Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Probabilities involving Mins and Maxes

  1. Sep 16, 2011 #1
    I know for sure that P(Max(X, Y) < 1) = P(X < 1)P(Y < 1) if the two RVs are independent, but what does P(Min(X, Y) < 1) equal?

    Would it be P(Min(X, Y) < 1) = P(X < 1) + P(Y < 1)?

    Thanks
     
  2. jcsd
  3. Sep 16, 2011 #2

    Stephen Tashi

    User Avatar
    Science Advisor

    The probability of "X < 1 or Y < 1" involves the union of two sets. Have you sudied the formula for [itex] P(A \cup B) [/itex] ?
     
  4. Sep 16, 2011 #3

    chiro

    User Avatar
    Science Advisor

    Hey muzihc and welcome to the forums.

    Have you ever come across order statistics?

    http://en.wikipedia.org/wiki/Order_statistic
     
  5. Sep 16, 2011 #4
    Hi,

    I've studied P(A or B) = P(A U B) - I had a class in probability/statistics. I've never formally studied order statistics, though maybe I've overlapped with it at some point.

    The maximum case is kind of intuitive - if the max is less than t, everything else is. We can treat it as P(A and B), in the independent case. On the other hand, if the minimum is less than t, the other random variable isn't necessarily.

    I guess we could use P(Min(X, Y) < t) = 1 - P(Min(X, Y) >= t), in which case I assume it's 1 - P(X >= t)P(Y >= t).
     
  6. Sep 17, 2011 #5

    chiro

    User Avatar
    Science Advisor

    Hey muzihc.

    Take a look at order statistics. Order statistics helps you calculate the distribution of the minimum and maximum of a value given the number of samples that you have. It will help you in the first part of your question.
     
  7. Sep 17, 2011 #6

    Stephen Tashi

    User Avatar
    Science Advisor

    Let [itex] A = [/itex] the events where X < t [/itex]
    Let [itex] B = [/itex] the events where Y < t [/itex]

    [itex] P(A \cup B) = P(A) + P(B) - P(A \cap B) [/itex]

    So for independent events [itex] A [/itex] and [itex] B [/itex]

    [itex] P(A \cup B) = P(A) + P(B) - P(A) P(B) [/itex]


    I agree.

    [itex] 1 - (1 - P(A)) (1-P(B)) = 1 - (1 - P(A) - P(B) + P(A)P(B)) = P(A) + P(B) - P(A)P(B) [/itex]

    As chiro says, "order statistics" are the topic you should study if you want to tackle more complicated versions of the situation. For example: "In a sample of 4 values drawn from 4 independent random variables A,B,C,D, what is the probability that the second largest value is less than 1?"
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook