According to this post, I found for $X \sim N(0,1)$, $x > 0$ the result that

\begin{align} \frac{1}{\sqrt{2\pi}}\big(\frac{1}{x}-\frac{1}{x^3}\big)e^{-\frac{x^2}{2}} \leq P(X>x) \leq \frac{1}{\sqrt{2\pi}}\frac{1}{x}e^{-\frac{x^2}{2}}. \end{align}

Now, an even stronger upper bound has to hold, namely \begin{align} P(X>x) \leq e^{-\frac{x^2}{2}}. \end{align} For $x \geq 1$ this is obvious. However, it seems to be rather difficult to show that this inequality does hold for $0<x<1$. What do you think?


$P(X > x) = P(e^{\lambda X} > e^{\lambda x})$ for all $\lambda$.

Now by Markov's inequality, \begin{align} P(e^{\lambda X} > e^{\lambda x}) &\leq \frac{\mathbb{E}[e^{\lambda X}]}{e^{\lambda x}} \\ &=e^{\frac{\lambda^2}{2} - \lambda x}. \end{align} So for $\lambda = x$, \begin{align} P(X > x) \leq e^{\frac{x^2}{2}-x^2} = e^{-\frac{x^2}{2}}. \end{align}


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.