# Sample Data from Distributions

#### statty

##### New member
I have a question that is confusing me.

If traffic in a network is being generated from a Pareto distribution,
And I am estimating this traffic rate at each packet arrival,
If I was to take the previous x estimated rates as a sample of data,
Would the distribution of these estimated rates still be Pareto or would they be more of a Normal distribution?

#### springfan25

##### New member
if you take random samples from a pareto distribution, the individual values you draw will each be pareto distributed.

However the central limit theorum says that the sample average will be approximately normal (if suitable conditions are met).

ie, suppose you draw random values $$x_1, x_2, x_3, x_4...x_n$$.

Each value $$x_i$$ is pareto distributed. However the sample average $$\bar{x} = \frac{\sum x_i}{n}$$ is approximately normal

Does that make the distinction clearer?

#### statty

##### New member
Yes thats great thank you