- experiment: a process whose outcome can’t be predicted in advance with certainty ( Randomness )
- sample space S: all possible outcomes; a coin is {H,T}, a die is {1,…,6}
- event: a subset of S; rolling a 1 is A={1}, an even roll is E={2,4,6}
- probability P(A): “the probability of event A occurring”, a number in [0,1]
- frequentist reading, the long run fraction of trials where A happens
P(A)=n→∞limnnA
- rules, all derived from the axioms
- 0≤P(A)≤1 and P(S)=1
- Complement: P(Ac)=1−P(A); S and ∅ are complements, so P(∅)=0
- disjoint A1,A2,… just add: P(⋃jAj)=∑jP(Aj)
- overlapping events need inclusion-exclusion: add the singles, subtract the pairs, add the triples, …
P(A∪B)=P(A)+P(B)−P(A∩B)
- three events: P(A∪B∪C)=P(A)+P(B)+P(C)−P(A∩B)−P(A∩C)−P(B∩C)+P(A∩B∩C)

- four events, written out

P(A∪B∪C∪D)=P(A)+P(B)+P(C)+P(D)−P(A∩B)−P(A∩C)−P(A∩D)−P(B∩C)−P(B∩D)−P(C∩D)+P(A∩B∩C)+P(A∩B∩D)+P(A∩C∩D)+P(B∩C∩D)−P(A∩B∩C∩D)
- mutually exclusive or not depends on the problem
- one card, A = red, B = face card: the 6 red face cards mean A∩B=∅, so not exclusive and P(A∪B)=5226+12−6=138
- a random song from 27,333; blues ( 330 ), jazz ( 537 ) and rock ( 8286 ) are disjoint genres, so P(blues, jazz or rock)=273339153≈0.335
- 20 five-star and 200 four-star songs out of 2000: P(5∗)=0.01, P(4∗)=0.10, P(5∗∪4∗)=0.11
- worked example: P(A)=.20, P(B)=.10, P(C)=.40, P(A∩B)=.05, P(A∩C)=.10, P(B∩C)=.03, P(A∩B∩C)=.01; probability none happen?
- none is (A∪B∪C)c by De Morgan’s Law, so take the complement of inclusion-exclusion
1−[.2+.1+.4−.05−.10−.03+.01]=1−.53=.47