mini normal 2226; who won

[mastina]

Hi I revi--errr, whoops, force of habit. *cough*
*ahem*

Hi I designed this game.

And I owe the players, especially the scum, an apology;

(Btw in hindsight I am having regrets in having second-guessed the setup balance into thinking that in hindsight, it's probably a fair bit more townsided than I had previously assumed. Mostly because in hindsight, the 2x rolecop is even stronger than I gave it credit for. Getting a guilty on the doctor/traitor is strong and the neighbor is an implied guilty, but it is also able to generate what amounts to hard-innocents because once the town gunsmith claims their information, it becomes obvious that a Vanilla result is on town so the 2x rolecop is functionally a 2-shot cop in strength. That, plus the fact that traitor gets endgamed if being the last scum alive. Ah well, it's too late to make a change now so all I can really do is offer the players an apology in postgame.)

I am informed that only one mafia has a gun

*flashbacks to that one normal i modded with 0 scum guns and a town gunsmith*

(For the record, yes, that setup did inspire this one. I obviously didn't want the setup to be guessable or easily solvable but I wanted to take inspiration from that game and do my own spin on it. In hindsight, per my 33, I fucked up, I

now

think the game is more townsided than it by all rights should be, because I underestimated the strength of the town PRs and overestimated the strength the scum would have in shutting down and outplaying the town PRs. I will fully own up to the fuckup when the game ends. But yeah, that was where I got the loose idea for this one.)

Not seeing how this passes a review.

Well scum aren't going to get endgamed N1 (tho D2 is theoretically possible) so not quite as dire as you imply, but reviewers are human and we all make mistakes; this setup as-is just so happens to have been one.

Oh was she a reviewer? Ngl that makes me more inclined to just forget about what’s balanced or not since she always explains her thought process from the review postgame

The core concept behind this game:
I wanted an informed gunsmith that knew there was only one mafia gun in the game.
To achieve that, I gave the mafia a doctor and a traitor.
I wanted the mafia to have a three-man unit so I gave the traitor a neighborhood with the otherwise-a-goon.
I didn't want the town to clear the neighborhood as being scum-town (as I have very very VERY strongly spoken against double-scum neighborhoods before), so I wanted to let the town know that there were exactly three players who could communicate in private.

Those were the core aspects of the setup.

Everything else was an attempt to balance out the concept behind it.

I wanted there to be very very few guilties the town could get on the scum, and for once the town got a guilty for them to be unable to generate more than just the one guilty pretty much. I wanted the town's power roles to mostly not be designed to catch scum, and yet for the town's power roles to also KNOW that they are not designed to catch most of the scum.

I wanted none of the town PRs to be clearly and unambiguously beyond any shadow of a doubt to be town, but for the town to still be able to bolster their strength by usage of said PRs, for the PRs to not be completely worthless or even negative utility.

I succeeded in all of those goals!
...Unfortunately, I made a genuine oversight, a fuckup on my end, a complete and total mistake from a lapse.
I knew that the town could potentially generate a few conftown from their power roles--in my opinion, this is not a bad thing; this is fine; this is still balanced given the setup. A few conftown when the town power roles aren't designed to get guilties on most of the scum is, in my opinion, fine.

The key word here being

few

conftown.

I fucked up in not realizing just how

many

conftown could be generated. It was genuinely a failure on my end to not see just how many conftown could be gotten and for that oversight I well and truly do apologize; the setup is a lot more townsided than I thought when I designed it because there's more conftown available readily and easily to the town than there should be in a healthily balanced game.

Wait. Are you starting your models now, after the normally almost useless role just happened to muddle out a guilty night 1 AND after the jailkeeper landed a save AND after the scum team apparently chose to no kill?

Because, yeah. All of those things before you start looking at the win chances, should push the win in towns favor.

While this is true, that this run of the game is about as good as it possibly gets for the town and as bad as it possibly gets for the scum, with all of the town PRs making some of the best decisions and the scum making bad calls, MathBlade in this case is, for once, actually right in thinking the setup is townsided beyond what it should be, because it IS townsided beyond what it should be, due to my fuckup in not realizing how many clears the town could get.

I don't think a few clears is bad design. But this setup allows for more than a 'few', it's actually too easy to get clears. The town have played as well as they could and generated more than they by all rights should, but they still can, just by random statistical chance, generate more conftown than healthy.

1-3 conftown in a game is, imo, not bad, but this setup allows for 3-6 in too short a span of time and me not realizing that is a drastically bad oversight which left the town with a lot more power than I gave them credit for, and again I feel the need to apologize to the players because as the designer and NRG member, I

should

catch that sort of hidden power as it were, and my failure to do so means MathBlade's criticisms are actually something I am fully onboard with in complete agreeance as they're beyond just fair, they're completely and totally 100% accurate.

Sorry for not noticing before the setup was finalized. <3

she will not take criticism xD

I do take criticism. It's just that players tend to not understand the math behind theoretical balancing and their criticism is usually mathematically wrong. They tend to think that scum had a lot less options than they did, that town had a lot more options than they did, that town power roles were a lot stronger than they actually were, that the scum were a lot weaker than they actually were, etc. I then point out the flaws in their assumptions and how by the math, they were in fact thinking of outcomes that are statistically speaking anomalies. And reviewers can't review for outcomes that are statistically speaking anomalies. We can only review what we believe the average overall most likely outcomes are and make our judgement calls there.

...However in the case of this setup: the players are perfectly right, their criticism

isn't

mathematically wrong; the scum did have plenty of options, but the town had more options than I thought they had; the town power roles WERE stronger than I gave them credit for, the scum were maybe slightly weaker than I thought they were, and the players' assumptions aren't flawed and are in fact dead on the money because by the math, the outcomes they are thinking of are NOT statistically speaking anomalies. The statistically speaking anomalies this game are the outcomes where the town PRs don't generate guilties or innocents, so as the designer of the game and NRG member I

should

have caught it and revised it, but I missed it, I made a mistake, I fucked up, and I didn't realize it until too late.

The only defense I have for this is that this is literally the first setup I've designed in over two years, and designing a setup is different in perspective from reviewing a setup, in that as the designer it is easier to not notice the flaws in the setup you designed (that's why you need a reviewer in the first place!) and that goes doubly so for someone rusty in designing a setup, but I still feel bad/guilty for the oversight and not noticing the mathematical odds behind the town generating an unacceptably high number of clears. But I still take full accountability for the mistake, it was still a fuckup on my end and I can only offer postgame apologies to the players for the oversight. I

should

have caught the hidden power and given further revisions to the town's power.

Making the roaming loyal jailkeeper an alien (so that there's an increased risk of messing with the town PR's investigative results), making the rolecop be even weaker e.g. a 2x vanilla cop, maybe gating it further to 1x vanilla cop, maybe gating the traffic analyst, maybe giving the scum something extra like a 1x roleblocker, etc. Probably not all of these at once, but done

something

to that effect, maybe multiple somethings, to lessen the odds of the town PRs getting too much bang for their buck.

I think the concept of the setup was fine and that the setup met 80% of its goals, but having too many town clears too easily is the remaining 20% that nerfs to the town and/or buffs to the scum could have accomplished, and my failure to realize that this was necessary is, explicitly, exactly that; a failure.

I don’t see how this passes a review

Through designer oversight.

If we work together in the future, do you want me to more strongly push back if I disagree with your assessment? I did think it was somewhat townsided even at the end, but I wanted to compromise and I figured it wasn't likely to be completely busted. But it seems pretty reasonable to ask implosion or try to get a third reviewer in if the two are at odds, instead of just forcing them to compromise.

Eh, a setup being 'somewhat townsided' isn't bad imo; if the setup was, in your opinion, ~55-45 town-scum (without the hidden town power in the game, this is about what I thought the setup was but see below for what I think the setup actually is with the hidden town power), that'd be an acceptable margin to me so if you thought it was "somewhat townsided, but within the acceptable margin as still close enough to balanced", then it's an acceptable setup to pass.

However, if the setup was closer to 60-40 or in your opinion was more than 60% town, then yes, absolutely, you should push back if you disagree with my assessment. Especially since there were multiple ways to preserve the core identity of the setup I wanted to design but with various possible adjustments to make the town be weaker and/or the scum be stronger.

After all, I already took some of your suggestions to weaken the town and get the game closer to balanced. I had concerns that if too much was removed from the town or too much was added to the scum the town would be too weak, but if you voiced your opinion that it was more likely that town was too strong, especially describing the why, I would have, once seeing it myself, agreed.

There wouldn't have been reviewers at odds in this case if I hadn't made the critical oversight in underestimating the town's power--and you pointing out your thoughts would've probably helped me actually see the hidden strength to the town's PRs that I didn't see before.

---

On the note of the big oversights in the hidden town power, the main things I didn't take into account:
1: The loyal jailkeeper's power. The roaming loyal jailkeeper was designed to be a protective role that could not be used offensively and could not enable follow the cop. It was meant to be a role that couldn't catch scum and couldn't indefinitely protect the town. However, I failed to account for a certain aspect of loyal, in that Loyal is, in of itself, an investigative--a Loyal protective role is a role that should, theoretically, not conftown players because if the protection is successful then balance of probability is that the player protected was town anyway (basically, if there's no kill, the protection part makes the player town in of itself so the loyal part doesn't come up in them being town), but Loyal Jailkeeper can conftown all of the investigatives. This was something I genuinely didn't think of in the role. I was still thinking of the loyal roaming jailkeeper as a pure protective role and did not take into account that the loyal modifier could conftown all of the town investigatives. This was a massive oversight.

2: The power of the rolecop being higher than I thought. The rolecop having 2-3 potential guilties, I knew about, but I failed to account for how a vanilla result to the rolecop is basically a hard-clear because given the public info (3 private communicators) and the information from the gunsmith (one scum gun), it is almost a certainty that the scum have no Goons. (A Goon could theoretically exist if the mafia neighbor was also the doctor, but the moment the mafia neighbor flips, a vanilla result to the rolecop is a hard clear.) I was, mistakenly, under the impression that the rolecop could get guilties, but no innocents, that any result not on scum could not be clearing. I failed to account for the rolecop basically being a full cop in strength. This was a massive oversight.

3: I underestimated how much agency a traitor removes from the scumteam. Usually, when a traitor is in the game, the main source of the lack of scum agency is apparent; the traitor cannot communicate with the scum and the scum don't know who the traitor is, so between those two factors, the traitor has less agency. With the traitor in a neighborhood with scum, I thought that there wasn't a reduction in scum agency because both of the traitor's main agency-reducing properties removed (able to communicate, full knowledge).
I didn't think about a third aspect of reduced scum agency with a traitor though, namely that a traitor being unable to endgame alone meant scum have less agency in who they can keep alive. Obviously, it registered that a traitor cannot endgame, that one scum needed to live, but it didn't register that DUE TO THIS, scum would have less freedom. This wasn't as massive of an oversight, but I would still call it an oversight all the same because I thought the scumteam was stronger than they actually were. (Basically, I know that traitors cannot endgame and that the game balance needs to account for their lack of ability to endgame, but I didn't take into account that a traitor by virtue of being unable to endgame reduces scum agency in of itself, just by virtue of being unable to endgame.)

With these factors in the game, I would rate it at ~65-80:20-35 town:scum. There's a 20-35% chance for scum to win this setup imo, if the scum play well and the town plays poorly. The scum need to not get caught and probably kill a PR or two; the town need to not get any guilties OR innocents or for said results to be minimal. So I still think that in the final setup, it's at least theoretically possible for the scum to win and the town to lose...

...But requiring BOTH the town to play poorly AND the scum to play well for a CHANCE of a scum win, is unacceptable, and this game has precisely that. Given that 60% is the hard-cutoff margin, a setup that is at minimum 65% and could be as high as 80% is too townsided and should not have been passed.

Unfortunately, I didn't catch these factors until well after the game had begun.

Subject: mini normal 2226; night time four

Math these last few pages feels like scum who is mad at a townsided set up, and then felt like he was in auto lose and is madly trying to widen the elim pool.

I mean, he was, but to be fair; he had every right to complain and I owe a big huge apology to him postgame.

The setup should not have been passed as-is because it's too townsided in hindsight given the interactions I didn't account for. Sorry. <3

[mastina]

Needless to play even though the setup was broken town deserved this win and played way better than scum.

I'm inclined to agree with this btw.