Binomial distribution, when variable isn't x

58 Views Asked by At

I've been using the formula $$p(x,N)=\frac{N!}{(\frac{N+x}{2})!(\frac{N-x}{2})!} p^{1/2(N+x)} q^{1/2(N-x)}$$ to determine the probability for a dog who walks in a straight line and can either move $x_{n+1}=x_n+1$ (to right) or x_{n+1}=x_n-1 ( to left) for different values of $p, N,$ and $x$.
$p$ is probability of success (in this case probability of moving to right),
$N$ is total number of steps, and
$x$ is final position after $N$ steps. How could I change this so that instead of steps to the right being $x_{n+1}=x_n+1$, steps to right is $x_{n+1}=x_n+4$, but steps to left is $x_{n+1}=x_n-1$?

For example, say the probability of $x_{n+1}=x_n+4$ is $p=0.3$, therefore probability of going $x_{n+1}=x_n-1$ is $q=0.7$. How could I change this formula for that case?? (I am assuming I can still use binomial distribution for steps that aren't of equal length but if this isn't the case someone let me know)

Thank you

1

There are 1 best solutions below

0
On

yes you can still use the binomial distribution! My advise to you is to start with the binomial distribution in its most general form:

$$ p(M,N) = \frac{N!}{M!(N-M)!} p^M q^{(N-M)} $$

where $p$ is the probability of success, $q$ the probability of failure, $N$ the total number of steps and $M$ the number of successes.

Now you can make specific choices. For example that success corresponds to $x = 4$ and failure to $x = -1$. Then the total length traveled equals $4M + (N-M)-1$. You are free to adjust your probabilities $p$ and $q$. Ultimately you can rewrite your equation in terms of $x$ instead of $M$.