The Probabilistic Method
The Notion of Probability
Let be a finite set of outcomes of some sequence of trials, so that all these outcomes are equally likely. Let . Then is called a sample space, and is called an event. The ratio
is called the probability of .
Proposition
Let be events from the same sample space. Then
Proof
We simply have to show that
This is true as the left-hand side counts each element of the sample space that is part of at least one of the exactly once, while the right-hand side counts each element of the sample space that is part of at least one of the at least once.
Non-constructive Proofs
Theorem
For all positive integers , the inequality holds.
Proof
Let , and let us color each edge of red or blue as follows. For each edge, flip a coin. If we get a head, we color that edge red, otherwise we color that edge blue. This way each edge will be red with probability one half, and blue with probability one half. We are going to show that the probability that we get no monochromatic -subgraphs in this way is more than zero. On the other hand, , the number of favorable outcomes divided by the number of all outcomes, where is the set of all possible 2-colorings of the edges of a complete graph on vertices. So implies that there is at least one favorable outcome, that is, there is at least one with 2-colored edges that does not contain any monochromatic -subgraphs.
Instead of proving that , we will prove that , which is an equivalent statement. Note that is the probability that we get at least one monochromatic subgraph in our randomly colored graph . The number of ways to 2-color the edges of a given -subgraph of is clearly as there are two choices for the color of each edge. Out of all these colorings, only two will be monochromatic, one with all edges red, and one with all edges blue. Therefore, the probability that a randomly chosen -subgraph is monochromatic is
The graph has subgraphs that are isomorphic to . Each of them has the same chance to be monochromatic. On the other hand, the probability that at least one of them is monochromatic is at most the sum of these individual probabilities, by the proposition In other words, if denotes the event that the -subgraph of has monochromatic edges, then
where ranges through all -subgraphs of . Now let us assume, in accordance with our criterion, that . Then the last term can be bounded as follows.
for all . The last inequality is very easy to prove, for example by induction.
Conditional Probability
If and are two events from the same sample space , and , then and are called independent events. Otherwise they are called dependent.
Let and be events from the same sample space, and assume . Let
Then is called a conditional probability, and is read "the probability of given ".
That is, is the probability of given that occurs. The following proposition is now immediate from the definitions.
Proposition
The events and are independent if and only if holds
Bayes' Theorem
Let and be mutually exclusive events so that . Let be any event. Then
Proof
As and are mutually exclusive, and are disjoint, and since , their union is exactly . Therefore,
and the proof follows as the first (resp. second) member of the right-hand side agrees with the first (resp. second) member of the right-hand side of the Bayes' Theorem.