ΘρϵηΠατπ

sample space
A sample space Ω is a non-empty set
event
Given a sample space Ω, we say that any subset EΩ is an event
Probability Measure
Given a sample space Ω and a function P:Ω[0,1] , then we say that P is a probability measure if the following holds
  • EΩ,0P(E)1
  • P(Ω) =1
  • E1,E2,Ω such that E1,E2, are all pairwise disjoint P(i=1Ei)=i=1P(Ei)
Finite Pairwise Disjoint Probability
Suppose that A1,A2,,An are pairwise disjoint, then P(A1An)=i=1nP(Ai)
In the definition of a probability measure, for i{1,,n} we set Ei=Ai and then for any j>n we set Ej=, then we can see that E1,E2, is pairwise disjoint, and thus
P(A1A2An) = P(A1A2An)
= P(i=1Ei)
= i=1P(Ei)
= i=1nP(Ei)+P()+P()+P()+
= i=1nP(Ei)+0+0+0+
= i=1nP(Ai)
probability of a single point with equally likely outcomes is zero
Suppose that our sample space is [0,1] and there is some c[0,1] such that for any x[0,1], P({x})=c, then c=0
Suppose that c0, and note that P(Ω)=P([0,1])=x[0,1]P(x)=x[0,1]c, but note that no matter how small the value of c, since we are summing it uncountably many times with itself the sum always goes to infinity, but at the same time, P(Ω)=1 and thus we have a contradiction, so c=0.
Union Superset yields Sum Inequality
Suppose that (En) is a sequence of sets and that Ei=1, then P(E)i=1P(Ei)

Define Fi=EiE, we claim that E=i=1Fi. Let xE, then there is some i1 such that xEi, therefore xEiE:=Fi, on the contrary suppose that aFi for some i1, therefore aEiE, so clearly aE, therefore E=i=1Fi

Define G1:=F1 and set Gk:=Fk(n=1k1Fi)C for k2. Firstly note that the G's are pairwise disjoint, suppose that we have Gi,Gj where without loss of generatlity i<j, then we know that GiFi, but at the same time we know that Gj(n=1jF)C, since Fi is part of that union, then when we take it's complement, clearly Gj cannot intersect with Fi and since FiGi it cannot intersect with Gi as well.

Additionally we claim i=1Gi=Fi=1. Let xi=1Gi, so there exists some k1 such that xGk, since GkFk then we know xFk that is there is some k such that xFk so by definition xi=1Fi. Moving the other way suppose that pi=1Fi in that case there is some m1 so that pFm, let I1 be the collection of indices such that for every jI, pFj, clearly I is non-empty since at least kI, therefore by the well ordering principle it has some least element lI, note that this means for any c[1l1] we have that pFc, so by definition pGl, so that we've shown pi=1Gi as needed. We conclude that i=1Gi=i=1Fi=E

P(E)amp;=P(i=1Gi)amp;=i=1P(Gi)amp;i=1P(Fi)amp;i=1P(Ei) The first line is justified by the paragraph above, the second by the fact that the Gi's are disjoint, the third by the fact that GiFi and the last because FiEi.
Random Variable
Suppose that Ω is a sample space. A random variable is a function X:Ω
Probability of a Random Variable being an Element of a Set
Suppose that X is a random variable and E some event , then we define P(XE):=P({aΩ:X(a)E})
Probability of a Random Variable being an Element of a Set using Inverse Image
P(XE)=P(X1(E)) where we're using the inverse image of X
Probability of a Random Variable being equal to An Element
Let y, then we define P(X=y) as P(X{y})
conditional probability
Suppose that A,BΩ, with P(B)>0, then we define the conditional probability of A given B as
P(A|B)=P(AB)P(B)
Uniform Probability Measure
Let Ω be a finite sample space, then we define the uniform probabilty measure as a probability measure such that given any event A we have P(A)=|A||Ω|

Note the division in the above always works since Ω is assumed non-empty

Increasing Sequence of Sets
Let (An) be a sequence of sets, then we say that it's increasing when A0A1A2...
A Sequence of Sets Increase to Another
Suppose that (An) is an increasing sequence of sets, then we say it increases to A when n=1An=A, and write (An)A