[MAFIA] GAME OVER - FREQUENCY MAFIA

16396

09/25/2019 03:14 PM

I'll dig out the code, it might not be statistically better to lynch. Depends on the numbers of scum to players.

Just got back from work, my co-workers forgot they're meant to be British... normally no one will eat the last slice because they're too polite and then I can enjoy it when I take the leftovers home but not this time :(

The TM is for Totally Magical.

32388

09/25/2019 03:17 PM

Here's the post where he talks about the code. Note in this post that he thinks that it's better to nolynch in the game in question.

https://rpgmaker.net/forums/topics/24686/?post=887297#post887297

Then follows this post where he realizes that the equation actually favors lynching.

https://rpgmaker.net/forums/topics/24686/?post=887300#post887300

Finally, follows this post, where he reveals the full breakdown.

https://rpgmaker.net/forums/topics/24686/?post=887303#post887303

Cap_H

DIGITAL IDENTITY CRISIS

6625

09/25/2019 03:18 PM

author=OzzyTheOne
author=Cap_H
You're all mad, Cave is an obvious Lynch choice after making me look like a fool again.

#lynch Cave Dogge

But to comment on previous comments too. Piano seems little jumpy, but not in a bad way. It's like he's immediately pushing people to act.
Hexa's little iffy with no Lynch as this is far from his first game.
While I appreciate your willingess to read people this is literally the start of the game, it's hard to read anyone's actions as anything from any vote this early in the game.

I guess holes are your favorite part of cheese as this post couldn't possibly include less fat.

The TM is for Totally Magical.

32388

09/25/2019 03:21 PM

We have 13 players, so the following lines are the relevant ones.

Fomar
lllllllllllllllll| 1 scum | 2 scum | 3 scum | 4 scum | 5 scum |
n8 town | 59.46% | 21.58% | 16.44% | 3.18% | 2.89% |
n9 town | 50.66% | 35.16% | 10.58% | 9.04% | 1.47% |
10 town | 63.13% | 26.28% | 20.68% | 5.26% | 5.02% |
11 town | 54.78% | 39.63% | 13.84% | 12.21% | 2.67% |
12 town | 65.74% | 30.28% | 24.50% | 7.52% | 7.11% |

16396

09/25/2019 03:24 PM

author=pianotm
Then follows this post where he realizes that the equation actually favors lynching.

https://rpgmaker.net/forums/topics/24686/?post=887300#post887300

Yes Psy corrected me because I'd misunderstood a rule. We're thirteen players again, I scanned the first post but didn't see a number for how many scum there are but assuming it's three again then those statistics should still be accurate, though they assume our instincts are no worse than randomly lynching someone every day.

The TM is for Totally Magical.

32388

09/25/2019 03:31 PM

Cap_H
author=OzzyTheOne
author=Cap_H
You're all mad, Cave is an obvious Lynch choice after making me look like a fool again.

(Cap's CAVE_DOG vote)

But to comment on previous comments too. Piano seems little jumpy, but not in a bad way. It's like he's immediately pushing people to act.
Hexa's little iffy with no Lynch as this is far from his first game.
While I appreciate your willingess to read people this is literally the start of the game, it's hard to read anyone's actions as anything from any vote this early in the game.
I guess holes are your favorite part of cheese as this post couldn't possibly include less fat.

You know what, Ozzy's referring to the fact that you choose to take everyone's joke vote dead serious every. single. game. You get called out every game for it, and you usually ignore it. Now you're just mocking someone for calling you out on it and at this point, you've done it so consistently that I can only assume you're just doing it to piss us off. And you know what, since you've decided to mock someone who's being very reasonable about it, I figure that on day one, you're just as good for a serious lynch vote as anyone else.

#cancel
#Lynch Cap_H

Cap_H

DIGITAL IDENTITY CRISIS

6625

09/25/2019 03:54 PM

So, this is the first serious vote of the game. Dunno, I'm usually nicer as scum.

JESEUS MIMLLION SPOLERS

3702

09/25/2019 05:13 PM

Well, that's odd. When I run the code, I get a very different result.

Assuming 10 town and 3 scum, 10000 simulated games ran at this with day 1 no lynch:
Town Wins = 3416
Scum Wins = 6584

And with day 1 lynch, I got:
Town Wins = 2077
Scum Wins = 7953

town = 10
scum = 3
town_wins = 0
scum_wins = 0
day1_scum = 0
no_lynch = false
stop = false
days = {}
math.randomseed(os.time())


for i=1, 10000, 1 do
  tmp_town = town
  tmp_scum = scum
  day = 1
  while (true) do
    if not (day == 1 and no_lynch) then
      if math.random(tmp_scum + tmp_town) <= tmp_scum then  --i changed this to a <= because this random func returns 1 and the parameter inclusive
        tmp_scum = tmp_scum - 1
      else
        tmp_town = tmp_town - 1
      end
    end
    if tmp_scum == 0 then
      town_wins = town_wins + 1
      --days.push(day)
      break
    end
    if tmp_scum > 0 then
      tmp_town = tmp_town - 1
    end
    if tmp_scum > tmp_town then
      scum_wins = scum_wins + 1
      --days.push(day)
      break
    end
    day = day + 1
  end
end

function love.draw()
string = ""
love.graphics.setColor(.5,.5,.5)
love.graphics.rectangle("fill",0,0,1000,1000)
love.graphics.setColor(1,1,1)
if no_lynch then string = string .. "Not Lynching on day 1\n" else string = string .. "Lynching on day 1\n" end
string = string ..  "town:" .. town_wins .. "\n"
string = string ..  "scum:" .. scum_wins .. "\n"
love.graphics.print(string)
--string = string ..  "days:" .. (days.inject{ |sum, a| sum + a }.to_f / days.size).to_s
end

Did I mess this up in the conversion?

JESEUS MIMLLION SPOLERS

3702

09/25/2019 05:16 PM

Oooh waait he changed it after that comment from psy... one sec.

16396

09/25/2019 05:26 PM

Also bear in mind that hopefully our lynches will be better than random guesses, usually meta reads popup through the game and we have power roles.

author=Cap_H
So, this is the first serious vote of the game. Dunno, I'm usually nicer as scum.

And this is your reply to the first serious vote?

I'd say scum but to be honest this is exactly how Cap_H was last time (town).

JESEUS MIMLLION SPOLERS

3702

09/25/2019 05:28 PM

Huh, damn. Really does work out that way, when the scum win condition is >=.

lynch 2117 vs 1100

Math is neat!

ESBY

extreme disappointment

1238

09/25/2019 05:35 PM

can we start lynching the math nerds

psy_wombats

13315

09/25/2019 05:36 PM

It's statistically better and practically better because we're not selecting targets via rand() and need actual info/evidence to hunt effectively.

Also the simulators are pretty unhelpful given the closed setup. Speaking of which my frequency is 531khz and I'll be busy until tomorrow so expect me to be posting ~infrequently~ until then.

psy_wombats

13315

09/25/2019 05:38 PM

re: votes in thread so far: nothing that would make me suspect anyone. cap looks fine. piano's vote looks bunk but not scummy. esby has a wise pinky toe.

#lynch gourd_clae for dredging up this old debate which isn't actually a debate. I guess we're here because there's no setup info to speculate over

16396

09/25/2019 05:39 PM

author=psy_wombats
It's statistically better and practically better because we're not selecting targets via rand() and need actual info/evidence to hunt effectively.

We could use pop instead.

author=ESBY
can we start lynching the math nerds

Psy, Hexatona should we all gang up on Esby for not being a maths nerd?

16396

09/25/2019 05:41 PM

author=psy_wombats
#lynch gourd_clae for dredging up this old debate which isn't actually a debate. I guess we're here because there's no setup info to speculate over

Also for stealing Libby's day 1 play.

JESEUS MIMLLION SPOLERS

3702

09/25/2019 05:47 PM

Hah! Sure

#Lynch Esby

But if he says "I'm a little piggy, oink oink oink" I'll pick Cavedog instead

16396

09/25/2019 05:55 PM

#cancel

#Lynch Esby

#mathsnerdunite

author=Hexatona
But if he says "I'm a little piggy, oink oink oink" I'll pick Cavedog instead

Cave did take us for a ride last game, though he hasn't posted yet this game.

He's not alone though, Libby, Frogge.