An evil salesman has offered you a magic stone. You don't know the value of this stone right now—the only thing you know is that it is worth between and coins inclusive—only integer values permitted—and that the probability that the stone is worth is . You have to pay the salesman between to coins, after which the salesman tells you the actual value of the stone. If the number of coins you give is greater than or equal to the value of the stone, you get the stone; if not, you don't get it. In either case, no change is given.
For example, if you pay the salesman coins and the stone is worth coins, you will receive the stone and have a net loss of coins. However, if you pay the salesman coins, you will not receive the stone and your net loss is coins.
You know that paying only one coin will ensure the least expected loss, but how many coins would you have to pay in order to have the greatest expected loss?
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
The expected loss when you pay p coins, E ( p ) , can be calculated as follows:
When p is smaller than the value of the stone, v : expected loss of p with probability 1 − 2 p ( p + 1 ) .
When p is greater than or equal to v : expected loss of p − v with probability 5 5 v for all 1 ≤ v ≤ p .
This means that E ( p ) = p ( 1 − 2 p ( p + 1 ) ) + ∑ i = 1 p 5 5 i ( p − i ) .
This can be simplified to E ( p ) = p − ∑ i = 1 p 5 5 i 2 = p − 3 3 0 2 p 3 + 3 p 2 + p .
When the derivative of this function is zero, we find a maximum point. The derivative is 1 − 3 3 0 6 p 2 + 6 p + 1 and is zero when p = 6 . 9 2 2 .
As E ( 6 ) = 4 . 3 4 5 and E ( 7 ) = 4 . 4 5 5 , you will have the greatest expected loss when you pay 7 coins.