1. EVENTS AND THEIR PROBABILITIES 1.1 Events Example of an experiment: Toss a coin and observe up face. Possible outcomes of this experiment: H (head), T (tail) Sample space (set of all possible outcomes): {H,T} Events (subsets of the sample space): {}, {H}, {T}, {H,T} The event {H} occurs, if a head is observed. The event {T} occurs, if a tail is observed. The event {H,T} occurs, if a head or a tail is observed. The event {} cannot occur. The probability that {H} occurs is 12 . The probability that {T} occurs is 12 . The probability that {H,T} occurs is 1. The probability that {} occurs is 0. In abbreviated form we may write P({H})= 12 , P({T})= 12 , P({H,T})=1, and P({})=0 Since the sample space (usually denoted by Ω) contains all possible outcomes and therefore occurs with probability 1, it is also called the certain event. Since the empty set (usually denoted by ∅) contains no outcomes and therefore occurs with probability 0, it is also called the impossible event. 1.1 1.2 Complementary events The complement of an event A (denoted by Ac) contains all possible outcomes that are not in A. Two events A and B are called complementary if Ac=B. The complements of the four events in the coin-tossing experiment are given by ∅c=Ω, {H}c={T}, {T}c={H}, and Ωc=∅. We observe that and P(∅)+P(∅c)=P(∅)+P(Ω)=0+1=1, P({H})+P({H}c)=P({H})+P({T})= 12 + 12 =1, P({T})+P({T}c)=P({T})+P({H})= 12 + 12 =1, P(Ω)+P(Ωc)=P(Ω)+P(∅)=1+0=1. In general, we have P(A)+P(Ac)=1 or, equivalently, P(Ac)=1-P(A). • Exercise 1.a: Suppose that a die is tossed. (i) List all subsets of Ω={1,2,3,4,5,6} that contain two outcomes. (ii) Find the complements of the events {4} and {1,3,6}. (iii) What is the probability of {4}c given that P({4})= 16 ? 1.2 1.3 Unions and intersections of events The union of two events A and B (denoted by A∪B) occurs if either A occurs or B occurs or both occur. The intersection of two events A and B (denoted by A∩B) occurs if both A and B occur. As an illustration, we form the unions and intersections of some subsets of the probability space Ω={1,2,3,4,5,6}. {1,2}∪{1,3,5}={1,2,3,5} {1,3}∪{1,3,5}={1,3,5} {2,4}∪{1,3,5}={1,2,3,4,5} The union of {1,2} and {1,3,5} occurs if {1,2,3,5} occurs. The union of {1,3} and {1,3,5} occurs if {1,3,5} occurs. The union of {2,4} and {1,3,5} occurs if {1,2,3,4,5} occurs. {1,2}∩{1,3,5}={1} {1,3}∩{1,3,5}={1,3} {2,4}∩{1,3,5}=∅ The intersection of {1,2} and {1,3,5} occurs if {1} occurs. The intersection of {1,3} and {1,3,5} occurs if {1,3} occurs. The intersection of {2,4} and {1,3,5} never occurs. 1.3 1.4 Mutually exclusive events Two events A and B are called mutually exclusive (or disjoint) if A∩B=∅. In the coin-tossing experiment we have and P({H}∪{T})=P(Ω)=1 P({H})+P({T})= 12 + 12 =1, but at the same time we have also and P({H}∪Ω)=P(Ω)=1 P({H})+P(Ω)= 12 +1= 32 . In the latter case, the probabilty of the union is not equal to the sum of the probabilities because the two events {H} and Ω are not disjoint. The outcome H is contained in both events and is therefore counted twice. General rules: P(A1∪A2)=P(A1)+P(A2) if A1∩A2=∅ P(A1∪A2∪A3)=P(A1)+P(A2)+P(A3) if A1∩A2=A1∩A3=A2∩A3=∅ P(A1∪A2∪…)=P(A1)+P(A2)+… if Ai∩Aj=∅ ∀i≠j 1.4 1.5 Summing simple event propabilities A simple event contains exactly one outcome. In the die-tossing experiment the simple events are {1},{2}, {3},{4},{5}, and {6}. In this experiment, it is natural to assume that all simple event probabilities are equal, i.e., P({1})=P({2})=P({3})=P({4})=P({5})=P({6})= 16 . Of course, two different simple events are always disjoint. To find the probability of the event A={2,5} we can therefore simply add the probabilities P({2}) and P({5}) of the simple events {2} and {5}, i.e., {2}∩{5}=∅ ⇒ P({2,5})=P({2})+P({5})= 16 + 16 = 62 = 13 . • Exercise 1.b: What is the probability that an even number is observed? {2}∩{4}={2}∩{6}={4}∩{6}=∅ ⇒ P({2,4,6})=P({2})+P({4})+P({6})= 16 + 16 + 16 = 36 = 12 . • Exercise 1.c: What is the probability that an odd number is observed? {1}∩{3}={1}∩{5}={3}∩{5}=∅ ⇒ P({1,3,5})=P({1})+P({3})+P({5})= 16 + 16 + 16 = 36 = 12 . Alternatively, we can find the solution with the help of 1.b. P({1,3,5})=P({2,4,6}c)=1-P({2,4,6})=1- 12 = 12 . 1.5 1.6 Additive rule of probability Suppose that a die is tossed. What is the probability of observing a prime number or an even number? P({2,3,5}∪{2,4,6})=P({2,3,4,5,6})= 56 In contrast, P({2,3,5})+P({2,4,6})= 36 + 36 =1, because 2 is contained both in {2,3,5} and in {2,4,6} and is therefore counted twice. Equality can be obtained by subtracting P({2}) from the sum of the probabilties P({2,3,5}) and P({2,4,6}), i.e., P({2,3,5}∪{2,4,6})=P({2,3,5})+P({2,4,6})-P({2}). In general, the sum of the probabilities differs from the probability of the union just by the probabilty of the intersection. Additive rule of probability: P(A1∪A2)=P(A1)+P(A2)-P(A1∩A2) 1.6 1.7 Properties of conditional probability Suppose that after having carried out an experiment we know that event B has occurred. We can then conclude firstly that the actual outcome of the experiment cannot be contained in any event that is disjoint to B and secondly that it must be contained in every superset of B. Denoting the conditional probability of event A, given that event B has occurred, by P(AB) we can write and (i) P(AB)=0 if A∩B=∅ (ii) P(AB)=1 if A⊇B. Once we know that the actual outcome is contained in B, the conditional probability of any other event A depends exclusively on those of its outcomes that are contained in A∩B. Clearly, none of the outcomes contained in A-B can possibly be the actual outcome. Accordingly, we have (iii) P(AB)=P(A∩BB). Finally, suppose that A⊆B. Then B can be divided into the two disjoint subsets A and B-A. Since the knowledge that B has occurred cannot change the odds between A and B-A, we also have (iv) P(AB):P(B-AB)=P(A):P(B-A) if A⊆B. 1.7 1.8 Formal definition of conditional probability Apart from the special properties and (i) (ii) (iii) (iv) P(AB)=0 if A∩B=∅, P(AB)=1 if A⊇B, P(AB)=P(A∩BB), P(AB):P(B-AB)=P(A):P(B-A) if A⊆B, the conditional probability must also meet the ordinary standards of probability. In particular, since the sets A∩B, B-A, A-B, and (A∪B)C are pairwise disjoint, we have P(A∩BB)+P(B-AB)+P(A-BB)+P((A∪B)CB) = P((A∩B)∪(B-A)∪(A-B)∪(A∪B)CB)=P(ΩB)=1. It follows from (i) that P(A-BB)=0, P((A∪B)CB)=0, hence P(A∩BB)+P(B-AB)=1. Dividing both sides of this equation by P(A∩BB), using B-A=B-A∩B, and applying (iv) gives P ( A ∩ B B) P ( B − A B) P(B− A ∩B B) P(B− A ∩B) 1 = + =1+ =1+ P(A ∩B) P ( A ∩ B B) P ( A ∩ B B) P ( A ∩ B B) P(A∩ B B) P ( A ∩B) P(B− A ∩B) P(A∩B)+ P(B−A ∩B) P ( B) = + = = . P(A∩B) P ( A ∩ B) P(A ∩B) P ( A ∩ B) Finally, using (iii) we obtain a universal formula for conditional probability. P(A∩B) P(AB)=P(A∩BB)= P(B) 1.8 1.9 Conditional versus unconditional probabilities • Exercise 1.d: Suppose that a die is tossed. What is the conditional probability of observing a prime number given that an even number is observed? P({2,3,5}{2,4,6})=P({2,3,5}∩{2,4,6})/P({2,4,6}) =P({2})/P({2,4,6})= 16 / 63 = 13 In contrast, the unconditional probability is given by P({2,3,5})= 63 = 12 . Obviously, the occurrence of {2,4,6} alters the probability that {2,3,5} has occurred! • Exercise 1.e: Suppose that a die is tossed. What is the conditional probability of observing a number less than 3 given that an even number is observed? P({1,2}{2,4,6})=P({1,2}∩{2,4,6})/P({2,4,6}) =P({2})/P({2,4,6})= 16 / 63 = 13 In this case, the conditional probability equals the unconditional probability, because P({1,2})= 62 = 13 . 1.9 1.10 Independent events In general, the conditional probability P(AB) equals the unconditional probability P(A) only if P(A∩B)=P(A)P(B). In this case, the probabilty P(B) is canceled out: P(AB)= P(A ∩ B) P(A)P(B) = =P(A) P(B) P(B) Two events A and B are said to be independent if P(A∩B)=P(A)P(B). Note: In contrast to the definition of P(AB), where it is required that P(B)≠0, the definition of independence is always meaningful. • Exercise 1.f: Suppose a die is tossed. (i) Are the events {1,3,5} and {2,4,6} independent? No, because whereas P({1,3,5}∩{2,4,6})=P(∅)=0, P({1,3,5})P({2,4,6})= 12 12 = 14 . (ii) Are the events {1,3,5} and {1,2} independent? Yes, because and also P({1,3,5}∩{1,2})=P({1})= 16 P{1,3,5})P({1,2})= 12 13 = 16 . 1.10 1.11 Further exercises • Exercise 1.g: Suppose a die is tossed. (i) What is the probability of observing a six or an odd number? P({6}∪{1,3,5})=P({1,3,5,6}= 64 = 32 (ii) Find P({2,4,6}{1,3,5}), P({3,5}{1,3,5}), P({2,3,5}{1,3,5}), and P({2,3,4,6}{1,3,5}). P({2,4,6}{1,3,5})=P({2,4,6}∩{1,3,5})/P({1,3,5}) =P(∅)/P({1,3,5})=0/ 63 =0 P({3,5}{1,3,5})=P({3,5}∩{1,3,5})/P({1,3,5}) =P({3,5})/P({1,3,5})= 62 / 63 = 32 P({2,3,5}{1,3,5})=P({2,3,5}∩{1,3,5})/P({1,3,5}) =P({3,5})/P({1,3,5})= 62 / 63 = 32 P({1,2,3,4}{2,4})=P({1,2,3,4}∩{2,4})/P({2,4}) =P({2,4})/P({2,4})=1 (iii) Are the events {2,3,6} and {1,2,3,4} independent? Yes, because P({2,3,6}∩{1,2,3,4})=P({2,3})= 62 = 13 and P({2,3,6})P({1,2,3,4})= 36 64 = 13 . 1.11 • Exercise 1.h: Suppose a coin is tossed twice and the sequence of heads and tails is recorded. (i) List the sample space. Ω={(H,H),(H,T),(T,H),(T,T)} (ii) Find all simple event probabilities. To calculate P((H,H)) we assume that the events H1={1st toss is a head}={(H,H),(H,T)} and H2={2nd toss is a head}={(H,H),(T,H)} are independent and that P(H1)=P(H2)= 12 . Then P({(H,H)})=P({(H,H),(H,T)}∩{(H,H),(T,H)}) =P(H1∩H2) =P(H1)P(H2) = 12 12 = 14 . Analogously, P({(H,T)})=P({(T,H)})=P({(T,T)})= 14 . (iii) Find P({(H,H),(H,T)}{(H,H),(T,T)}). P({(H,H),(H,T)}{(H,H),(T,T)}) =P({(H,H),(H,T)}∩{(H,H),(T,T)})/P({(H,H),(T,T)}) =P({(H,H)})/P({(H,H),(T,T)}) = 14 / 24 = 12 1.12 • Exercise 1.i: Discuss the four properties and (i) (ii) (iii) (iv) P(AB)=0 if A∩B=∅, P(AB)=1 if A⊇B, P(AB)=P(A∩BB), P(AB):P(B-AB)=P(A):P(B-A) if A⊆B using the die-tossing experiment and A={4,6}, B={1,3,5} for (i), A={1,2,3,4}, B={1,2} for (ii), A={2,3,4,5}, B={2,4,6} for (iii), and A={6}, B={2,4,6} for (iv). 1.13