Math 2.0

problem

Gambler's ruin

Problem

A gambler starts with kk units of money, where 0≤k≤N0 \le k \le N for an integer N≥1N \ge 1. In each round, the gambler wins one unit with probability p∈(0,1)p \in (0, 1) and loses one unit with probability q=1−pq = 1 - p. The rounds are independent. The game stops when the fortune reaches 00 (ruin) or NN (the target). Find the probability PkP_k that the gambler reaches NN before 00.

Remarks

The fortune is a random walk on {0,1,…,N}\{0, 1, \dots, N\} that stops at 00 and NN. By the law of total probability for the first round, Pk=pPk+1+qPk−1P_k = p P_{k+1} + q P_{k-1} for 0<k<N0 < k < N, with P0=0P_0 = 0 and PN=1P_N = 1. For a fair game, p=q=1/2p = q = 1/2, the solution is Pk=k/NP_k = k/N. For p≠qp \ne q, the solution is Pk=1−(q/p)k1−(q/p)NP_k = \dfrac{1 - (q/p)^k}{1 - (q/p)^N}. In a fair game against an opponent with unlimited money (N→∞N \to \infty), the gambler is ruined with probability 11.

History

An early form of the problem appears in a letter of Blaise Pascal to Pierre de Fermat in 1656.

Explanations

L1 human

A gambler who plays a fair game against a much richer opponent will almost surely lose everything in the end. With 1010 units and a target of 100100 in a fair game, the chance to reach the target is only 10/100=0.110/100 = 0.1.

What this node needs for W2
  • edge bounds both (0 of 1)

The writing levels say how complete the text of a node is. They do not measure its truth.

Sources

  1. Web https://en.wikipedia.org/wiki/Gambler's_ruin, Fair coin flipping
  2. Web https://en.wikipedia.org/wiki/Gambler's_ruin, Unfair coin flipping