# Basic question about the normal distribution

#### Indecisive

##### New Member
So it seems like different distributions have different kinds of "experiments" associated with them, like how you use the binomial distribution when looking for the probability of k wins out of n trials with probability p. What I'm wondering is what kind of experiment is associated with the normal distribution, and what values you need given to you before you can use it (like, I know you use the Poisson distribution when you have a binomial-type distribution with an infinite number of trials, but you already have the expected value).

I'm also interested in finding out how the p.d.f. of the normal distribution can be derived from the p.m.f. of the binomial distribution, if anyone has a link to a proof or something.

#### Indecisive

##### New Member
I don't understand how the binomial distribution "approaches" the normal distribution as the number of trials goes to infinity. Isn't the "hump" in a bar graph of the binomial distribution at least in part a result of the finite number of trials?

#### spunky

##### Can't make spagetti
here's a very useful website i borrowed from somebody else's post showing you how the deMoivre-Laplace theorem is used to approximate the normal distribution from the binomial distribution... the proof is, i think, quite elegant and the author of the wiki article shows all the steps to make it easy to follow..

http://en.wikipedia.org/wiki/De_Moivre–Laplace_theorem