1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Question about infinite probability?

  1. Jun 10, 2012 #1
    i was just wondering (theoretically thinking), if I had to pick "option 1" from an infinite amount of options and my pick was random and I have an infinite amount of time (kind of like taking a ball from a bag, don't know which one is which) but I have to get exactly "option 1" and not any other.
    Is their any physical chance that i will ever get the option i want or is it impossible because an infinite amount of options means that your probability is to small to exist?

    Just wondering this, it isn't part of my homework or textbook.:biggrin:
  2. jcsd
  3. Jun 10, 2012 #2
    In general, if there are infinitely many events and you are considering the probability to select finitely many, then the probability is zero. Else way, it is the density of the set of selections in the set of events.

    Given that you have infinite time though, you will absolutely pick option 1. Same goes with all other finite sets of options. If you had finite time, then we could not say the same.
  4. Jun 10, 2012 #3
    Everything depends on the actual probability distribution. How likely is it to pick option 1??

    Also, are you working with countable infinite or uncountable??

    These things are needed to give a suitable answer to your question.
  5. Jun 10, 2012 #4


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    The trouble is you cannot have a uniform distribution over a countably infinite set. All probabilities would be zero, and the total probability would be zero, which is not valid.
    You could redefine it as a limit problem as the size of the set tends to infinity. To complete the definition, you need to specify how the time (number of trials) available depends on the set size, n, and whether the ball is replaced after each trial. E.g. if you specify n trials, with replacement, the probability of success tends to 1 - 1/e.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook