Basic Probability
If we want to analyze the averagecase performance of an algorithm, we need to incorporate some basic facts from probability.
Sample space (S) : It is defined as the set of all possible outcomes from a certain experiment.
ExAMPLE # 1
Suppose 2 independent dice are rolled, there are 36 possible outcomes.
Event (E) : Every subset of S ( sample space) is an event.
EXAMPLE # 2:
Let D1 denote the no. that comes up on dice 1 when rolled.
Let D2 denote the no. that comes up on dice 2 when rolled.
Then E : D1 + D2 = 8 is an example of an event.
Probability of an event( E ) :It is the chance of occurrence of a particular event . It is denoted by Pr (E).
It can be measured as ( #outcomes where E occurs / #outcomes of the sample space)
Where # denotes “number of “
Ex :
Pr(E) for the event stated in the previous example is 5 / 36.
Pr(E) is a function whose value varies from 0 to 1.
If E is a subset of the sample space , then Pr(E) = ∑Pr (s) (for all s € E)
Probability of an event (E ) is the sum of the probabilities of all the elementary events belonging to the event.
The conditional probability that an event A occurs , given an event B, is denoted as
Pr ( A / B ) and is defined as
Pr ( A / B ) = Pr ( A ∩ B ) / Pr ( B )
assuming that Pr ( B ) > 0 .
Independence of events : Two events A and B are said to be independent events if the outcome of one does not effect the outcome of the other i.e.
Pr ( A / B ) = Pr ( A )
Random variable : Let S be a sample space , Pr denote the probability function associated with the sample space.
Random variable( X ) is defined as a function that maps outcomes from a sample space to real no.s .
X : S → R
Pr[ X = x ] = ∑ Pr ( w ) such that X (w) = x for all w € S
EXAMPLE # 3
Suppose two dice are rolled as stated in the Ex 2 and let the random variable
X = D1 + D2 , then X can take any value from 2 to 12.
For X = 8 , Pr ( X ) = 5/ 36.
For X = 2 , Pr ( X ) = 1 / 36.
So the probability of a random variable is not unique as in the case of an event. It takes different values depending on the value of X taken.
An indicator random variable( Z ) is a random variable that maps the outcomes to either 0 or 1 i.e.
Z = 1 if E ( event ) occurs
= 0 otherwise
EXAMPLE # 4
Consider the tossing of 3 coins
Z = 1 if all coins show heads up.
= 0 otherwise
Independence of Random Variables :
Let X and Y be two random variables
X : S → R and Y : S → R
Two Equivalent Definitions of Independence:

Pr[ X=x / Y=y ] = Pr[ X=x ]

Pr[ X=x ∩ Y=y] = Pr[ X=x ].Pr[ Y=y ]
EXAMPLE # 5
Let X = D1 + D2
And Y = 1 if D1 = 5
If X = 2 , then Pr[X] = 1/36
If Y = 1, then Pr[ X=2 / Y=1 ] = 0
So these two random variables X and Y are not independent.
EXAMPLE # 6
Let D1 be a random variable already defined in EXAMPLE # 2
Let I7 = 1 if D1 + D2 = 7
= 0 otherwise
Pr[ I7 = 1 ] = 1/6
Pr[ I7 = 1 / D1 = x ] = 1/6
Since these two values are equal, it is clear that D1 and I7 are independent random variables.
Expected Value of a Random Variable :
E[ X ] = ∑_{x€X} x.Pr[ X=x ]
Expectation of Random Variable is a weighted average of the probability of all its values.
E[D1] = 1(1/6) + 2(1/6) + 3(1/6) + 4(1/6) + 5(1/6)+ 6(1/6)
= 21/6
E[I7] = 1(1/6) + 0(5/6)
= 1/6
To find the expected value of the sum of 2 random variables X1 and X2
E[X1 + X2] = E[X1] + E[X2]
E[X + Y] = ∑_{e€S} Pr[e]( X[e] + Y[e] )
= ∑_{e€S} Pr[e] X[e] + ∑_{e€S} Pr[e]Y[e]
= E[X] + E[Y]
EXAMPLE # 7
X = D1 + D2
Y = 1 if D1 = 5
E[X] = 2(1/36) + 3(2/36) + 4(3/36) + 5(4/36) + 6(5/36) + 7(6/36) +8(5/36) + 9(4/36) + 10(3/36) + 11(2/36) + 12(1/36)
= 7
E[Y] = 1/6
E[X + Y] = 43/6
= E[X] + E[Y]
This shows that the proof stated above is valid for any 2 random variables irrespective of their independence or not. 