Experiment. An experiment is any process whose result is well defined. It is deterministic when the outcome can be predicted with certainty under ideal conditions (e.g. genetic determination, hitting a fixed target under fixed conditions). It is a random (non-deterministic) experiment when (i) every possible outcome is known in advance, (ii) no individual outcome can be predicted in advance, and (iii) it can be repeated under identical conditions -- rolling a die and tossing a coin are the standing examples.
Sample space and elementary events. The sample space S of a random experiment is the set of ALL its possible outcomes; each point of S is called a simple event, elementary event, or sample point, and cannot be broken down any further. S can be:
- finite -- S={1,2,3,4,5,6} for a die, S={H,T} for a coin;
- countably infinite -- tossing a coin until the first head gives S={H,TH,TTH,TTTH,…}, which can be listed and put in one-to-one correspondence with N;
- uncountably infinite -- choosing a real number x with 0<x<1 gives S={x:0<x<1}, which cannot be listed at all.
This chapter, from here on, restricts to a finite sample space.
Events. When S is finite, ANY subset of S is an event -- formally, every element of the power set P(S) is an event. Named types:
- Sure (certain) event -- S itself (it always happens).
- Impossible event -- the empty set ∅ (it never happens).
- Complementary event Aˉ (also written Ac or A′) -- 'not A', the event that A does not occur.
- Mutually exclusive events -- A1,…,Ak with Ai∩Aj=∅ for every i=j: they cannot occur simultaneously. …