You are viewing a single comment's thread from:

RE: Want to solve a real world math problem about Pokemon Go? Here you go!

in #mathematics7 years ago

I'll give you an answer and try to explain where it comes from and how it can be generalized to any collecting problem.

It will take on average 100 "Unknown" for a player to collect all the 26 different forms. The reason is the following :

  • It takes only one "unknown" to find the first form, no problem with that.
  • Suppose that you have found one form already.For each different try, you will have 1 out of 26 chances to pick the same one you already have. So the probability for it to take more than n tries to find a new one is (1/26)^n. In other terms, if we write T the number of tries, P(T>n)=(1/26)^n. This is a very usefull thing to know, because we can use it to compute the expected value of the number of days it will take to find a different form than the first one. The exepected value of T is noted E[T], and the formula is

E[T]=P(T>0)+P(T>1)+P(T>2)+P(T>3)+...
=1+(1/26)+(1/26)^2+(1/26)^3+...

(the dots mean that the sum goes on forever)

  • In the same way, if you already have found k forms and want to find a new one among the 26-k that remains : there are k chances out of 26 to pick on you already have for each try. The probability for it to take more than n tries to find a new one is (k/26)^n, and the expected value is

E[T]=1+(k/26 )+(k/26)^2+(k/26)^3+....

Fortunately, we don't have to make all these calculation by hand, thanks to a very well known formula :

1+q+q^2+q^3+...+q^n=(1-q^(n+1))/(1-q)

Here, q=k/26 where k is the number of forms we have already found. As q<1, the limit of q^n as n approaches infinity is 0, so the formula becomes

1+q+q^2+q^3+...= 1/(1-q)

  • Let's get back to our problem : when we have already found k forms, the expected value of the number of days it will take to find a new one is 1+1/26+...=1/(1-k/26)=26/(26-k). Say you have found 13 forms out of 26, it will take an average of 26/(26-13)=2 days to find a new one.

  • The final answer : the expected value of the number of days it will take to find ALL of the forms is the sum of the expected value to find a (k+1)th once you have found k forms, that is

E=1+26/25+26/24+...+26/2+26/1

E=26*(1+1/2+...+1/26)

The result of this computation gives approximately 100, and this formula works for any problem of collecting n different things with equal chances of finding each one of them at each tries : it takes on average n*(1+1/2+...+1/n).

  • For large values of n, a good approximation of this number is n*ln(n)

To learn more about this, you can look online for "Coupon collector's problem" ;)

I apologize in advance for any english mistake as it is not my mother tongue

Sort:  

That's awesome - thanks for taking the time to provide a solution :-)

I can confirm 100 days with a numerical simulation to find all 'unown's' on average !

Glad I could help! I was about to do a numerical simulation as well just to make sure I didn't make any mistake hahah

That's great! Where did you learn this? I want to know more :D