Mini Normal 2224 | Garfield Minus Garfield | Game Over

[MathBlade]

Respectfully I would disagree Mastina I think that in this case this setup was a bit too townsided.

1) There is no clear path to victory. Assume that not only scum knew who was which of
Town Motion Detector
Town Loud Follower
Town Loud Neapolitan
Town Lazy Roleblocker
VT*6
Mafia Informed Visitor (knows the setup)
Mafia Goon *2

Assume that scum magically know which player is which role. They have to predict which guilty to lose a scum to. The odds of three townies never checking a scum in 3 nights is astronomically low. (Yes some roles don’t check townies I am simplifying this way down)

This means that bussing is 100% off the table.

This setup handcuffs how scum play which is imho a hallmark of a townsided setup.

My general rule of thumb is hard guilties should always be T-2 at most and soft guilties/play kills for the remainder.

2) Loud follower does not have a counter. The counter cannot simply “be good and creative”. Imagine for a moment loud follower and loud roleblocker + goons is the whole setup. Once down to the last goon follow the cop takes place. Roleblocker and follower work in tandem and scum can’t do shit. Scum have to be given a way to prevent follow the cop (follower).

3) Knowledge of how screwed you are and no ability to do anything about it can be rather demoralizing. This has been a trend rather sitewide. Calculasia was also guilty of this too on a worse scale. Informed is not some fix for the amount of power it just tells scum they can’t bus 99% of the time how it’s used here. Unless informed actively helps day play strategy it just tells scum “we gave town a lot of power good luck”.

[MathBlade]

Mafia:
1. The Mean Value Theorem (Mafia 1-Shot Strongman)
2. The Chain Rule (Mafia 1-Shot Redirector)
3. Derivatives (Mafia Multitasking Jack of All Trades (Flavor Cop, Doctor))

Town:
1. Limits (Town Doctor)
2. First Derivative Test (Town Odd Night Cop)
3. Second Derivative Test (Town Even Night Rolecop)
4. The Product Rule (Town 2-Shot Roleblocker)
5. The Ratio Test (Town Informed Neighbor) --> Knows flavor name of neighbor
6. The Root Test (Town Informed Neighbor) --> Knows flavor name of neighbor
7. The Intermediate Value Theorem (Town Vanilla)
8. Euler's Method (Town Vanilla)
9. Integration by Partial Fractions (Town Vanilla)
10. Tangent Line (Town Vanilla)

Calculasia was this problem on admittedly a worse scale.

Newbies are balanced with cop + doc + full roleblocker.

Here scum had less power against a split every role cop, full doc, town had the roleblocker and a secret “masonry” if used properly.

Against a one shot strongman and a power that relied on guessing the cop correctly and the flavor cop was useless to someone who had a degree in math.

Informed does not fix problems like OP shit. It just doesn’t.

Informed should help scum with their day play like “You know that there is a one shot and full of a same role” this gives scum a hint to pr fish may be beneficial as town usually thinks those are in conflict.

Or “You are informed there is a N2 role that can create confirmed townies”.

This tells scum they have a time limit and they have to find a certain power before creep gets them. Again, not limiting to play but helpful.

[MathBlade]

If town mass claims D1 they win 100% of the time which is another sign of a bad setup.

[skitter30]

) Loud follower does not have a counter. The counter cannot simply “be good and creative”. Imagine for a moment loud follower and loud roleblocker + goons is the whole setup. Once down to the last goon follow the cop takes place. Roleblocker and follower work in tandem and scum can’t do shit. Scum have to be given a way to prevent follow the cop (follower).

3) Knowledge of how screwed you are and no ability to do anything about it can be rather demoralizing. This has been a trend rather sitewide. Calculasia was also guilty of this too on a worse scale. Informed is not some fix for the amount of power it just tells scum they can’t bus 99% of the time how it’s used here. Unless informed actively helps day play strategy it just tells scum “we gave town a lot of power good luck”.

^

[mastina]

Assume that scum magically know which player is which role. They have to predict which guilty to lose a scum to. The odds of three townies never checking a scum in 3 nights is astronomically low.

Oh? Are they?

Let's do the math then.

For the follower to get a guilty:
They need to follow the scum making the nightkill.
On N1, there is a 1/11 chance of that occurring. (9.090909090909091%)
On N2, there is a 1/9 chance of that occurring. (11.11111111111111%)
On N3, there is a 1/7 chance of that occurring. (14.28571428571429%)

For the roleblocker to get a guilty:
They need to roleblock the scum making the nightkill.
And keep in mind that if they block scum that didn't make the nightkill--they can't block that scum again even if the scum DOES make the nightkill later. (Say they target the scum not nightkilling N1 but said scum nightkills N2--the roleblocker will fail for the entire rest of the game.)
This one is thus something I don't know how to calculate for sure, but let's use the formula above for simplicity's sake;
On N1, there is a 1/11 chance of that occurring. (9.090909090909091%)
On N2, there is a 1/9 chance of that occurring. (11.11111111111111%)
On N3, there is a 1/7 chance of that occurring. (14.28571428571429%)

For a motion detector to get a guilty:
They need to detect the scum making the nightkill, AND have no town PR target the player who makes the nightkill (not to mention, the scum have an easier time talking their way out of an MD guilty than any other type of guilty).
On N1, the odds of that happening are (10/11) * (10/11) * (10/11) * 1/11: 1000/14641, a 6.830134553650707% chance.
On N2, the odds of that happening are (8/9) * (8/9) * (8/9) * 1/9: 512/6561, a 7.803688462124676% chance.
On N3, the odds of that happening are (6/7) * (6/7) * (6/7) * 1/7: 216/2401, a 8.996251561849229% chance.

What do you get when you add all three percentages up?
34.48773448773449% for the follower, assumed 34.48773448773449% for the roleblocker, and 23.63007457762461% for the motion detector.

Keeping in mind also: the town does not know that a roleblocker guilty is a hard guilty until no killstop roles are claimed (because a doctor, jailkeeper, babysitter, rolestopper, bulletproof, commuter, and hider all give possible explanations for no kill).
The town does not know that a motion detect guilty is a guilty.

So between the three, if all three were left alive you might get a 92.60554355309359% chance of getting

one

guilty between them, but that's if all three were left alive for three nights straight and with the lens of the above that not all guilties are hard guilties. Only the follower is a guilty that cannot be argued against. (Counterbalanced by being Loud and thus letting the scum know precisely when they have been guiltied.)

The Neapolitan has different math, admittedly:
3/11 N1 (27.27272727272727%)
3/9 N2 (33.33333333333333%)
3/7 N3 (42.85714285714286%)

But again, the Neapolitan while guiltying scum is Loud and thus means the scum targeted by it know they have been guiltied and know they need to claim a power role. Any power role. Literally any power role--they can claim an unprovable role and the Neapolitan has no way of telling otherwise. And if the claim is appropriately low in power, it can fit into the setup fairly easily. And to reiterate, this is only IF the Neapolitan guilties scum.

And all of this math? Is in the worst case scenario where scum do not identify any of the threatening power roles to them.

The scum have a 4/9 shot N1 at hitting a power role (assuming a VT dead D1).
If they miss and a VT dies D2, then on N2 it's 4/7.
If they miss and a VT dies D3, then on N3 it's 4/5. (Granted, if there's literally only one VT left in the game, the scum have done something very, very, VERY wrong.)

This means that bussing is 100% off the table.

Is that a bad thing? Scum should not be rewarded by the setup for killing scum, they should be fully aware they should be punished for killing scum. Town should not be punished by the setup for killing scum; they should be justly rewarded for killing scum.

2) Loud follower does not have a counter.

The loud follower does not have a good counter if they get a guilty, that is true enough--however, as shown above, the statistical odds of the follower getting a guilty are quite low.

The counterplay to the loud follower is also plenty apparent BEFORE a nightkill--select someone to perform the nightkill you don't think the loud follower will target. You've got little in the way of counterplay after a guilty, but you've got plenty of counterplay before a guilty exists by guaranteeing that no guilty is generated.

Once down to the last goon follow the cop takes place. Roleblocker and follower work in tandem and scum can’t do shit.

Bluntly if the game gets down to one scum alive with both the roleblocker and follower alive, they deserve to lose and the town deserves to win.

The mafia are the informed minority--both of those are key. They are the minority, so statistically speaking it is hard for the town to get guilties on them especially given the scum are informed. By knowing what roles are in the setup and by knowing who the scum are, the scum have the best idea of who among them are least likely to be caught by 3/4 of the town power roles. Because as it turns out, follower/roleblocker/motion detector all have the same type of preferred target.

And as the informed minority they have the advantage of communicating with each other, collaborating, and coordinating. This can allow them to hone in on the power roles (with full knowledge of what they are, it is easier to pick up on identifying behaviors) and to coordinate who does the nightkill as to dodge the town PRs and to, in their dayplay, play in a way to help maximize their night action gameplay.

After all--you don't need to bus in order to look town as scum. Bussing is one way to look town as scum; it is not the

only

way to look town as scum. Plus you don't necessarily need to bus; distancing without bussing is still viable in this setup. If scum, with full knowledge of the setup, do nothing to capitalize on this knowledge and strategize, then they are neglecting their advantages as the informed minority by not taking advantage of the information given to them.

[MathBlade]

I owe a more detailed response to this but the gist is:

>> Percentages of individual events are correct. The odds calculator done after is not.

Even if I go with those numbers that’s 90% chance of a guilty. And sure it’s loud but you literally cannot have long term power balance play. Everything in the setup relies on having quick wit and reactions and getting through a moment versus being able to investigate and plan. Elos are won and lost from being able to have a narrative which you can’t build if you’re just surviving.

>> Scum should not be rewarded by the setup for killing scum. I agree. They should not be punished either. Scum being limited in their day play is OP as fuck. It says okay miselim 3 townies while having hands behind your back. It’s doable as I have done it but that doesn’t make it balanced at all.

[MathBlade]

Good scum play is about pulling each player’s strings at exactly the right time and exactly the right moment. Turning each person against another without them noticing. Taking away the ability to use half of those strings is broken. At some point it’s not even about the guilties it’s that you’ll be forced to take a PR to elo in this setup which is almost GG you lose.

Again loud imho is not a counter balancing mechanic.

Knowing your screwed and not able to do a thing about it and having to sacrifice them is not balancing.

[MathBlade]

Furthermore scum not finding any PR is viable in this setup as none of them have any reason to claim here until said guilty. With no investigative power scum not finding a PR until mass claim time is very probable.

I owe a more typed up response but even without bussing I do not see a path for scum to win this.

[MathBlade]

Oh man I'm due this now the other game is over

We'll go with the numbers you had for the roleblocker for N1.

Odds of Follower not getting a gpuilty is 10/11
Roleblocker not getting a guilty is 10/11
Motion Detector is 93% not getting a guilty (using your formula for simplicity)
Neopolitan has an 8/11 odds of not getting a guilty

So everyone not getting a guilty is 10/11 * 10/11 * 93/100 * 8/11 = 55%

Scum being guiltied outside of their control is 45%

Way too high without any sort of defense.
Each individual has a low chance, but scum are almost certainly guiltied more often than not.

[MathBlade]

N2 and N3 are slightly more complicated, but the informed of the whole setup just tells scum you're screwed you have to pretend to be a PR never lose and if you're checked expect to die.

Any scum claiming PR with an action will be elimmed within 3 days for not dying so therefore cannot end game.

So yeah this wouldn't pass review for me.