gambler's ruin: one construct, many systems

Gambler's ruin is the classic problem of a player who stakes a fixed amount on each of a sequence of fair or unfair bets and asks the probability of losing the whole bankroll before reaching some target. Posed by Huygens and Pascal and solved in general form in De Moivre's The Doctrine of Chances, it is modelled as a random walk between two absorbing barriers, with the striking result that a gambler with finite capital playing against a much richer opponent is ruined with probability approaching one, even in a fair game. The referent is the first-passage of a bounded random walk to an absorbing boundary, quantified by ruin probability and expected duration, the same structure that governs allele fixation and insurer insolvency.[1]

Unit

The probability of reaching zero capital before a target, and the expected number of bets

Recovery means

Not applicable

Measured by

Computed as the ruin probability and the expected duration of play for a random walk between absorbing barriers

In population genetics the same construct governs the fate of an allele under random genetic drift: in a finite population the count of copies of an allele performs a random walk whose only resting states are zero (loss) and the population size (fixation), both absorbing boundaries. Moran's 1958 Random processes in genetics formalized this as a Markov chain, giving fixation probabilities and expected times to absorption that are the biological counterparts of ruin probability and duration of play. It is the identical absorbing-boundary first-passage problem as gambler's ruin, here realized in allele frequencies and measured by fixation probability and time to fixation or loss.[2]

Unit

The probability an allele reaches fixation, and the expected time to fixation or loss

Recovery means

Not applicable

Measured by

Derived as fixation probabilities and mean absorption times from the Markov chain of allele counts

In actuarial science, ruin theory asks the probability that an insurer's surplus (premiums coming in, claims going out) ever falls below zero, which is the company's analogue of a gambler losing the bankroll. Dickson's Insurance Risk and Ruin develops the surplus process and the probability of ultimate ruin, treating the reserve as a stochastic path that may or may not cross the absorbing level of insolvency. It is the same first-passage-to-an-absorbing-boundary construct as in gambling and in allele fixation, transposed to an insurer's capital and measured by the probability and timing of ruin given premium rate, claim distribution, and starting surplus.[3]

Unit

The probability the surplus process ever falls below zero, given starting reserve

Recovery means

Not applicable

Measured by

Assessed as the probability of ultimate ruin as a function of initial surplus, premium rate, and claim distribution