On de Movire's solutions of the problem of duration of play 1708–1718

Summary I attempt to solve the enigma of how de Moivre derived his recursion formulae and the formulae for the probability of a duration of exactly n games. I believe that he obtained the latter probability by differencing Nicholas Bernoulli's formula for the probability of ruin and that he obtained the recursion formulae by a simple algebraic method and induction.