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!

Homework Help: Tsebyshev's probability

  1. Sep 16, 2016 #1
    1. The problem statement, all variables and given/known data
    The average lifetime of a product ##T=7.5## (years). The variance of the lifetime ##\sigma^{2} = 0.41##.

    Using Tsebyshev's inequality, determine the lower bound for the probability, that the product lasts at least 5 years.

    2. Relevant equations

    Tsebyshev's inequality:
    P(|X-\mu | \geq t) \leq \frac{\sigma^{2}}{t^2}
    P(|X-\mu | < t) \geq 1 - \frac{\sigma^{2}}{t^2}
    where ##\mu## is the expected value, ##\sigma^2## is the variance and ##t## is the deviation from the expected value.

    3. The attempt at a solution

    We want
    P(T \geq 5)
    &\geq P(5 \leq T \leq 10)\\
    &= P(|T-7.5| \leq 2.5) &|t=2.5\\
    &\geq 1 - \frac{\sigma^2}{t^2} &| Tsebyshev\\
    &= 1-\frac{0.41^2}{2.5^2}\\
    &\approx 0.973
    This is apparently not the correct answer, and I'm not sure what I'm doing wrong. We obviously want the deviation from the mean to be less than 2.5; otherwise ##|T-\mu|## would produce a value that is below 5.
  2. jcsd
  3. Sep 16, 2016 #2


    User Avatar
    2017 Award

    Staff: Mentor

    Edit: Why did you square the 0.41?

    Some manual tuning: The worst case is some fraction p failing "just before" 5 years and all other machines failing after X years, X a bit larger than 7.5

    The condition for the mean is then 5p+(1-p)*X = 7.5.
    The condition for the variance is p*2.5^2 + (1-p)*(X-7.5)^2 = 0.41

    Solving gives p=0.062 and X=7.66.
    Last edited: Sep 16, 2016
  4. Sep 16, 2016 #3
    You can use it for both. One gives a lower bound, the other gives an upper bound.
  5. Sep 16, 2016 #4
    You are given ##\sigma^2 = 0.41##. But later on, you write ##0.41^2## at the end of your computation. There is no need for that square.
  6. Sep 16, 2016 #5
    So what I could do is calculate ##P(T<5)## and go from there?

    &\leq P(T<5 \text{ and } T>10)\\
    &= P(|T-\mu| >= 2.5)\\
    &\leq \frac{\sigma^2}{2.5^2}\\
    &= 0.026896
    But then I'm stuck with the shaded area shaded blue in this picture:
    and I need the one in the middle. At least that's my understanding.
  7. Sep 16, 2016 #6
    Oh. My. Gauss!

    Thank you. How typical of me...
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted