strength lies and where the weaknesses in the town's strength lied.

You always have to remember that for every perfect storm of how things could go well for a faction there's also a perfect storm of how things could go disastrously for a faction, and how likely each is to happen.

The chances of the town roleblocker/rolestopper fucking with the town's other power is fairly high. (And, sure enough, Datisi blocked town!)
The chances of the scum identifying the cop early and eliminating them are very high. (Ehhh, not going to say this happened in the way which it'd have been most likely to have.)
The chances of the scum's strongest role being lynched on D1 are very, very low. (And yet it happened.)

[mastina]

Chemist1422 (7)

: Iconeum, GuiltyLion, bji, bob3141, Menalque, Datisi, Looker

Looker (4)

: Chemist1422, wilky, emps, AaronFrost
not voting: Pine, Tchill13

(To illustrate what I mean here--both scum were on Chemist here. If both scum were instead on Looker, then Looker would've been at L-1 with Tchill being a player who probably would hammer him, and Chemist would've peaked at at-most 5 votes and wouldn't have even needed to claim. Or if he did need to claim, your team would've had more time to discuss what the claim should've been, how to claim, when to do it, etc. The double-bus on the scum's strongest PR directly led to the death of said strongest PR, which was easily 'avoidable'*. Avoid the double-bus of the scum's godfatheresque role, and the town has a much, much harder time getting a lynch on scum and a harder time throughout the game.)

*As I said, a town

can

get seven votes onto scum, even the scum's strongest power role! Can. In very, very rare cases. Statistically speaking, it has happened before but it is a statistical abnormality. First off, games where towns lynch scum on D1 are in the minority. Say the town lynches scum 30% of the time. Then you need to,

of that 30%

, analyze how many have no scum on the lynchwagon. Of the, in this hypothetical 30%, say only 30% of those games have no scum on the lynchwagon.

30% of 30% of games played, is 9% of games played, where town would have a D1 lynch on scum without the help of scum. I'm totally pulling these statistics out of thin air since I'm not going to go through 2000 games' worth of data to find how many games had a D1 scumlynch and how many of said D1 scumlynches were all town, but you get the idea.

And then, you could go one step further--there's a 1/3 chance that a D1 scumlynch is on the scumteam's strongest power role. (This is an exact statistic due to three scum in a game, obviously. So, this one's not pulled out of my ass and is for sure 33.33% without any shadow of a doubt.) 33% of the hypothetical 9%, is 2.97%.

If there were only a 2.97% chance of the strongest scum PR being lynched on D1 in an all-town wagon...

...Yeah, that's a statistical abnormality; that's something the town absolutely should be rewarded for.
So, like. When I say "avoidable", it's not 100% avoidable. But it'd be something like 97.03% avoidable. A 97.03% chance, you could avoid the lynch of your strongest scum PR on D1.

If there's a 97.03% chance the lynch of the strongest scum PR is avoidable on D1...then uhhhhhhh...what does it say when the strongest scum PR is lynched on D1?

[mastina]

While I'm not going to do 2k games' worth of research, I did research comparable-sized games that have not been archived:
2085*--no scum lynched D1 (12p, three groupscum)
2088*--no scum lynched D1 (12p, three groupscum)
2092--no scum lynched D1
2076--scum lynched D1; double-bus (hard to say if strongest)
2117--no scum lynched D1
2115--scum lynched D1; one scum bussed (strongest role)
2113--no scum lynched D1
2111--no scum lynched D1
2114--scum lynched D1; no scum bussed (strongest role*; scum had UB so losing their role D1 didn't hurt them)
2121--no scum lynched D1
2124--scum lynched D1; one scum bussed (negative-utility scum role; scum wanted it dead)
2122--no scum lynched D1

So, of 12 other completed mini normals beyond this one, four have had D1 scum lynches. In this small, 12-game, sample size, that's 25% of games having a D1 scum lynch.

Of those D1 scum lynches, only one had it be the strongest scum role, but with the caveat of scum having a Universal Backup and thus, not having lost anything. (And another one of those games had the scum lynch a tracker-enabler buddy, thus, disabling the tracker for that game.)

Of the 25% of D1 scum lynches, only one had no scum involved. 25% of 25%. In our sample size of 12 games, only 6.25% of them had a D1 scum lynch that scum could NOT avoid. Only 6.25% of them had a D1 scum lynch that scum had no contributions towards.
It happens that of that 6.25%, 100% of it just so happens to have been the scum's strongest scum PR, albeit with that caveat that when the scum have a Universal Backup, that's the one time losing their strongest scum PR immediately isn't inherently necessarily detrimental. (And could be, maybe, potentially, theoretically, possibly advantageous if the town thinks that due to the scum PR being dead that with it out of the game scum couldn't be using its powers anymore.)

Like I said, very small sample size, too lazy to dig up statistic beyond this, but this should give you a basic idea of how much of a statistical abnormality it is for the town to lynch the scum's strongest PR without it basically being hand-fed to them.

Town's are, by and large, not as smart as scum think they are. Scum have this weird tendency to bus scumbuddies on D1 far more often than what would actually be necessary since the town on D1 has a much much much harder time than people think of lynching scum.

[mastina]

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.

[mastina]

Whoops, forgot to delete some of Menalque's post. Here's my post, fixed:

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.

[mastina]

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.

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.

[mastina]

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.

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.

[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.

[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.)