Quit while you're ahead

You play the following game:

Roll a fair 6-sided die as many times as you like.

  • If it's any integer between 1 to 5 inclusive, you win $100.
  • If it's a 6, you lose all your money and the game ends.

You can continue to roll, winning $100 every time it's not a 6. You can also decide to quit at any time, having not yet rolled a 6, and keep all of your winnings.

What is the optimal number of rolls at which you should "call it quits"? i.e. For what number, n n , should you decide ahead of time, "I'm gonna continue to roll until I've done n n rolls (assuming I haven't lost all my money yet) and then I should quit."

Give your answer as the sum of all the numbers for which you maximize your expected value of the game. So, for example, if there are two such numbers for which you maximize your expected payoff, m m and n n , please give your answer as the sum m + n m+n .


Image credit: www.geekvintage.com


The answer is 11.

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.

2 solutions

Geoff Pilling
Jul 12, 2016

In general, your expectation value for a strategy where you quit after rolling n n times is given by:

E n = n × ( 5 6 ) n × $ 100 E_n = n \times (\frac{5}{6})^n \times \$100

This is maximized for n = 5 n = 5 and n = 6 n = 6 .

5 + 6 = 11 5 + 6 = \boxed{11}

Kai Ott
Jul 13, 2016

We have the function E ( n ) = 100 $ n ( 5 6 ) n E(n) = 100\$ \cdot n \cdot {(\frac{5}{6}})^n . There are 3 ways to find out when this function is maximized.

Way 1: Plug in the numbers 1,2,3,4... until you found the maxima

Way 2: Find the derivative E ( n ) = ( 5 6 ) n ( l n ( 5 6 ) + 1 ) E'(n) = ({\frac{5}{6}})^n \cdot (ln( \frac{5}{6} ) + 1) and solve for E'(n)=0. This will yield a number between 5 and 6 which means you have to check if E(5) or E(6) is or if they are equal.

Way 3: From the function for E(n) we can find the recursive definition E ( n ) = E ( n 1 ) 5 6 n n 1 E(n) = E(n-1) \cdot \frac{5}{6} \cdot \frac{n}{n-1} we then solve for 5 6 n n 1 = 1 \frac{5}{6} \cdot \frac{n}{n-1} = 1 because that's the value for n where E(n) is just about to decrease again. Solving will yield n = 6 n=6 which means that E ( 5 ) = E ( 6 ) E(5) = E(6) are the maxima of the function E (for natural n).

In my opinion way 3 > way 2 > way 1.

Thanks for your insights, Aaa! (I used Way 3... ;) )

Geoff Pilling - 4 years, 11 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...