Probability and Statistics: Lums Undergraduate SS 11
Probability and Statistics: Lums Undergraduate SS 11
LUMS
Undergraduate
SS 11
Rules of Probability
• Conditional Probability: Probability of event A given that
event B has happened i.e. Pr ( A|B).
Pr ( A|B) = Pr (A ∩ B)/ Pr (B) (assuming P(B) > 0)
– If we have observed that event B has occurred, what is the probability
that A will also occur.
• Logic of this (fundamental) formula:
– Denominator: sample space now restricted to event B circle (since
we know event B has happened)
– Numerator: part of A that is in B.
• Example: Amongst the population of 16-21 year olds in the
USA the Pr (Unemployed) = 13.5%; Pr (Dropout) = 29.05% and
Pr (Unemployed and Dropout) = 5.32%. What is the
probability that a given person from this population is
unemployed given that you know he is a dropout?
Applying the Rules of Probability
1. Write down the probabilities you know in notational form:
– Pr (U) = 13.%; Pr (D)= 29.05% and Pr (U ∩ D) = 5.32%
2. Write down the probability you need to find: Pr (U|D)
– This is the more crucial step in conditional probability problems!
3. Invoke the appropriate rules of probability:
– Pr (U|D) = Pr (U ∩ D)/Pr (D) = 5.32%/29.05% = 18.3%
4. Note: knowing someone is a high-school drop out alters the
probability of him being unemployed!
– Conditional probabilities can tell us whether two events are related
(dependent) to one another or not (independent).
Rules of Probability
• The law of multiplication for dependent events states:
Pr (A ∩ B) = Pr ( A|B) . Pr (B)
– Joint probability is the product of conditional and marginal probability.
• Events A and B are said to be dependent if Pr (A) i.e., marginal
probability differs from Pr (A|B) i.e. conditional probability.
– Intuition: Knowing event B has occurred changes probability of event
A occurring e.g. , events dropout and unemployment were related to
one another (i.e. dependent) because being a dropout increases the
probability of being unemployed.
– Example: 18.7% of the members of US congress are senators and 49%
of senators are Democrats. What is the probability that a randomly
selected member of US Congress is a Democrat and a Senator?
Applying the Rules of Probability
• Write down the probabilities you know in notational form:
– D event member selected is a Democrat and S event member
selected is a senator. We know P(D|S) = 0.490 and P(S) = 0.187;
• Write down the probability you need to find: Pr (D ∩ S)
• Invoke the appropriate rules of probability and then apply it:
– Pr (A ∩ B) = Pr ( A|B) . Pr (B)
– Pr (D ∩ S) = P(D|S) . P(S) = 0.490 . 0.187 =0.092
• Interpret your answer:
– The probability that a randomly selected member of the US Congress
is a Democrat and a Senator is 0.092.
Rules of Probability
• Two events are independent if knowing that one event has
occurred does not effect the probability of the other event
Pr ( A|B) = Pr (A)
• If two events are independent then conditional probability
i.e. Pr ( A|B) is the same as marginal probability i.e. Pr (A)
• Example: Probability of getting a six on throw of die if I got a
six the last time I threw a dice.
• The law of multiplication for independent events states:
Pr (A ∩ B) = Pr ( A) . Pr (B)
– Joint probability is the product of marginal probabilities.
• Important application of this concept in gambler's fallacy
Applying the Rules of Probability
Question: Increasing the robustness of critical systems
within an organization’s operations is a major goal of supply
chain managers. To this end, internet giants like (Amazon,
Google etc.) maintain independent databases at different
locations around the world to minimize chances of
downtime in the face of natural disasters etc.
Probability that any given server at an international location
fails is 2% (once every 50 days). How many international
servers are needed to ensure that the probability of
downtime is approximately once every 125,000 days? Note:
Downtime happens when all servers fail at the same time.
Applying the Rules of Probability
• Write down information provided in probability notation:
– Pr (Server Fail on a given day) = 0.02
– Target Downtime Probability = 1/125,000
• Find probability of downtime i.e. all servers fail on same day?
– Applicable formula: Joint probability for independent events.
– 1/125,000 = 0.02 x 0.02 x 0.02 x …….. (hit and trial)
– 1/125,000 = (0.02)^N (where N is the no. of independent servers)
• Turns out having just 3 independent servers can make the
chances of downtime virtually zero!
• Example: What is the probability that your plan succeeds
when your plan depends on seven independent variables and
the probability of favorable outcome for each variable is 0.90?
Applying the Rules of Probability
Mutual Fund Mutual Fund does
Outperforms Market not Outperform
Index Market Index
Top-20 MBA 0.11 0.29