Is This Expected?

Game A
Roll a fair dice repeatedly until we land on a 1.
Let A denote the expected sum of all the values (including the final 1).

Game B
Roll a fair dice repeatedly until we land on a 6.
Let B denote the expected sum of all the values (including the final 6).

Which is larger, A or B?

B They are equal A

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.

1 solution

In general, suppose we keep rolling until we get the number N N . Then the expected value of each roll up until we roll N N is A N = 21 N 5 A_{N} = \dfrac{21 - N}{5} . As the die is fair, the probability that we roll N N on any given roll is 1 6 \dfrac{1}{6} , and the probability we don't is 5 6 \dfrac{5}{6} . If we take m 1 m \ge 1 rolls to get N N then the final sum will be N + ( m 1 ) A N N + (m - 1)A_{N} . The expected sum E N E_{N} is then

1 6 N + 1 6 5 6 ( N + A N ) + 1 6 ( 5 6 ) 2 ( N + 2 A N ) + . . . + 1 6 ( 5 6 ) m 1 ( N + ( m 1 ) A N ) + . . . . = \dfrac{1}{6}N + \dfrac{1}{6}*\dfrac{5}{6}(N + A_{N}) + \dfrac{1}{6}\left(\dfrac{5}{6}\right)^{2}(N + 2A_{N}) + ... + \dfrac{1}{6}\left(\dfrac{5}{6}\right)^{m-1}(N + (m - 1)A_{N}) + .... =

N 6 k = 0 ( 5 6 ) k + A N 6 k = 1 k ( 5 6 ) k = N 6 × 6 + A N 6 × 5 6 ( 1 5 6 ) 2 = N + 5 A N = N + ( 21 N ) = 21 \displaystyle\dfrac{N}{6}\sum_{k=0}^{\infty} \left(\dfrac{5}{6}\right)^{k} + \dfrac{A_{N}}{6}\sum_{k=1}^{\infty} k\left(\dfrac{5}{6}\right)^{k} = \dfrac{N}{6} \times 6 + \dfrac{A_N}{6} \times \dfrac{\dfrac{5}{6}}{\left(1 - \dfrac{5}{6}\right)^{2}} = N + 5A_{N} = N + (21 - N) = 21 .

So regardless of which number we roll until, the expected (inclusive) final sum is invariant, and thus the answer is A and B are equal \boxed{\text{A and B are equal}} .

(Note: A shortcut could have been to note that the expected number of rolls to get any given number is 6 6 , so the expected sum would be N + 5 A N = 21 N + 5A_{N} = 21 as found above, but I wasn't certain that this was airtight so I went with the long version to be sure.)

Nice solution (+1)

There's another solution that uses Doob's optional stopping theorem. It's a fairly advanced tool, but it yields the answer almost immediately.

Brian Moehring - 4 years, 2 months ago

Log in to reply

I hadn't heard of that theorem before, so thanks for mentioning it. From what I've just read, as this is a martingale the expected sum will be the same regardless of the designated "stopping time", or in this case stopping number, so we can conclude A and B are equal without making any calculations.

Brian Charlesworth - 4 years, 2 months ago

Log in to reply

We do have to be a little careful in defining the martingale, but if you let S n S_n denote the sum after n n rolls of the die, then M n = S n 7 2 n M_n = S_n - \frac{7}{2}n is a martingale with bounded increments. If you then let the N N th roll be the first time any chosen face shows up, then the optional stopping theorem tells you E [ M N ] = E [ M 0 ] = 0 E [ S N ] = 7 2 E [ N ] = 7 2 ( 6 ) = 21 \mathbb{E}[M_N] = \mathbb{E}[M_0] = 0 \implies \mathbb{E}[S_N] = \frac{7}{2}\mathbb{E}[N] = \frac{7}{2}(6) = 21

Brian Moehring - 4 years, 2 months ago

There is an easier way to do the calculation using law of iterated expectation .

If we let X be the expected value of the sum before landing on a 1, we have X = X + 2 6 + X + 3 6 + X + 4 6 + X + 5 6 + X + 6 6 X = \frac{ X+2}{6} + \frac{X+3}{6} + \frac{X+4}{6}+ \frac{X+5}{6}+ \frac{X+6}{6} . Then, A = X + 1 A = X + 1 .

Calvin Lin Staff - 4 years, 2 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...