Pascal and Fermat discuss probability

In 1654, Pascal and Fermat exchanged letters about games of chance. One question asked how to divide stakes when a match stops before either player reaches the agreed number of wins. Simply dividing by the present score does not always reflect each player’s chance of eventual victory. Earlier writers had discussed the problem, and French conversations about dice and gambling brought it to the two mathematicians. Future results had to enter a present calculation.

References: [1]

Fermat enumerated possible sequences of wins and losses and compared the outcomes. Pascal could work backward from the two possible results of the next round. Both approaches turned an uncertain future into a finite set of cases. With specified conditions, the players’ expected shares could be calculated. Their correspondence also showed how different methods could arrive at the same result and provide an opportunity for mutual checking.

References: [1]

Huygens and other authors subsequently developed these questions in print. Insurance, annuities, and population records supplied additional problems, while expectation, distributions, and laws of large numbers acquired fuller treatment. Games offered models with relatively simple conditions in which possible outcomes could be handled numerically. Applications to society required attention to data, samples, and independence. Probability gradually became part of statistical practice and decisions made under uncertainty.

References: [1]