[MAY CHALLENGE]

[Jingle]

drama

:jazzhands:

[Jingle]

The lazy part of me wants to submit innocent evil people. The slightly less lazy part of me wants to find the balance points for innocent evil people. The rest of me wants to make a new setup.

Eh, might as well put in the numbers I just worked:

Innocent Evil People (Daystart)

Setup is VT v Goon. Each night, the mafia chooses one of their own to leave the game. Each day, a player is lynched. Mafia wins if they all escape, town wins with a single mafia lynch. Mafia escape is compulsory. Lynching is not.

5p: 3v2 60% town EV.
6p: 4v2 50% town EV.
7p: 5v2 43% town EV.
8p: 6v2 38% town EV.
9p: 6v3 62% town EV.
10p: 7v3 56% town EV.
11p: 8v3 52% town EV.
12p: 9v3 48% town EV.
13p: 10v3 44% town EV.

Innocent Evil People (Nightstart)

Setup is VT v Goon. Each night, the mafia chooses one of their own to leave the game. Each day, a player is lynched. Mafia wins if they all escape, else town wins. Mafia escape is compulsory. Lynching is not.

6p: 3v3 60% town EV.
7p: 4v3 50% town EV.
8p: 5v3 43% town EV.
9p: 6v3 38% town EV.
10p: 6v4 62% town EV.
11p: 7v4 56% town EV.
12p: 8v4 52% town EV.
13p: 9v4 48% town EV.
14p: 10v4 44% town EV.

[Jingle]

Any thoughts on this? I'm wondering if 1/3 P is too high in high-player-count games, like maybe it should just be a constant = 3 or 4

Yes. Probably 3. 3 has a 25% chance of any given triplet winning the game for town or scum. # of potential triplets is P-2(first two can't complete a triplet)-scumkills. As player list increases, the players have a harder time organizing, more than making up for the scum wincon of Oh shit we ran out of lynches decreasing in power. Honestly, I suspect even at 3 that it's too scumsided.

Also, might be interesting to have the scumkill be compulsive every night but always have to target someone of the opposite alignment of the lynch. Gives scum more agency, and trims a large playerlist to be more readable faster.

[Jingle]

Two extra VTs added to avoid losing D2 by parity
Capturing a goon now blocks the Mafia's Imprison outright
Mafia's Capture, renamed Imprison, now reduces vote weight to 0 and is unlimited
The Dragon is no longer bulletproof, and wins when nightkilled

Imprison might as well be public knowledge, inability to vote is pretty confirmable unless you want arbitrary modkills as part of the setup.

If capture can be hammered before lynch, dragon D1 happens every game.

(Either Dragon is captured D1, or someone else is captured and the Dragon goes "Hey yall, Imma dragon. If you don't lynch me maf imprisons me tonight and the game is over. Peace!"

Both Capture and Imprison are functionally investigative results. Investigative results that can be fucked with, sure, but capture and then imprisonment clears two players. Capture and no imprisonment implies a scum capture, because imprisoning the dragon is probably worth it. It's definitely worth it late game. Also, both investigatives are public.

[Jingle]

Titanzar Mafia

VTs and Goons: 5-3, 8-4, or 11-5

A formshifting member of the Skalg alien race is found in the engine room destroying the cooling core and engine shutoff switch. This luxury spaceship is going to explode, and there's not enough escape pods, so the captain keeps chanting "Humans and human children first!"

Each day, players vote for someone to escape. A player with majority or deadline plurality vote jets off in an escape pod and lands on the nearby desert planet. They lose access to the spaceship thread and gain access to the desert planet thread. Later escapees can summarize discussions on the ship but cannot quote it, like neighborhoods.

The mod has secretly chosen a population size for the desert planet before the game. When the desert planet has 4 + 2z players, the captain admits there is only one escape pod left and suggests that someone use the ship's DNA radar to help use that pod wisely. The vote for the last day is to elect a DNA radar operator. Once elected, the DNA radar operator becomes mute and learns the alignment of every player on the desert planet. Then she selects one last escapee. She may not select herself as the escapee, unless she is scum, then she may.

Then the space ship explodes killing all occupants. The residents of the desert planet panic. They play a standard game of mafia complete with factional nightkill.
Perished players still win with their faction regardless of where they perished.

Seems likely to be broken by designating an order before sending the first player. You don't actually get information from who leaves, so picking all of the 'saves' before voting just makes sense. If you pick the radar op first, it becomes autowin if town AND town has a majority on the planet. So basically its just a flipless game of vote for town that all happens on D1.

[Jingle]

Because attempting to track all of these votes as the game progresses by hand is rather annoying (and difficult), the moderator instead only tracks the number of players who wish to end the day earlier. Once the day is over, the game enters a special Twilight phase where players send the moderator a ranked list of their choices for the lynch, including themselves (i.e.: it must have the entire player list), (with the person they want lynched most at the top). Players that fail to submit a list in time rank players in the order of the player list (i.e.: not random). The moderator then determines the Condorcet winner (likely using a calculator of some sort) and posts the lynch as well as everyone's lists. In the case of a tie or an undeterminable winner, no lynch occurs, but the moderator should still post everyone's lists. Then, the next day begin. Standard win conditions apply.

You could make that trackable without TOO much work on the mod's part, actually. Give everyone a PT and post a list of all pairs. Whenever they want to vote, they just copy the list and bold the names they choose.

You could even make it public PTs by giving playerlist PT access and individual players mod access, banning PT edits. That way the player voting in that PT can vote while it's locked.

As a mod I'd find that less annoying than PM barrage at the end of the night, and as a player I'd prefer the ability to see people's reads change over the course of the day.

Also it's a significant enough mutation that I'd say go 50% EV and judge based on a couple of runs.

Edit: Oh misread. You don’t need to post a list of all pairs, just playerlist. Somehow I thought this was some weird gladiator condorcet hybrid originally.

[Jingle]

Actually, just getting rid of the Radar operator is probably enough to make it not broken. It's a group decision in a mini on who plays the micro, basically. (obviously not necessarily a mini and a micro, but w/ever)

[Jingle]

I could fix the setup by making both the vote Capture and the Mafia factional Capture a treestumping power, but I'd probably have to make the setup a Large (and force the Lynch to be hammered before the Capture) and I don't see it as a fun setup.

The Jingle Giveth, and the Jingle Taketh Away.

More importantly, you can just make capture straight plurality and resolve it immediately after the lynch.

[Jingle]

I can't figure out how the dragon should play.
That seems like a good thing.

Protown, probably. Town only needs to find dragons if they're not finding scum, so the faster the scum is caught the better the odds of a dragon lynch.

Oh, and to solve the issue of :oopsallclears: make imprisonment invisible (no votesteal) and the capture actually a straight up jailkeep. Public jailkeep is pretty weak for town, but the strength after the first scumlynch is amazing. Dragon is probably scumsided inherently. (Town guesses first, Scum has >>info when guessing). You could probably drop the autowin from Maf too and just have it end up being a double kill if they hit dragon.

Balance turns into 8v2 with a lategame only town PR, a win mechanic for town, and a 3p with a semibenevolent wincon who slightly townsides.

[Jingle]

Dragon exit on imprison IS functionally an extra kill that night. Not sure if the additional daykill is warranted, tbh, although it does put the game back on odds, so... probably not the worst thing for the game?

[Jingle]

Probably fair, but 2v9 mountainous is boring AF.

I'd rather give scum additional power as a counterbalance.

[Jingle]

So...

Recruitable traitor, Kingmaker, and rotating PGOs? Seems really scumsided. Also, there should never be a case when the executioner chooses a lynch on a werewolf, so that rule is superfluous. You could probably drop a goon and be closer to balanced, tbh.

[Jingle]

I really really really wanna break that. But I don't see a way to.

NK15, CFJ ATTACK!



(I feel like I should clarify that I don't try to break setups because I don't like them, but because I think the only way to make a good game is to start with a mediocre game and then break/fix that game repeatedly until it's better. In fact, I'm more likely to try to break a game I like than one that just isn't interesting to me.)

[Jingle]

You could bump up the Dragon win % by dropping multitasking from the scum.

[Jingle]

As is, Dragon has an ~23% chance of winning in the first cycle (D1 Lynch 1/11 N1 NK 1/7) and an ~24% chance of losing in the first cycle, assuming scum isn't captured D1. The second cycle gets complicated, based on whether there are scum lynches, whether scum kills prior captures, whether town lynches prior captures. Dragon has a really low chance of winning via survival unless town hits autowin early though (scumlynch then scum capture).

[Jingle]

Actually, assuming you take town autowin on D1 as a dragonwin, that's a 25% win percent for the dragon on D1, which is perfectly acceptable for a Jester-like role.

Although RC never wins as dragon because he 100% wants to be the D1 capture as either scum or town.

[Jingle]

Broken by no flip.

If you just never flip a tile, you run 0 risk of an increase in scum players.

[Jingle]

Broken by no flip.

If you just never flip a tile, you run 0 risk of an increase in scum players.

I realized that now.

And then if you give ties to scum, the town just have to flip a tile, lock it, see that the new result is town, then no-flip.

I edited it to a different probably still broken setup now.

EDIT: The new one is still terrible. I give up completely.

Changes to make it viable if potentially boring AF.

All players, regardless of side, can vote. Locked scum loses their vote, locked town does not.

1 random pair is white on both sides.

Scum may choose to flip a piece each night, cannot flip a piece flipped during the previous dayphase.

Wincon is have majority of your color locked in.

Each day is either flip or lock, not a specific one of the two.

[Jingle]

On a side note, having fully half of the player list not playing at any given time is probably boring AF. I would suggest at least giving inactive players a PT to discuss in.

[Jingle]

Probably the wrong place to put this, but not a mafia game:

Hidden Reversi.

N players, N is odd.

Each player is a reversi tile who can choose to be black or white. At the end of [deadline], the color which is chosen by fewer players wins.


Which lead me to an actual submission:

Hidden Reversi Mafia

3 Goons
10 VT

Each day, players choose White or Black. Standard lynching. All players with the more popular choice are lynched and all choices are revealed at the beginning of the next day. Each night, scum chooses a player to nightkill. White Flag.

[Jingle]

Baron autowins D2 imo. Better to have them claim D1 and be crowned D2 than risk a scumwin out of the blue. Scum should never cc because while the player can be smuggled out if they survive D1, they have a 50% chance of instalose.

4/11 scum lynch D1. For ease of calculation, we can look at both scum teams escape and then both scumteams holster.

Both escape, either 7v1 or 6v1v1 on D3. 1/8 or 2/9 scumlynch.

Both holster, 6v2v2 or 7v2. 2/5 or 2/7 scum lynch.

Alternatively, it's an option to just no lynch D1 and 3 and treat it as a vote for town setup. you have a 7/11 chance for the first town coronation and a 6/10 for the second (in this case the escape is pointless), which means a 38% win rate for town. If you scumlynch D1, it turns into 58%. If you scum lynch and have an extradition, it turns into 75%. Town lynch and single extradition is 42% Town lynch and double extradition is 54%. Town lynch and no extradition is 33%. Basically, scum should never escape as just doing so makes it far more likely for a town win by coronation. Scum cannot win by majority if they escape. Scum win by majority is: (Lynch Town 7/11, Baron Escape, Lynch Town 6/10, Crown Town 5/9, Lynch Town 1/2 OR 10.6%)

Scum win by coronation (Assumed no escapes, Baron Crown D2): Each individual scumteam has a chance of 24.8%.

Town EV is 40.8%

Game ends after D6 regardless.

Multiball, with theoretical win ratio of 40.8:35.4:35.4 doesn't seem bad at all. Escaping skews towards town, an opposite team scumlynch increases a team's chance of winning. Probably wouldn't play it personally, but seems fairly balanced. Probably more fun without the Baron, tbh. Could start with 11p, make it Coronations on D3 and 5 and not have a player who is basically guaranteed a win without changing your EV.

[Jingle]

Oh, yeah, definitely leave the escape in. It's a WIFOM-y trick that can hail mary shot for a team. It's just irrelevant to the EV calcs because the only way to suitably calc EV is to assume random lynching, and random lynching says it's never the right play to make.

And yeah, town should always coronate the Baron D2 as is. It ups their win by 3%, (which admittedly is not much) over the course of the game. Town should actually have a higher win % than 40.8% because they have two optimal strategies and should always follow the one with the higher winrate. But that is a more complicated probability tree than I'm willing to look at at the moment because it requires value judgments.

[Jingle]

what do you think it's kinda unfun or not very attractive about the setup?

Multiball is iffy. Nightless is iffy. This particular setup is multiball nightless in a way that gets rid of the best part of multiball (trying to dodge the opposing NK) but keeps the worst part (It's definitely to scum's advantage to get rid of the other team, so scumhunting via "Are they scumhunting?" doesn't really work well).

It's not that it's a bad setup, necessarily, just not the type of setup that I'm personally interested in.

[Jingle]

A player/slot cannot be crowned twice.

This is a notable addition that is good to have, although I had assumed it to be the case.

Scum can no longer viably reach a majority scum win (Game ends after at most 2 lynchs). Town winning through scum lynches is similarly unlikely.

Assuming Baron claim D1 (which is still probably the likely to be correct option) EV shifts to 45:28:28 with no possibility of scum teams joint win. Yes I'm aware that adds up to 101%. Rounding.

Honestly, I think setup two was better if you include the no doubling up on crowns line.

[Jingle]

In V2 and the version where you just have coronations on D3 and 5 and no baron, there is a 10.6% chance of a mutual scumwin through endgaming.

This not being there isn't necessarily a bad thing, just worth noting in that a mutual scumwin possibility adds to the win percentages of both scum teams allowing for a larger town win percentage that is close to the individual scumwin percentage. It's an easy answer to the old multiball question of "Is it correct to balance for 50:25:25 or 33:33:33?"

As far as the Baron-nation changing the EV, it's actually irrelevant. The Baron being coronated reduces the number of players, but it doesn't reduce the number of mislynches or lynches required to win, nor does it change the portion of the cycle you're in.

The assumption that is afforded by Optimal play is that the Baron is never lynched. This is because claiming Baron as scum necessarily lowers your win chance (while raising the win chance of the opposite scumteam), so any Baron claim can be treated as trustworthy and thus town, regardless of the day it comes on.

The subjective influence of Baron reads is also not something that is relevant to EV, as the individual player in the slot needs to be considered in order to figure out the impact. To simulate this, you use a fully random lynch pattern among all slots that can reasonably be lynched. This, coincidentally, is why 50% EV =/= Perfectly Balanced. There are things in the game that don't impact EV but do impact the actual game.

Consider, for a moment: 7v2 mountainous. From an EV standpoint, that's identical to 50v2 where 43 of the town are D3 suicidal ICs. I'd bet that no one would think they're exactly the same town win chance.

[Jingle]

FTR, this isn't a "Good at calculating EV" setup. It would take longer to write the program than to brute force the probability tree for a scum team and a joint scum win (which is what I've been doing for this setup).

Baron actually doesn't have any incentive to claim prior to threat of being lynched in the current incarnation.

I could see changing Baron to: If it would be lynched, instead, the Baron loses, Deadline is extended by :blah: and the Baron exits the game in shame. If it would be crowned, instead, the Baron wins, Deadline is extended by :blah: and the Baron exits the game in glory. Either way, the day continues as before.

I believe this would have a slight bump to both scum wincons while not meaningfully dropping the town wincon, but can't be sure without checking the numbers. It does mean the Baron still has the option of being a semi-IC from the outset, but has a reasonable course of action of hide-and-coronate while leaving it as a protown role. It also means both lynching and coronating a Baron are theoretically more valuable for scum the later in the day it happens, especially if the deadline extension is short.

[Jingle]

edit: I'll be trying to run this setup in marathon btw, if anyone is interested

Advertise in the open setup marathon group (link in my sig) when you know when you're able to run it and you'll probably get a few interested parties

[Jingle]

Not only are they allowed, they were explicitly required a couple of months ago.

OTOH, if you wouldn't mind updating your setup with major changes into new posts it helps me to see where your thought process leads and also I'm more likely to actually realize you updated your setup. (I don't look back at the early pages as much.) If not it's cool, I just think it's easier if we can look back at what has changed to get an idea of what you actually want in your setup while giving advice.

[Jingle]

Toriel: As long as no player has been Killed, Toriel can protect another player from a kill each night.
Alphys: If the Core is reached in a Pacifist Route and Alphys is still in the game, she can immediately choose another player still in the game to be Spared, causing the Pacifist Ending to be played out with five players instead of four.
Papyrus: On each even night, if at least one player has been Spared and at least one player has been Killed, Papyrus will learn the alignment of all Spared players.
Asgore: If the Core is reached in a Neutral Route and Asgore is still in the game, he may jailkeep one player each night (starting Night 5).
Undyne: On Night 2, if no player has been Spared, Undyne may choose any player, including herself, to be permanently bulletproof. They are not informed of this. This goes away if a player is ever Spared.
sans: If the Core is reached in a Genocide Route and Sans is still in the game, he can immediately kill one player.

First of all, the numbers on this seem . If you go with the pacifist route, it's the goal to BARELY have a majority of town players for the last phase, which is pretty :/

Alphys is a named townie. If you go pacifist route, she ALWAYS claims day 4 and gets saved. If she chooses to add to the Spared pool, odds are she's adding mafia. (2 mafia spared ever means that there's a maf win, so you're going from 1/4 max mafia to potential autolose if Alphys picks the second scum. If there aren't scum in the 4 original spares, it's already a town autowin.)
Papyrus and Asgore both seem strong if you go into Neutral from 3 spare/1 kill.
sans is a vig or a named townie, which means he's identical to Alphys in the pacifist route.
Undyne is a deathproof townie (assuming they survive D1)

Optimal strat is probably Spare->Spare->Spare->Claim if not Papyrus/Asgore ->Spare claim or kill if no claim, obviously tailored if someone flips. If a named player (other than the mentioned) is about to be spared, they claim and get spared anyway. Kill D4 if no claims made at that point.

Undyne is the only one for which this strategy doesn't get the most out of the role (1/6) and even then, the

chance

of a bulletproof IC isn't worth losing the utility from every other role IMO.

[Jingle]

Undertale mafia seems like it has lots of even player number days in it.

It shouldn't, no. It still has a Lynch/NK, it's just that some of the lynched players can come back later.

Only reason you have evens is pacifist ending (in which case it's just MYLO, unless pacifist becomes nightless), successful Doc or Successful JK. Which is fine.

[Jingle]

sans is the only character for which you go genocide route ever, and he vigs.

Also, genocide route is Lynch 4 Kill 3, so you enter the core at 2 alive. Sans literally says if you have gotten to the genocide route and Sans is still alive, town wins.

[Jingle]

That setup would probably just be better as a closed, tbh. (Which is a thing, we've had an undertale mafia before and it worked well iirc.)

Also, named townie isn't an 'issue' here. It's just a thing with opens.

[Jingle]

blah

Eh, given that scum is considerably more likely to catch the dragon than town is and they get 2 functional extra kills from it, I don't really see the admittedly high utility of the towncapture as an issue. It's only super good for town after a scum lynch. If scum is lynched D1, the setup turning into 1v8 with every scumkill confirming a town and a potential for a free doublekill down the line seems harsh, but not unreasonable.

[Jingle]

I can go with the convention of reposting.
This is what I currently favor the most as a submission.

V2 was the same as this, but with secret captains and key re-gifting. Bad idea.

Titanzar Mafia V3

Semi-open
1

Town Voting Captain

2

Voting

Captains

of

Random

Alignment

: Town, Scum, Spaceship Lyncher, or Spaceship Unlyncher

8 Town Passengers

2 Scum Passengers

Evil saboteurs rigged the engine of the Titanzar cruise spacecraft to explode soon! Captains must go down with the ship, but some passengers can be saved in the limited escape pods. A good captain would help save the passengers that aren't part of the organized crime that ruined the engine... but it's not clear all of our captains are good..

Each day, all players collectively vote one player to grant an escape pod key. The voted player will receive the key and be scheduled to escape when the ship is going to explode. A captain can never escape in an escape pod, they must go down with the ship. They can vote, but can't be voted.
When six escape pod keys have been distributed each passenger holding a key gets into an escape pod. Then, one of the town captains yanks one escape pod passenger back into the ship and sets the autopilot of the escape pod to a well-known comet instead. (The survivors will never be discovered and rescued unless we make something noticeable happen in this star system, too bad if that means one fewer pod!)

Then the pods land on a nearby desert planet, and the space ship explodes killing and flipping all 3 captains and the 5 remaining passengers.

Each captain beyond the first has a randomized alignment, decided by a d18.
1-9 Town
10-12 Scum
13- 14 Evil Spaceship Lyncher
15 Good Spaceship Lyncher
16-17 Evil Spaceship Unlyncher
18 Good Spaceship Unlyncher

Lyncher and Unlyncher do not receive their "good" or "evil" flavor in their role pm, only in their flip. Spaceship Lynchers win the game if their target dies in the spaceship explosion and lose if their target survives the spaceship explosion. Spaceship Unlynchers win the game if their target survives the spaceship explosion and lose if their target dies in the spaceship explosion. "Good" flavored roles are targeted in a way that benefits the town win condition.
When Lynchers and Unlynchers flip, the survivors learn whether they were Good or Evil, but the survivors do not learn who the target was.
After escaping the explosion, the five players living on the desert planet play a standard game of mafia with nightkills.

Spaceship lyncher and unlyncher probably just always claim, tbh. Lyncher target has a 2/3 chance of being scum, which is better than random by a considerable amount. Unlyncher target (if claimed) is confirmed post explosion. Both seem like they're really annoying for scum. To fix this, I'd suggest making the target's alignment irrelevant to the roll and just having Lyncher/Unlyncher Captains. Even still, they probably should just claim and agree to townside if the town agrees to let them win with them.

[Jingle]

You still as town want the unlynchers to claim and to save their targets, because when they flip as good/evil it confirms the alignment of their target. At worst, that turns 5p LYLO into 5p LYLO with a conftown. At best, it makes the game autowin for town.

Seems like other than an interesting fakclaim choice for scum they don't actually change much in the setup though, so removing them or making their target unconfirmed and normal randomization both seem like completely fine options to fix that and maybe make them interesting.

[Jingle]

FWIW, I think scum captain is the most interesting role to play in the setup.

[Jingle]

Yes. Does the game end after Lynch 5 or is the last surviving player lynched? I assume the former, given the draw criteria, but it is worth asking.

5 lynches, each team has a 25% of a D3 win, a 12.5% chance of a D4 win, and a 6.25% chance of a D5 win. Therefore there's a 12.5% chance of a tie. Within that 12.5 is a slight preference towards a scumwin, but not a large one.

[Jingle]

Undertale Brain Dump:

I'm fine with starting another thread for this, I just had already decided to do this when I saw that.

Spoiler: Trivial analysis setups

Genocide: Toriel, Alphys, Papyrus, Asgore

7v2 with a named townie, significantly scumsided.

Pacifist: Alphys, Papyrus, Asgore, Undyne, sans

7v2 with a named townie, vote to save mechanic for first 4 days.

Neutral: Alphys, sans, Undyne (inactive), Toriel (inactive)

7v2 with a named townie. One day (at least) is a no lynch, no kill on N4 Significantly scumsided.

Spoiler: sans Genocide

7v2 with a named townie. If the named townie would be endgamed, instead the named townie wins. Sans should never claim unless about to be lynched or 3p LYLO. Still significantly scumsided.

Spoiler: Toriel

Pacifist: As with trivial pacifist, but with Doctor.

Neutral: As with trivial Neutral, but with X shot doc. Still kind of scumsided.

Spoiler: Papyrus

Neutral: If Papyrus survives N4, town probably wins. Otherwise, strictly worse than Genocide route, thus Papyrus NEVER CLAIMS pre N4.

Spoiler: Asgore

Neutral: If Asgore survives N4, difficult slog for town. If town lynch, and mafia saved, D4 starts with 3v2. If 2 townlynches, already endgamed. If Asgore doesn't survive N4, scum probably ROFLstomp. Very swingy based on whether town ever spares scum.

Spoiler: Undyne

Active:

Neutral: Same as Neutral but with full BP named town. Strictly worse than Genocide.
Genocide: 7v2 with full BP IC.

tl;dr: most of the setups are probably scumsided. I think I've convinced myself that Asgore probably claims and gets spared D4 since the potential for scum sweep with Asgore neutral is really strong. Pacifist unless Papyrus is probably still the most viable choice for town.

[Jingle]

I’ll run numbers tonight, CFJ. First thought though is that it’s probably breakable by organizing clear chains.

Edit: Thinking it over, it's not particularly breakable. I feel like it's probably slightly scumsided, but I'd encourage you to run it as a marathon over the weekend to get a bit of real data. Numbers are, as you mentioned, a nightmare.

[Jingle]

Wrt connections, each townie who makes 4 connections has a 7% chance of winning the game for town. Knowledge is gained whenever this process fails, in the vein of there being scum in a pool of more than 50% of the players. 6 players manage to make 4 connections.

If 2/3 of the lynched players are scum town wins. That’s a 23% chance; the other 77% of the time, there are between 3 and 4 chances at a town win. Assuming 4 for the sake of argument, that’s a 44% total town win rate. It’s 39 if you assume all scum will make 4 connections.

Additionally, scum can choose to pair to prevent themselves from being the 0 elimination, so 23 is actually a high number.

[Jingle]

I think I preferred Titanzar with the variable number of town/scum captains and the yanking a player mechanics, tbh.

Both flavor wise and mechanically.

You could just do 0-2 Scum Captains and 1-3 Town Captains. Weight the two random alignment ones however you want.

And the 6, but town captain yanks one idea was pretty cool, tbh.

The new setup is very similar to BNL's branching mafia, which wasn't a bad setup by any means, but wasn't good enough to need the variation. Also, the whole first phase is pretty pointless, AND requires you to have 5 days that might as well not be there.

[Jingle]

How is May already over? WTF kinda sorcery is this?


Dargon Karpturer v3 TempLich - I actually kinda dig the setup. I'm not convinced it's wholly balanced, but I think it definitely deserves to be run.
Hidden Reversi Jingle - Quick and dirty. probably biased.
Titanzar v6 pops - Interesting way to handle a mostly vanilla setup while keeping the endgame exciting. Probably great for those weirdos that like mountainous. Weirdos.
Will o/t People Ircher - Interesting mutation. Definitely playable.
Undertale S_S - I still think this needs more polish before being playable, but I haven't had the time to work on how to do so.
Connections cfj - I think this is probably still scumsided? Would probably play in a marathon setting though.
Defender BNL - Public PR is always a plus, but it's kinda boring for an open, tbh. Not much going on.
Coronation 2 Rainn - This is a strictly better setup than F&I imo. OTOH, fuck F&I.
Chaos V Order Relly - Functional, but meh.
Wolves NK15 - Seems really scumsided, tbh.

[Jingle]

Hidden Reversi: Broken for town: just have every player choose a random colour (without claiming it). If scum all choose the same colour, it'll almost certainly be the most popular, wiping the scum out (and yet scum will be forced into this after a D1 scum lynch, if it happens, so lynching scum makes things even more townsided). So scum split 2:1. There's still a >50% chance two scum die (again, winning for town, as it's White Flag). The rest of the time the setup probably ends up at 3:2, which is still winnable for town, so town are doing pretty well in every scenario.

Hmmm. Not to be included with voting, but probably better to drop white flag from the setup, tbh. I think it was a holdover from considering the setup without an actual lynch.

[Jingle]

I was asked to double check numbers, but it looks like SS did so and I'm not going to bother double checking a double checked set.

:p