The question:

Suppose we roll a fair 6 sided die with the number [1,6] written on them. After the first die roll we roll the die $k$ times where $k$ is the number on the first die roll. The number of points you score is the sum of the face-values on all die rolls (including the first). What is the expected number of points you will score?


There are six different cases for the first roll

  • Case 1: Roll 1

    • $P(D_1=1) =\frac{1}{6}$
    • The expected number on the next roll is $E(D_2)=\frac{1}{6}\times(1+2+3+4+5+6)=3.5$

    • $E(S_1) = 3.5+1=4.5$ points

  • Case 2: Roll 2

    • We know what the expected value of one roll is, and since rolling the die is independent we can use the previous expected value for the next roll.

    • $E(S_2) = (3.5\times 2)+2=9$ points

  • Case 3: Roll 3

    • $E(S_3) = (3.5\times 3)+3=13.5$ points
  • Case 4: Roll 4

    • $E(S_4) = (3.5\times 4)+4=18$ points
  • Case 5: Roll 5

    • $E(S_5) = (3.5\times 5)+5=22.5$ points
  • Case 6: Roll 6

    • $E(S_6) = (3.5\times 6)+6=27$ points

Therefore, the expected number of points scored is

$=\frac{1}{6}\times(4.5+9+13.5+18+22.5+27)=15.75$ points

  • 1
    $\begingroup$ I think this is a law of total expectation problem, $E[X]=E[E(X|K)]$. Where $K$ is the face of the first roll, and $X$ is the sum of the $K$ rolls $\endgroup$ – gd1035 Nov 16 '18 at 15:49
  • 1
    $\begingroup$ I think your attempt is correct. That is how I would have tried it anyways. $\endgroup$ – Sauhard Sharma Nov 16 '18 at 16:22

The expected value of a sum of random number $N$ of iid random variables $X_i$ is $$E\left[\sum_{i=1}^N X_i\right]=E[N]E[X_i]$$ In your case you add $E[N]$, so the answer is $$E[N]E[X_i]+E[N]=3.5\cdot 3.5+3.5 =4.5\cdot 3.5 = 15.75$$

  • $\begingroup$ Why do we add $E[N]$ to the summation? $\endgroup$ – Arthur Green Nov 16 '18 at 19:13
  • 1
    $\begingroup$ Because you include the 1st (actually zeroth) roll which determines $N$ into the sum, and because of linearity of expectations it adds $E[N]$. $\endgroup$ – kludg Nov 16 '18 at 20:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.