Calculating Chances

Once you start working with probabilities, you quickly realize that the assumption of all possible outcomes being equally likely isn’t always reasonable. For example, if you think a coin is biased then you won’t want to assume that it lands heads with the same chance as tails.

To deal with settings in which some outcomes have a higher chance than others, a more general theory is needed. In the 1930’s, the Russian mathematician Andrey Kolmogorov (1903-1987) formulated some ground rules, known as axioms, that covered a rich array of settings and became the foundation of modern probability theory.

The axioms start out with an outcome space $\Omega$. We will assume $\Omega$ to be finite for now. Probability is a function $P$ defined on events, which as you know are subsets of $\Omega$. The first two axioms just set the scale of measurement: they define probabilites to be numbers between 0 and 1.

• Probabilities are non-negative: for each event $A$, $P(A) \ge 0$.
• The probability of the whole space is 1: $P(\Omega ) = 1$.

The third and final axiom is the key to probability being a “measure” of an event. We will study it after we have developed some relevant terminology.