Calculating Probabilities in Mafia Party Game

  • Context: Undergrad 
  • Thread starter Thread starter Gerenuk
  • Start date Start date
  • Tags Tags
    Probabilities
Click For Summary
SUMMARY

The forum discussion focuses on calculating probabilities in the Mafia party game, specifically using a mathematical model to determine the likelihood of good players winning against bad players. The formula presented is P(g,b)=(g-1)/(g+b-1)*P(g-2,b)+b/(g+b-1)*P(g-1,b-1), with the conditions P(g<=b)=0 and P(b=0)=1. The author highlights the surprising non-monotonicity in the probabilities, noting instances where P(g=4,b=1) exceeds P(g=5,b=1), and questions the implications of player parity on game outcomes. A JavaScript implementation is provided for users to visualize the probabilities based on varying player counts.

PREREQUISITES
  • Understanding of probability theory and basic mathematical modeling
  • Familiarity with the Mafia party game mechanics
  • Basic knowledge of JavaScript for running the provided code
  • Experience with game theory concepts, particularly in strategic decision-making
NEXT STEPS
  • Explore advanced probability distributions relevant to game theory
  • Learn about Monte Carlo simulations for modeling complex game scenarios
  • Investigate the implications of player dynamics in multiplayer games
  • Study non-monotonic functions and their applications in strategic games
USEFUL FOR

This discussion is beneficial for game theorists, mathematicians, game developers, and enthusiasts of the Mafia party game seeking to deepen their understanding of probability calculations and strategic outcomes in multiplayer settings.

Gerenuk
Messages
1,027
Reaction score
5
Does anyone know the game
http://en.wikipedia.org/wiki/Mafia_(party_game)
?

I was trying to calculate probabilities with the assumptions that all players are dumb and slay a random person at day. Basically every night one of the good people dies and then a random one of the rest dies. So I wrote down (g=good, b=bad)
P(g,b)=(g-1)/(g+b-1)*P(g-2,b)+b/(g+b-1)*P(g-1,b-1)
P(g<=b)=0
P(b=0)=1
for the probability that the good people win.

Now I was surprised about the non-monotonicity. In my calculations it seems for example
P(g=4,b=1)>P(g=5,b=1)

Have I missed something out? Why does it depend on odd/even so much?
By this calculations werewolves should sometime be decreased if the number of players is increased?
 
Physics news on Phys.org
I don't understand what happens if it arrives at 1 of each on a day slot (either would have to be very dumb indeed to vote to chop himself), but assuming they draw lots the lack of monotonicity is real.

E.g. it may be surprising that starting with a day slot the goodies do better to have one goody v one baddy than two goodies v one baddy but if you think about it it's right.

I append a javascript. (To run paste into a .htm file and drop it into a browser. Set the range values to lowest and highest numbers of players for the two sides.) The first number shown is the goodies and the 'n' or 'd' represents night or day start.

<script>

range=[1,9]

function p(g,b,n){
if(g<0||b<0||(b==0&&g==0))return NaN
if(b==0)return 1
if(b>g)return 0
if(n)return p(g-1,b)
else return (g*p(g-1,b,1)+b*p(g,b-1,1))/(b+g)
}

for(k=0;k<2;k++)
for(i=range[0];i<=range[1];i++)
for(j=i;j<=range[1];j++)
document.writeln('('+j+','+i+','+(k?'n':'d')+') '+p(j,i,k).toFixed(3)+'<br>')

</script>
 
Last edited:

Similar threads

  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 14 ·
Replies
14
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 9 ·
Replies
9
Views
6K
  • · Replies 13 ·
Replies
13
Views
3K
  • · Replies 1 ·
Replies
1
Views
3K