Here is the problem I am trying to figure out:

Someone starts with X amount of money that they can bet on this betting game. The game is the person can wager whatever he wants, and when he places the bet he has a 49.8% chance to double his money, and 50.2% chance to lose it. If this player is up Y amount, so his total wealth is X + Y, what is the total expected amount of money he must bet in order to get back to having a total of X money.

Heres an example of the question:

If I have 100 dollars, and I play this game enough times where I’m up to 150 dollars. How much total money must I wager, whether it is 5 dollars ten times or 50 dollars once, in order to expect to be back at 100 dollars


After betting $\$X$, you get back on average $\$X(0.996)$, because you double $0.498$ of the time. So after betting $\$250Y$, you get back on average $249Y$.
On the other hand, that doesn't answer your question. The time that your expected return is $-Y$, might not be the expected time that your return reaches $-Y$.


Suppose that M_n is the amount of money with start with before the n bet and that b_n is the proportional amount of that money that we bet on that specific bet. Our expected leftover money for that bet are:

$$ E(L_n) = M_n \cdot (1 - b_n) + 0.498 \cdot b_n \cdot M_n \cdot 2 = (1 - 0.004 \cdot b_n) \cdot M_n $$

Of course it's very easy to notice that our expected leftover money of bet n is equal to the expected value of money we start with at bet n + 1:

$$ E(L_n) = E(M_{n+1}) $$ $$ E(M_{n+1}) = (1 - 0.004 \cdot b_n) \cdot E(M_n) $$

Starting from the first bet we know:

$$ E(M_1) = X + Y$$ $$ E(M_2) = (1 - 0.004 \cdot b_1) \cdot (X + Y)$$

Right off the bat we can tell what's the amount we need to bet to return to X right away after the first bet.

$$ X = (1 - 0.004 \cdot b_1) \cdot (X + Y) $$ $$ b_1 = {{250 \cdot Y} \over {X + Y}} $$

Of course keep in mind that you should accept the b_n solution only if its equal or lower than 1. You cannot bet more money that you have of course. In your example with 100 + 50 dollars, b_1 would be 83,333... so that means it would be impossible to return to X in expected value after a single bet.

The procedure continues on with iterative substitutions...

$$ E(M_3) = (1 - 0.004 \cdot b_2) \cdot (1 - 0.004 \cdot b_1) \cdot (X + Y)$$ $$ E(M_4) = (1 - 0.004 \cdot b_3) \cdot (1 - 0.004 \cdot b_2) \cdot (1 - 0.004 \cdot b_1) \cdot (X + Y)$$

In the end it's about finding at least one solution of the following:

$$ X = (X + Y) \cdot \prod _{n=1}^{N} (1- 0.004\,b_{{n}}) \text { subject to } [0 \le b_n \le 1] \forall n$$

Hopefully that helped.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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