# Recurrent random walk using binomial theorem

I wonder if someone can please check if I am on the right lines.

I want to show that a random walk with a probability of a right step of $$\frac{1}{4}$$ is recurrent. I think I might have done something wrong in my deductions.

I have that the probability of being at the origin at the nth step is

$$\left( \begin{array}{c} n\\ \frac{n}{2} \end{array} \right) \left(\frac{1}{4} \cdot \frac{3}{4}\right)^\frac{n}{2}$$

Then because this could be for any $$n \ge 1$$, I have decided that the probability of being at the origin at step $$n$$ for some $$n$$ could be

$$\sum_{n=1}^\infty \left( \begin{array}{c} n\\ \frac{n}{2} \end{array} \right) \left(\frac{1}{4} \cdot \frac{3}{4}\right)^\frac{n}{2}=\sum_{n=1}^\infty\left( \begin{array}{c} n\\ \frac{n}{2} \end{array} \right) \left(\frac{1}{4}\right)^{n-\frac{n}{2}} \left( \frac{3}{4}\right)^\frac{n}{2}$$

Is this correct? And if it is, this then looks similar to the binomial theorem for $$(\frac{1}{4}+\frac{3}{4})^n$$ which is of course 1, meaning the random walk would be recurrent. If this is what I need to be doing, can anyone guide me as to how I get this in the correct format to use the binomial theorem.

• Are you using the gamma function in this summation to define n choose n/2 ? – Peter Foreman Feb 2 at 16:44
• "this then looks similar to the binomial theorem for $(\frac{1}{4}+\frac{3}{4})^n$" This might "look similar" to your eyes but this is a completely different object, right? – Did Feb 2 at 16:53

Technical hint: $$\binom n {n/2} \sim c 2^n/\sqrt n$$. Which you can see by using Stirling's formula or by just knowing that about half of the mass in Pascal's triangle is within $$\pm \sqrt n$$ of the middle.
Conceptual hint: you are calculating the expected number of returns to $$0$$ ever. If it is finite, that means ...