Open for Micro Queue

[Jingle]

I'm a little feverish now, but I had an idea for a game to throw into the mostly empty Micro queue and wanted a quick and dirty balance analysis before deciding whether to scrap it or move forward.

8 Player

6x Town
2x Mafia

1 Cop
1 Jailkeeper
1 Roleblocker
1 Tracker
4x Vanilla

Roles randomized independently of alignment. If town receives all PR's, Mafia chooses one of their own to become ascetic at the beginning of Night 1. Ascetic flips as goon.

Roleblocker takes preference over JK should they cross target.

Any obviously breaking strategies/setup issues I'm overlooking?

[Jingle]

I lean against having the scum draft a role. Not only does it fundamentally change the style of the game (WIFOM of role selection instead of roles being wholly null) it could lead to an unfortunate situation of lynching one scum and having a ridiculously small lynch pool. Particularly a goon flip means that scum can’t afford to shoot prs and can’t afford not to shoot prs, making the game functionally unwinnable day 1.

I’ll think about it a little more and maybe post some numbers tonight. Tracker could maybe become rolecop, but jk and rb together helps to prevent any hard innos/guilties.

Thanks for the feedback guys. I’ll see what I can do to fix the setup. Honestly I was more worried about scum getting two roles and stomping because of it.

[Jingle]

Okay, I'm not gonna do analysis for every case, because there are a ton of them and I frankly don't have that kind of time. Most of the 1v3's seem pretty balanced to me on the premise that it's not a drafted pick. I'll look at the cop one particularly, the best case scenario for scum (JK AND RB) and the no PRs for scum case right now for ease of my mind.

Scum Cop

Setup is Tracker, RB, JK V Cop.

Tracker cannot get a guilty assuming Cop makes nightkills until cop is lynched. Cop is unlikely to be D1 lynch.
RB and JK cannot get innos or guilties as long as either is alive and both scum alive.

Best case for scum:

JK lynch D1.
N1 RB hits Tracker.

D2 Mylo with no additional information and no town claims. Cop can fake guilty for victory, or let things play out as is, depending on situation. Probable scumwin D2.

VT Lynch- (Not hard numbers, approximations) 50%
JK targets scum who kills or nightkill target (1/7 + 1/6) and is not targeted by RB (-1/7 P) -> 85/294
RB targets scum who kills (1/7) and is not targeted by JK -> 6/49

Successfully stop NK -> 121/294 -> 41% for one stopped nightkill, which likely outs both JK and RB.

Tracker gets soft guilty 1/7 -> Not insta lynch due to cop safeclaim, not possible if tracker gets rb'd, jk'd, or scum nightkill stopped. (1/7 * 30%)

Odds of Scum nightkill hitting 59%. Odds of scum nightkill hitting town power 50%.

Assuming scum nightkill is stopped D2 is

5v2 with very little usable information. Similar to 7v2 nightless, which is scum sided, but with the possibility to get more information future days. Reasonably balanced. No innocents possible. No hard guilties possible. Soft guilty is approx 4% of the time.

In the case that scum is lynched Day 1 (25%), odds of prevented kill remain the same. Odds of 3 innocents: RB targets not tracker, not scum, not JK and not kill, 3/7 AND JK targets not tracker, not scum, not RB, not kill, and not rb target 2/7, AND tracker targets not kill not rb target not JK target, not scum, not kill 3/7 AND scum kill not tracker not JK not RB 3/7. 2% Scum autolose.

Tracker guilty, scum lose. 1/7 x 6/7 x 6/7 x 5/7. 7% scum autolose. So, if Scum is lynched D1, scum has a 10% chance of autolosing D2, which is reasonable based on superb day play D1 by town. Elsewise, scum have a severe uphill battle, but that is a reasonable expectation from a scumlynch D1 in a 6v2 setup. This scenario seems incredibly fair, unless I'm misreading something massively.

Best case scenario for Scum:

JK+RB vs 6 VT Tracker Cop. 1/32 of games. (3/32 games scum will have 2 PRs, which is ~9% of setups.)

No protections possible, meaning only one mislynch. Scum probably not lynched D1 based on claims.

This looks crazy scumsided. Possible mitigations: could force scum to use power or kill. (Doesn't affect scum cop, only really onerous in cases where one scum is lynched or both scum have powers, and in fact I really like this option.) reroll setups where scum have 2 PRs. (Dislike this immensely, for the same reason as I dislike the draft option. Lynching a scum PR basically ends the game in town's favor by clearing 3 players.) Give town some advantage they don't know about in the case of 2 scum PRs (Vt's all get 1x BP but aren't told?) needs thought.

Any ideas here would be immensely helpful.

I'll do the ascetic later, cause I have to leave five minutes ago, but basically, the ascetic is immune to cop and tracker as long as rb and jk are alive because rb OR jk targeting and lying about it (or actually targeting) is more immediately likely than them being the ascetic. Cop becomes most important town power. RB is negative utility unless ascetic is lynched. NK prevention becomes an only if the jk targets the nk target, and town can count on 0% of this information. Honestly, I think this is the optimal roll for the setup because it relies almost exclusively on Day play with only a slight chance of PRs being immensely impactful, but still leaves scum with a lot to play around.

[Jingle]

8 Player

6x Town
2x Mafia

1 Cop
1 Jailkeeper
1 Roleblocker
1 Vanilla Cop (Investigates a target, if their target is Cop, Jailkeeper or Roleblocker they receive Non-Vanilla. Otherwise, they receive Vanilla.)
4x Vanilla

Roles randomized independently of alignment.

If town receives all PR's, Mafia chooses one of their own to become ascetic at the beginning of Night 1. Ascetic flips as goon.
Each scum can use a PR or Nightkill.
If scum receives two PR's, all Vanilla town are Uninformed 1x BPs that flip as vanilla town.

Roleblocker takes preference over JK should they cross target.

I believe this solves all of the issues regarding 0 and 2 scum power roles. I don't think it unduly changes the numbers for any of the 1 scum power role cases, though if anyone has a reason one of the cases is unbalanced I'm happy to listen. Otherwise I'll probably run it in the micro queue in the near future.