Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Disclaimer, I'm a computer person, not a math person, so surely I'm not going to use the proper math lingo. I apologize in advance. Feel free to edit to make this question more readable to others. This is a 2 part question:

  1. What is the probability of achieving a goal that has a probability of $1/n$, with $n$ tries, as $n$ goes to infinity?

  2. The actual question I'm trying to solve is related to a hypothetical poker game, and I believe its answer hinges on the answer to question 1 above. The game is as follows: Player 1 $(P_1)$ begins with a score of $x$, Player 2 $(P_2)$ begins with a score of $N*x$. Both players "go all in" on the flip of a coin comparison, and wager the amount of the lower player's score. If $P_1$ gets down to 0, he starts with $x$ again. We continue until $P_2$ has 0. What is the expected number of times this game can be played until $P_2$ hits 0, in terms of $x$ and $N$?

I wrote a simple computer program to illustrate the game, and played it with x = 1 and N = 100. After 20 or so tries, here was one result that was short enough to paste here:

Starting Position - $P_1:1, P_2:100$

Hand Count: 1 $P_1:2, P_2:99$

Hand Count: 2 $P_1:4, P_2:97$

Hand Count: 3 $P_1:0, P_2:101$

Hand Count: 4 $P_1:0, P_2:102$

Hand Count: 5 $P_1:2, P_2:101$

Hand Count: 6 $P_1:4, P_2:99$

Hand Count: 7 $P_1:0, P_2:103$

Hand Count: 8 $P_1:2, P_2:102$

Hand Count: 9 $P_1:4, P_2:100$

Hand Count: 10 $P_1:0, P_2:104$

Hand Count: 11 $P_1:2, P_2:103$

Hand Count: 12 $P_1:4, P_2:101$

Hand Count: 13 $P_1:8, P_2:97$

Hand Count: 14 $P_1:16, P_2:89$

Hand Count: 15 $P_1:32, P_2:73$

Hand Count: 16 $P_1:64, P_2:41$

Hand Count: 17 $P_1:105, P_2:0$


$P_1$ stated with 1. $P_1$ Ended with 105.

$P_2$ started with 100. $P_2$ Ended with 0.

Total Hands played until $P_1$ busts $P_2$: 17

Total amount $P_1$ invested: 5

Notes: Hand #4 shows the result after Player 1 performed a "rebuy" of 1 and immediately lost. Hand #5 show the result after $P_1$ performed a rebuy of 1 and won, etc. In order for Player 1 to ever win this game, he must win 7 in a row, and I'm thinking he has about 128 tries to do it before he then has to hit 8 in a row, then 256 tries to achive the $1/256$ before getting 512 tries at $1/512$, etc.

share|cite|improve this question

The probability of not succeeding in question 1 is $\left(1-\frac1n\right)^n$ and $$\lim_{n\to\infty}\left(1-\frac1n\right)^n=e^{-1}$$ is "well-known", thus giving a limit probability of $\approx 63.212\,\%$.

share|cite|improve this answer
Thanks! I edited my question and removed my guess, which was pointless for me to include to begin with. – TTT Apr 13 '13 at 17:16
Do you think the answer to question 2 is related to Euler's number, or should I strip that assumption as well? – TTT Apr 13 '13 at 17:28

The value of $x$ will have no effect on the probability (since P2's money is just a multiple and the rebuy is $x$ again), but $N$ obviously will.

Let's just set $x=1$ and $X_n$ as the number of times P1 lost by the $n$th try, so $X_n$~$Bin(n,\frac{1}{2})$. We then have $N+1+X_n$ money in the game. To win, P1 will have to get a winning streak of lenght $log_2(N+1+X_n)-1$ (base $2$ because he doubles every time, $-1$ because the oppenent loses that amount, so he has to win one game less).

So let's say $Y_n=1$ iff the $n$th try starts a winning streak and $0$ elsewise. So $P(Y_n=1)$ is $\frac{1}{2}(\frac{1}{2})^{log_2(N+1+X_n)-1} =\frac{1}{N+1+X_n}$ under the condition that P1 did not win beforehand.

It gets ugly from then on, so we will do some approximations. You can start with approximating $X_n$ by its expected value of $\frac{n}{2}$. Then $P(Y_n=1)=[\prod_{j=1}^{n-1}(1-\frac{1}{N+1+\frac{j}{2}})](\frac{1}{N+1+\frac{n}{2}})$ and thus, roughly, $$E[M]=\sum_{n=1}^{\infty}nP(Y_n=1)=\sum_{n=1}^{\infty}n[\prod_{j=1}^{n-1}(1-\frac{1}{N+1+\frac{j}{2}})](\frac{1}{N+1+\frac{n}{2}})$$ for the expected number of tries $M$ until P1 starts a winning streak.

This is just a quick glance at the problem with some possible errors and the approximations may cause problems, but I hope it goes into the right direction.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.