Htykuut

Question: For the random variable X, where E(X) = 0 and Var(X) = 1, derive the upper bound on the probability of the event ${|X-.6| > .2}$.

My attempt: Using Chebyshev's Inequality- Prob($|X-E(X)| > epsilon) leq frac{Var(X)}{epsilon^2} = frac{1}{(.2)^2} = 25.$

A) With no distributional assumption is this correct? I am thinking that because the probability is greater than 1 then this is a trivial inequality. I have seen alternate definitions of Chebyshev's inequality where $epsilon$ is $k sigma$ and if k is greater than or equal to 1 then this happens. However, 25 seems like an unusual number to get. Am I missing something?

B) As a follow up and for further understanding, how would the probability change if X was normally distributed with mean = 0 and variance = 1? Using the probability normal approximation integral I get the probability to be 3.96953... Although 25 and 3.96 are very different, is accuracy even a question since both probabilities are greater than 1?

You misused Chebyshev's inequality. You computed $P(|X|> .2)$ not $P(|X-.6|>.2)$. Using Markov's inequality (i.e. Chebyshev's) we get $$P(|X-.6|>.2)=P((X-.6)^2>.04)lefrac{E(X-.6)^2}{.04}=frac{1-0+.36}{.04}$$

Still, this is greater than $1$. This is to be expected. Consider $X$ to be a coinflip with $50/50$ shot of $pm 1$. Then $E(X)=0, operatorname{var}(X)=1$. And for all $varepsilon<1$ we have $$P(|X|>varepsilon)=1$$

You can modify this for the case of $|X-.6|$.

So the best bound we can get is $1$ without further information.