## Wednesday, February 29, 2012

### The Game Of Life

Did I manage to drag your attention with the title? Good. Let's proceed to the lyrical part of this post then. Few years ago I watched this film and found the statistics starting at 1:16:56 being quite alarming. Not that I didn't know about most of those details, but it is totally different when you see them altogether. I encourage you to watch it as well. However, for the sake of this particular post, let's retain the following statistics: "20% of the world's population consumes 80% of its resources".

Now, let's proceed to the technical part of the post. My son started to learn Pascal at school and from time to time I give him "advanced" home assignments, to polish his programming (and thinking) technique. After few complaints from my son, aka "I am tired of doing nonsenses like calculate π, sorting etc. I want to do something more real" (well, he is just 12 years old), I decided to give him something "more real", a very basic Monte-Carlo simulation that I called "The Game Of Life". The reaction was quite unexpectedly expected, he got dragged by the title :)

So, what is the game about? Basically, let's imagine a world with limited resources (I don't think it is difficult to imagine this), which in the language of the game is called simply "wealth". We have a population of players with the total number of PLAYERS=100. Each player has two attributes, "wealth" and "power". The game starts with the equal distribution of wealth between the players and the "power=0" for each player. With each iteration, we can call this "year", each player plays against another randomly picked player (the players can't play with themselves). The rules for winning a play are quite simple:

• The player with greater "power" wins.
• If the players have the same power, the richest one wins.
• If the players have the same power and wealth, the winner is picked randomly.
• The winner gets a reward, which means their "power" increases by 1 and if the loser has sufficient "wealth", the winner's "wealth" increases by 1, the loser's wealth decreases by 1 respectively (so we keep the resources constant).
The game continues for 10000 iterations.

Here is the code (slightly adjusted with my help):


program montecarlosimulation;
uses wincrt;
const
PLAYERS = 100;
YEARS = 10000;

var
i,j:integer;
x,y:integer;
wealth:array[1..PLAYERS] of longint;
power :array[1..PLAYERS] of longint;
tmp:longint;
poor:integer;

// p1 is the winner!!!
procedure reward(p1, p2:integer);
begin
if wealth[p2] > 0 then begin
wealth[p1] := wealth[p1] + 1;
wealth[p2] := wealth[p2] - 1;
end;
power[p1] := power[p1] + 1;
end;

// decide who is the winner and apply the reward
function decision(a, b:integer):integer;
var
d:integer;
begin
// same power
if power[b] = power[a] then begin
// same wealth
if wealth[a] = wealth[b] then begin
// pick the winner randomly
d := random(2);
if d = 0 then reward(a,b)
else reward(b,a);
end
else begin
if wealth[b] > wealth[a] then reward(b,a)
else reward(a,b);
end;
end
else begin
if power[b] > power[a] then reward(b,a)
else reward(a,b);
end;
end;

begin
// initialise the players' values
// total wealth is PLAYERS*1000
for i := 1 to PLAYERS do begin
wealth[i] := 1000;
power[i] := 0;
end;

// the actual game
randomize;
for i := 1 to YEARS do begin
for x := 1 to PLAYERS do begin
y := random(PLAYERS) + 1;
// the actual play
if x <> y then decision(y, x);
end;
end;

// sorting the results, wealthiest on top
for i := 1 to PLAYERS do begin
for j:=i+1 to PLAYERS do begin
if wealth[i] < wealth[j] then begin
tmp := wealth[i];
wealth[i] := wealth[j];
wealth[j] := tmp;

tmp := power[i];
power[i] := power[j];
power[j] := tmp;
end;
end;
end;

// printing the results
poor := 0;
for i := 1 to PLAYERS do begin
if (wealth[i] = 0) then poor := poor + 1;
writeln('person[', i,'] has ', wealth[i],' wealth and ',power[i],' power');
end;

// printing statistics
writeln(poor, ' of ', PLAYERS, ' are poor.');
end.

This code was compiled using Free Pascal IDE.

Now, let's analyse the results. After executing this code several times, the number of wealth players (those with "wealth" > 0) is in between 20-23, from 100. This is ~20%. Setting YEARS = 15000 will decrease the result to 16%.But let's get back to the 20% figure, is this a coincidence with the 20% from the lyrical part on top of the post? The decision to set YEARS = 10000 was somewhat a random one, but later we (I and my son) figured out that ... check this out "Human prehistory begins in the Paleolithic Era, or "Early Stone Age". Later, during the Neolithic Era (New Stone Age), came the Agricultural Revolution(between 8000 and 5000 BCE) in the Fertile Crescent, where humans first began the systematic husbandry of plants and animals".

Couple of interesting facts as you can see. Obviously the code is a way too simple to pretend to model a real game of life, e.g. the number of players is constant, there is no chance for the poor players to ever win (which isn't the case in our reality), lots of rich people are also philanthropists, "brain" (and "faith") attribute isn't considered at all, two or more poor players with total "power" greater than the "power" of a rich player can compete against that player, an "insufficiently" rich player can bribe a richer and more powerful player to win the play with another player etc. This will obviously make the game more real and complicated (for my son, who still learns Pascal, while his father "plays" the game).

Anyway, if I managed to drag your attention to this subject, feel free to write a better version of "The Game Of Life" and drop me a note with your results. I am really curious.