Expectations Between Jokers

A standard 52 52 -card deck with n n extra jokers (where n 2 n\ge 2 ) is shuffled. Given the expected number of cards between the first joker and the last joker exclusive is a positive integer E E , find the sum of all possible values of E E .

Details and Assumptions

The 52 52 -card deck with n n extra jokers implies that there are 52 + n 52+n cards in total.


The answer is 447.

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

Daniel Liu
Apr 16, 2014

Seeing that we want to find all possible integer n n such that E E is an integer, we should find a general formula for E E in terms of n n .

We can see that by the Law of Symmetry, the n n jokers will evenly divide the 52 52 -card deck into n + 1 n+1 sections. Each of these sections would thus have 52 n + 1 \dfrac{52}{n+1} cards in them.

To count the number of cards between the first joker and last joker exclusive, we can use complementary counting. Instead, let's try to count the number of cards before the first joker inclusive, and after the last joker inclusive.

Clearly, the number of cards before the first joker exclusive is 52 n + 1 \dfrac{52}{n+1} , as said before. Adding in this extra joker gives 52 n + 1 + 1 \dfrac{52}{n+1}+1 .

This is exactly the same as the expected number of cards after the last joker inclusive. Thus, the total number of cards in our complementary counting is 2 ( 52 n + 1 + 1 ) = 104 n + 1 + 2 2\left(\dfrac{52}{n+1}+1\right)=\dfrac{104}{n+1}+2 .

The total number of cards is 52 + n 52+n . Thus, the number of cards between the first joker and the last exclusive is E = ( 52 + n ) ( 104 n + 1 + 2 ) = 50 + n 104 n + 1 E=(52+n)-\left(\dfrac{104}{n+1}+2\right)=50+n-\dfrac{104}{n+1} .

Now we must make 50 + n 104 n + 1 50+n-\dfrac{104}{n+1} an integer. Clearly, 50 + n 50+n is always an integer; thus it suffices to make 104 n + 1 \dfrac{104}{n+1} an integer.

Note that the factors of 104 104 are 1 , 2 , 4 , 8 , 13 , 26 , 52 , 104 1,2,4,8,13,26,52,104 . We can see that by letting n + 1 n+1 equal any of these factors, E E turns into an integer.

Thus, n + 1 = 1 , 2 , 4 , 8 , 13 , 26 , 52 , 104 n+1=1,2,4,8,13,26,52,104 , or n = 0 , 1 , 3 , 7 , 12 , 25 , 51 , 103 n=0,1,3,7,12,25,51,103 . Heeding the restriction n 2 n\ge 2 , we see the possible values of n n is n = 3 , 7 , 12 , 25 , 51 , 103 n=3,7,12,25,51,103 .

Plugging each of these in and solving for E E , we get E = 27 , 44 , 54 , 71 , 99 , 152 E=27,44,54,71,99,152 . Thus, our final answer is 27 + 44 + 54 + 71 + 99 + 152 = 447 27+44+54+71+99+152=\boxed{447} and we are done.

Given that this question is essentially the Law Of Symmetry, can you prove it? Why would there be 52 n + 1 \frac{52}{n+1} cards on expectation?

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

Yesterday I was wondering how to prove it but today I realized a very easy way:

Suppose that instead of inserting jokers randomly in the deck of cards, we insert cards randomly in between the jokers. Obviously, the expected value of the number of cards between each joker would be the same (think of it as randomly putting marbles into urns); therefore our original claim is correct.

Daniel Liu - 7 years, 1 month ago

Log in to reply

Great explanation. Sometimes, choosing the correct perspective makes the problem accessible.

Calvin Lin Staff - 7 years, 1 month ago

Log in to reply

@Calvin Lin Hmm. I'm almost there, but how do you explain how we know it's equivalent to placing marbles in urns? How do we know that the "end" urns are just as likely as the middle ones?

Kevin Bourrillion - 7 years, 1 month ago

Log in to reply

@Kevin Bourrillion Well, the jokers act as a divider between the two urns. If you thought of the urns as a row in an ice tray, would it make more sense?

Daniel Liu - 7 years, 1 month ago

Log in to reply

@Daniel Liu Although the use of symmetry is fully justified: it is easy to see why it is troubling in this example. Unlike most simple urn models where a new card goes depends on the results of earlier choices. The probability of a particular card slotting between a given pair of jokers is not fixed at 1 k + 1 \frac{1}{k+1} , if you keep track of where the other cards have gone.

Here is a little more detail about the urn model which is completely analogous to this problem, and which I think helps one see exactly why symmetry is a valid argument.

Start with a bag of k + 1 k+1 balls each of a different colour. (The colours correspond to the k 1 k-1 slots between the jokers and the 2 2 slots outside.) Then draw 52 balls from the bag, and whenever a ball is drawn it is replaced along with a new ball of the same colour. (This corresponds to the fact that when you place a card between the a pair of jokers you increase the number of slots where you can place subsequent cards between those jokers by one.)

The results of different draws are dependent but since the rules are the same for each different colours, you would expect 52 k + 1 \frac{52}{k+1} of each, and hence 52 ( k 1 ) k + 1 \frac{52(k-1)}{k+1} of the cards will lie between the jokers.

Daniel is however fully justified in being relaxed about the dependencies involved, because of the nice result that the sum of expectations is the expectation of the sum whether or not the random variables are independent . If you do not take care of the result from previous rounds then the overall probability that a card lies between a given joker pair is fixed at 1 k + 1 \frac{1}{k+1} and you can just sum to get the expected number of balls.

David Vaccaro - 6 years, 10 months ago

Wait, wouldn't it have to be 52 + n n + 1 \frac{52+n}{n+1} ? I solved that and got that 2 2 and 16 16 both have integral solutions. You mentioned that when you add jokers, the amount in the deck rises. :P

Finn Hulse - 7 years, 1 month ago

Log in to reply

No, the 52 n + 1 \dfrac{52}{n+1} is the expected amount of cards between consecutive jokers. So no jokers are counted.

Daniel Liu - 7 years, 1 month ago

Log in to reply

Oh... I get it. Dang it.

Finn Hulse - 7 years, 1 month ago

Ah dangit, on the third try I figured it, but made a small arithmetic error and got 448. :-) Epic problem...

Kevin Bourrillion - 7 years, 1 month ago

I'm slightly confused as to why n = 2 doesn't work:

there are 54 cards in the deck with 2 jokers.

On average, we'd expect joker number 1 to be at card number 18, and joker number 2 to be at card number 36... => E = 17 in this case.

Where have I gone wrong here...?

Jeremy Ho - 5 years, 5 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...