I have a problem:

I can choose twice randomly numbers from 1 to 3 and probabilities to choose each numbers can be adjusted for best end result. Choosing a number is independent. But not equally likely. I can define probability for each number.

If I choose twice same number I get 0 points otherwise I get sum of two choosen numbers. How to calculate best, optimal probability for each number to get max possible sum? And how to calculate expected value after choosing twice?

I tried like this: - Choosing numbers twice from 1 to 3 is actually getting me sum of 2 to 6 - so I got sum matrix: \begin{matrix} 0 & 3 & 4 \\ 3 & 0 & 5 \\ 4 & 5 & 0 \\ \end{matrix}

I think then it would be probabilities for numbers that represent sum like for 3,4 and 5 probability is 2/9, but that looks awfuly wrong.

  • 1
    $\begingroup$ Please include your thoughts and efforts in this and future posts. Formatting tips here. $\endgroup$ – Em. Jan 16 '16 at 23:25
  • $\begingroup$ What do you mean by the best, optimal probability for each number to get max possible sum? $\endgroup$ – Jendrik Stelzner Jan 16 '16 at 23:30
  • $\begingroup$ probablyme Sorry for posting badly formatted and without my toughts / tries, will try to do better in next posts. @JendrikStelzner best in way that I get highest possible expected value, but so I don't get high probability of getting same number twice coz then I get 0 as end result. $\endgroup$ – Cazper Jan 16 '16 at 23:34

If choosing a number is independent and equally likely, then if $S$ is the number of points you get, then using your table, the distribution of $S$ is \begin{array}{r|ccc} s&0&3&4&5\\\hline P(S=s)&\frac{3}{9}&\frac{2}{9}&\frac{2}{9}&\frac{2}{9} \end{array} So, to calculate the expectation of $S$, we have $$E[S] = 0(3/9)+3(2/9)+4(2/9)+5(2/9)=\frac{8}{3}$$

  • $\begingroup$ Thanks for help! Choosing a number is independent. But not equally likely, I can adjust that some numbers are less likely and some more likely. So I can define my probabilities of how often each number will be choosen. That bothers me, as I don't know what distribution of probabilities should I use. Obviously highest number will have higher probability to be choosen, 2nd highest a little less, and the rest either 0 or something low. But I can't figure out logic how to choose these probabilities. $\endgroup$ – Cazper Jan 16 '16 at 23:45
  • $\begingroup$ @Cazper You should include in your post that the numbers are not equally likely that you are defining the distribution on the numbers $\{1,2,3\}$. At least, that's what it sounds like you are saying. $\endgroup$ – Em. Jan 16 '16 at 23:52
  • $\begingroup$ Yes, you are right. I added clarification to problem. Ty! $\endgroup$ – Cazper Jan 16 '16 at 23:55

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.