2
$\begingroup$

We know that, for a 1D symmetrical random walk, ${\displaystyle p = 1/2, q = 1/2}$, with equal walk step length, after n steps, its probability distribution will be proportional to binomial coefficient:

${\displaystyle f(k) = {n \choose k}}$

or in the continuous limit, its probability distribution will be Gaussian-normal distribution.

My question is, which type of random walk can have "scaled" binominal coefficient distribution ? that is, its distribution is:

${\displaystyle g(k) = f(ak), a > 0}$

I tried different walk step length for "walk to the left" and "walk to the right", (asymmetrical random walk), I could not get above distribution.

Can anyone help ?

Thank you in advance.

$\endgroup$
2
  • $\begingroup$ So you mean $g(k)=\large{{\binom{n}{a\cdot k}}}$? $\endgroup$
    – callculus42
    Aug 6, 2018 at 18:31
  • $\begingroup$ That is what I mean $\endgroup$
    – david
    Aug 6, 2018 at 21:46

1 Answer 1

Reset to default
2
$\begingroup$

First of all, the random walk probabilities are proportional to $$\binom{n}{(n+k)/2},$$ not $\binom{n}k$. The walk is symmetric, so the formula needs to be symmetric when the sign of $k$ changes.

If you want to scale the distribution, just scale the step size. The simple random walk whose step sizes are $\pm 1/a$ will have distribution proportional to $$ \binom{n}{(n+xa)/2} $$ where $x$ is the position you are trying to find the probability of.

$\endgroup$
3
  • $\begingroup$ Thank you for the post. $\endgroup$
    – david
    Aug 6, 2018 at 22:35
  • $\begingroup$ Thank you for the post. Maybe my question is not clear. Here is what I am looking for. For example, original symmetric random walk, after walk 100 steps, the possible end positions can be [0, 1, 2, ...., 100] with probability f(k) where k = 0, 1, 2, ..., 100. I am looking for another random walk, after 100 steps, the possible end position can be [0, 1, ..., 10] with distribution g(k) = f(10k). What kind of random walk can satisfy above requirements ? Maybe I shall not call this "scaled" binomial coefficient, maybe i shall call it "squeezed" binomial coefficient. $\endgroup$
    – david
    Aug 6, 2018 at 22:47
  • 1
    $\begingroup$ I now understand your question and have no idea how to answer it. $\endgroup$
    – Mike Earnest
    Aug 6, 2018 at 23:05

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.