Jensen’s Inequality: \(\phi(\mathbb{E}X) \leq \mathbb{E} \phi(X)\) for convex \(\phi\), with strict inequality for strictly convex \(\phi\) and nondegenerate \(F_X(x)\).
Markov’s Inequality: \(\Pr(Y \geq t) \leq \frac{\mathbb{E}Y}{t}\) for \(Y, t\) non-negative. Prove by defining dummy variable \(f(y) = 1\) if \(y \geq t\), \(0\) otherwise, then taking expectation over that.
Only really useful for deriving other inequalities. Sharpest bound possible given only non-negativity.
Chebyshev’s Inequality: \(\Pr(|X - \mu| \geq t\sigma) \leq \frac{1}{t^2}\). Prove by considering the square of both sides in the probability.