Expected Value

thepillow

New member
Joined
Sep 12, 2012
Messages
34
I was hoping someone could tell me if my work on the question below is correct.



Q: If we roll a normal die \(\displaystyle n\) times, what is the expected number of different sides of the die that we observe?




And here's my attempt at a solution:

Let \(\displaystyle S_i = 1\) if side \(\displaystyle i\) is observed at least 1 time, and \(\displaystyle S_i = 0 \) if side \(\displaystyle i\) is not observed, where \(\displaystyle i = 1, 2, \dots, 6 \). Then for all six possible values of \(\displaystyle i\) we have


\(\displaystyle Pr(S_i = 1) = 1 - Pr(S_i = 0) = 1 - \left(\frac{5}{6}\right)^n \),​

since the probability that a side is not observed on each roll is 5/6 and we roll \(\displaystyle n\) times.

And the expected value of each of the six \(\displaystyle S_i\)'s is therefore

\(\displaystyle E(S_i) = 0\cdot Pr(S_i = 0) + 1 \cdot Pr(S_i = 1) = 1 - \left(\frac{5}{6}\right)^n\).​

Now, let \(\displaystyle N \) be the number of different sides observed. Then we can write \(\displaystyle N \) as

\(\displaystyle N = S_1 + S_2 + \dots + S_6 \).​

So the expected number of different sides observed is just the expected value of \(\displaystyle N\), which is

\(\displaystyle E(N) = E(S_1 + \dots + S_6) = E(S_1) + \dots E(S_6) = 6\cdot E(S_1) = 6\left(1 -\left(\frac{5}{6}\right)^n\right) \).



Thanks for your comments!​
 
I was hoping someone could tell me if my work on the question below is correct.



Q: If we roll a normal die \(\displaystyle n\) times, what is the expected number of different sides of the die that we observe?




And here's my attempt at a solution:

Let \(\displaystyle S_i = 1\) if side \(\displaystyle i\) is observed at least 1 time, and \(\displaystyle S_i = 0 \) if side \(\displaystyle i\) is not observed, where \(\displaystyle i = 1, 2, \dots, 6 \). Then for all six possible values of \(\displaystyle i\) we have


\(\displaystyle Pr(S_i = 1) = 1 - Pr(S_i = 0) = 1 - \left(\frac{5}{6}\right)^n \),​

since the probability that a side is not observed on each roll is 5/6 and we roll \(\displaystyle n\) times.

And the expected value of each of the six \(\displaystyle S_i\)'s is therefore

\(\displaystyle E(S_i) = 0\cdot Pr(S_i = 0) + 1 \cdot Pr(S_i = 1) = 1 - \left(\frac{5}{6}\right)^n\).​

Now, let \(\displaystyle N \) be the number of different sides observed. Then we can write \(\displaystyle N \) as

\(\displaystyle N = S_1 + S_2 + \dots + S_6 \).​

So the expected number of different sides observed is just the expected value of \(\displaystyle N\), which is

\(\displaystyle E(N) = E(S_1 + \dots + S_6) = E(S_1) + \dots E(S_6) = 6\cdot E(S_1) = 6\left(1 -\left(\frac{5}{6}\right)^n\right) \).



Thanks for your comments!​
Of all concepts in statistics, "Expectation Value" comes closest to following common sense. You can get away with some lack of rigor and still have the right answer. This all looks great to me! The only worry I have is correlation .. if one side is seen, then the others are not. I haven't completely convinced myself that the six S_i are independent, but the symmetry demands that all be equal - which must (?) in turn show that any correlation effects cancel out. I believe it, but haven't eliminated all the fuzz in my mind.

So I will say, "Very nice and convincing proof"!
 
Top