So here's the question I'm trying to solve:

A stock price movement model supposes that if the current stock price is s, then, after one period, the stock price will be $us$ with probability $p$ and $ds$ with probability $1 - p$. Assuming that successive movements of the stock are independent, approximate the probability that the stock’s price will be up at least $50\%$ after the next $1000$ periods if $u = 1.1$, $d = 0.95$, and $p = 0.49$. What is the minimal value of $p$ to guarantee that the stock price will be over $60\%$ up after $1000$ periods?

And I don't even know where to start. Any suggestions as to how I should approach this problem?

thank you.

  • $\begingroup$ I'm guessing that you mean the colon to be a decimal point. $\endgroup$ Apr 30, 2012 at 1:52
  • $\begingroup$ It's not at all clear why a uniform distribution is mentioned in this question. $\endgroup$ Apr 30, 2012 at 2:09

1 Answer 1


You're multiplying 1000 times by something that's equal to either $u$ or $d$ each time. That means you're adding something equal to $\log u$ or $\log d$ 1000 times. The central limit theorem applies to independent random variables that are added; that's why we're taking logarithms.

We have $$ \mathbb{E}\left.\begin{cases} \log 1.1 & \text{with probability } 0.49 \\ \log 0.95 & \text{with probability } 0.51 \end{cases}\right\} = 0.49\log 1.1 + 0.51\log 0.95, $$ and $$ \operatorname{var}\left.\begin{cases} \log 1.1 & \text{with probability } 0.49 \\ \log 0.95 & \text{with probability } 0.51 \end{cases}\right\} = (0.49)(0.51)(\log 1.1 - \log 0.95)^2. $$ Hence $$ \frac{\left.\begin{cases} \log 1.1 & \text{with probability } 0.49 \\ \log 0.95 & \text{with probability } 0.51 \end{cases}\right\} - (0.49 \log1.1+0.51\log 0.95)}{\sqrt{(0.49)(.051)(\log1.1-\log0.95)}} $$ has expected value $0$ and standard deviation $1$.

Adding up $1000$ independent copies of this, we get $$ \frac{\text{sum}-1000(0.49\log1.1+0.51\log0.95)}{\sqrt{1000(0.49(0.51)(\log1.1-\log0.95)}} $$ has an approximately standard normal distribution. The problem is: what is the probability that this is more than $$ \frac{\log 1.5 - 1000(0.49\log1.1+0.51\log0.95)}{\sqrt{1000(0.49(0.51)(\log1.1-\log0.95)}}. $$ (I.e. going up at least $50\%$ is the same as multiplying by $1.5$ or more.)

So that last item is what you put into the table of values of the c.d.f. of the standard normal distribution.


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.