On the variance in the Bernoulli scheme

43 Views Asked by At

I'm trying to solve the following problem.

There are $755$ cards, each of which has a number from $1$ to $755$ (each number occurs exactly once). Of these cards, $20$ pieces are randomly selected and a random value equal to the sum of the numbers written on all $20$ cards is considered. I want to find the variance of this random value.

I think in this problem there is a Bernoulli scheme in which the space of elementary events is a set of numbers from $1$ to $755$. But are there ways to calculate the parameters of such a distribution (e.g. variance)?

1

There are 1 best solutions below

0
On

With $n$ the number of cards and $k$ the number drawn we have the OGF

$$[u^k] \prod_{q=1}^n (1+ux^q).$$

Differentiate once for the expectation

$$[u^k] \prod_{q=1}^n (1+ux^q) \sum_{q=1}^n \frac{q u x^{q-1}}{1+ux^q}$$

and set $x=1$:

$$[u^k] \prod_{q=1}^n (1+u) \sum_{q=1}^n \frac{q u}{1+u} = [u^k] (1+u)^{n-1} u \frac{1}{2} n (n+1) \\ = \frac{1}{2} n (n+1) {n-1\choose k-1}.$$

This gives for the expectation

$$\frac{1}{2} n (n+1) {n-1\choose k-1} {n\choose k}^{-1} = \frac{1}{2} n (n+1) {n-1\choose k-1} \frac{k}{n} {n-1\choose k-1}^{-1}$$

or

$$\bbox[5px,border:2px solid #00A000]{ \mathrm{E}[X] = \frac{1}{2} k (n+1).}$$

Differentiate again for the next factorial moment

$$[u^k] \prod_{q=1}^n (1+ux^q) \left(\sum_{q=1}^n \frac{q u x^{q-1}}{1+ux^q}\right)^2 \\ + [u^k] \prod_{q=1}^n (1+ux^q) \sum_{q=1}^n \frac{q (q-1) u x^{q-2}}{1+ux^q} - [u^k] \prod_{q=1}^n (1+ux^q) \sum_{q=1}^n \frac{q u x^{q-1}}{(1+ux^q)^2} q u x^{q-1}$$

and once more set $x=1$:

$$[u^k] (1+u)^n \left(\sum_{q=1}^n \frac{q u}{1+u}\right)^2 \\ + [u^k] (1+u)^n \sum_{q=1}^n \frac{q (q-1) u}{1+u} - [u^k] (1+u)^n \sum_{q=1}^n \frac{q u}{(1+u)^2} q u \\ = {n-2\choose k-2} \frac{1}{4} n^2 (n+1)^2 \\ + {n-1\choose k-1} \frac{1}{3} (n-1) n (n+1) - {n-2\choose k-2} \frac{1}{6} n (n+1) (2n+1).$$

This gives for the second factorial moment

$$ \frac{k(k-1)}{n(n-1)} \frac{1}{4} n^2 (n+1)^2 \\ + \frac{k}{n} \frac{1}{3} (n-1) n (n+1) - \frac{k(k-1)}{n(n-1)} \frac{1}{6} n (n+1) (2n+1) \\ = \frac{k(k-1)}{n-1} \frac{1}{4} n (n+1)^2 + k \frac{1}{3} (n-1) (n+1) - \frac{k(k-1)}{n-1} \frac{1}{6} (n+1) (2n+1).$$

or

$$\bbox[5px,border:2px solid #00A000]{ \mathrm{E}[X(X-1)] = \frac{1}{12} k (n+1) (3 k n + 2 k + n - 6).}$$

Finally recall that

$$\mathrm{Var}[X] = \mathrm{E}[X(X-1)] + \mathrm{E}[X] - \mathrm{E}[X]^2$$

so that

$$\bbox[5px,border:2px solid #00A000]{ \mathrm{Var}[X] = \frac{1}{12} k (n+1) (n-k).}$$